WEBVTT 00:00:00.000 --> 00:00:00.630 align:middle line:90% 00:00:00.630 --> 00:00:03.600 align:middle line:84% ATUL J. BUTTE: So I can and have in the past 00:00:03.600 --> 00:00:05.970 align:middle line:84% talked for about 6 hours on this subject. 00:00:05.970 --> 00:00:07.530 align:middle line:84% Today, we're just going to talk about the first of these, 00:00:07.530 --> 00:00:09.330 align:middle line:84% microbiology for the [? bioinformaticist. ?] 00:00:09.330 --> 00:00:11.700 align:middle line:84% And if we have time, then we can talk about gene measurement 00:00:11.700 --> 00:00:12.960 align:middle line:84% techniques, not just microarrays, 00:00:12.960 --> 00:00:14.585 align:middle line:84% but all sorts of different technologies 00:00:14.585 --> 00:00:16.180 align:middle line:84% depending on how fast or slow we go. 00:00:16.180 --> 00:00:19.090 align:middle line:90% So let's just get started then. 00:00:19.090 --> 00:00:22.410 align:middle line:84% So I have about 10 slides of basic biology, OK? 00:00:22.410 --> 00:00:25.770 align:middle line:84% If I see you start to yawn, then we'll just go faster. 00:00:25.770 --> 00:00:27.900 align:middle line:90% There's a lot that we can cover. 00:00:27.900 --> 00:00:30.660 align:middle line:84% So how many of you remember this kind of thing? 00:00:30.660 --> 00:00:31.720 align:middle line:90% Do you remember this? 00:00:31.720 --> 00:00:32.400 align:middle line:90% Do you remember this? 00:00:32.400 --> 00:00:32.940 align:middle line:90% Do you remember this? 00:00:32.940 --> 00:00:34.440 align:middle line:90% OK, so we can go pretty fast. 00:00:34.440 --> 00:00:37.290 align:middle line:84% So obviously, the key problem in biology 00:00:37.290 --> 00:00:39.300 align:middle line:84% or in all of life science is that organisms 00:00:39.300 --> 00:00:41.850 align:middle line:84% need to produce proteins for a variety of different things 00:00:41.850 --> 00:00:43.558 align:middle line:84% that they need to do over their lifetime. 00:00:43.558 --> 00:00:46.590 align:middle line:84% So the goal is proteins, but the organisms 00:00:46.590 --> 00:00:48.510 align:middle line:84% have to make the same proteins again and again 00:00:48.510 --> 00:00:51.378 align:middle line:84% in a very set pattern type of way. 00:00:51.378 --> 00:00:52.920 align:middle line:84% Some of these proteins are important. 00:00:52.920 --> 00:00:54.450 align:middle line:84% They're enzymes that catalyze reactions. 00:00:54.450 --> 00:00:57.033 align:middle line:84% These reactions would otherwise take 10,000 years to complete. 00:00:57.033 --> 00:00:58.742 align:middle line:84% Because you have the enzyme [INAUDIBLE],, 00:00:58.742 --> 00:01:00.420 align:middle line:84% it can take 10,000 microseconds instead. 00:01:00.420 --> 00:01:02.580 align:middle line:84% Structural support-- otherwise, a bag of water 00:01:02.580 --> 00:01:03.420 align:middle line:90% would just collapse. 00:01:03.420 --> 00:01:06.045 align:middle line:84% These are things that are going to hold the cell wall together, 00:01:06.045 --> 00:01:07.110 align:middle line:90% to give it some shape. 00:01:07.110 --> 00:01:08.870 align:middle line:84% And certainly, we have hormones to signal 00:01:08.870 --> 00:01:11.370 align:middle line:84% from one part of an organism to another part of the organism 00:01:11.370 --> 00:01:13.630 align:middle line:84% or for one organism to another organism. 00:01:13.630 --> 00:01:16.470 align:middle line:84% So the key problem in life science over the past 50 years 00:01:16.470 --> 00:01:18.030 align:middle line:84% is how to encode the instructions 00:01:18.030 --> 00:01:19.988 align:middle line:84% for making these specific proteins because they 00:01:19.988 --> 00:01:23.220 align:middle line:84% have so many disparate shapes, disparate lengths, 00:01:23.220 --> 00:01:24.600 align:middle line:90% characteristics. 00:01:24.600 --> 00:01:26.730 align:middle line:84% How does the organism know or how do the cells 00:01:26.730 --> 00:01:27.850 align:middle line:90% know how to make these proteins? 00:01:27.850 --> 00:01:29.820 align:middle line:84% And the first step is obviously nucleotides. 00:01:29.820 --> 00:01:32.850 align:middle line:84% That's the most basic element of these blueprints that 00:01:32.850 --> 00:01:34.700 align:middle line:90% goes into making the proteins. 00:01:34.700 --> 00:01:37.635 align:middle line:84% So have adenine, cytosine, guanine, and thymine, H, C, G, 00:01:37.635 --> 00:01:41.790 align:middle line:84% T. And as we all know, 50 years ago we've 00:01:41.790 --> 00:01:44.670 align:middle line:84% learned that these were actually arranged in a chain, actually 00:01:44.670 --> 00:01:48.420 align:middle line:84% two anti-parallel chains where the As base pair with a Ts 00:01:48.420 --> 00:01:50.070 align:middle line:90% and the Cs go along with the Gs. 00:01:50.070 --> 00:01:52.260 align:middle line:84% And these are not just arbitrary chains, 00:01:52.260 --> 00:01:54.580 align:middle line:84% but these are chains with polarity. 00:01:54.580 --> 00:01:56.880 align:middle line:84% So you have a start and an end to one chain. 00:01:56.880 --> 00:01:59.940 align:middle line:84% It's lined up with the start and the end to the other chain. 00:01:59.940 --> 00:02:01.150 align:middle line:90% And this is the double helix. 00:02:01.150 --> 00:02:02.790 align:middle line:84% This naturally forms a double helix 00:02:02.790 --> 00:02:05.160 align:middle line:84% if you melt this to the right temperature. 00:02:05.160 --> 00:02:07.458 align:middle line:84% If you heat the DNA, it starts to denature. 00:02:07.458 --> 00:02:09.000 align:middle line:84% But if you cool it again, it'll start 00:02:09.000 --> 00:02:10.560 align:middle line:90% to come back together again. 00:02:10.560 --> 00:02:12.018 align:middle line:90% It'll re-nature. 00:02:12.018 --> 00:02:13.560 align:middle line:84% So it's a natural formation depending 00:02:13.560 --> 00:02:16.740 align:middle line:84% on the temperature of the bonds here, 00:02:16.740 --> 00:02:19.120 align:middle line:84% the temperature needed to break these bonds. 00:02:19.120 --> 00:02:21.540 align:middle line:84% And so you get an idea of what this double helix looks 00:02:21.540 --> 00:02:22.140 align:middle line:90% like here. 00:02:22.140 --> 00:02:25.330 align:middle line:90% 00:02:25.330 --> 00:02:27.000 align:middle line:90% So now, let's take a step back. 00:02:27.000 --> 00:02:28.080 align:middle line:90% We have the DNA. 00:02:28.080 --> 00:02:30.030 align:middle line:84% We know that's going to be the end point, 00:02:30.030 --> 00:02:33.360 align:middle line:84% but we haven't had to even sequence the As, 00:02:33.360 --> 00:02:38.020 align:middle line:84% Gs, and Ts, and Cs to make genomic or genetic diagnoses. 00:02:38.020 --> 00:02:40.110 align:middle line:84% So even for the past more than 50 years, 00:02:40.110 --> 00:02:42.450 align:middle line:84% we've been able to make genetic diagnoses 00:02:42.450 --> 00:02:44.340 align:middle line:84% even before sequencing was invented. 00:02:44.340 --> 00:02:44.880 align:middle line:90% How? 00:02:44.880 --> 00:02:46.200 align:middle line:84% Well, obviously, you can actually 00:02:46.200 --> 00:02:47.575 align:middle line:84% look at the chromosome structure. 00:02:47.575 --> 00:02:49.620 align:middle line:84% So you can get a blood sample from any human, 00:02:49.620 --> 00:02:52.080 align:middle line:84% isolate the white blood cells, which still has DNA, 00:02:52.080 --> 00:02:53.970 align:middle line:84% and stain them in a particular way, 00:02:53.970 --> 00:02:56.190 align:middle line:84% take a photograph, a photomicrograph of that, 00:02:56.190 --> 00:02:58.180 align:middle line:84% cut out the chromosomes, and line them up. 00:02:58.180 --> 00:03:00.420 align:middle line:84% And so people have been doing this for decades, 00:03:00.420 --> 00:03:02.200 align:middle line:90% more than half a century. 00:03:02.200 --> 00:03:03.990 align:middle line:84% And basically, these chromosomes have 00:03:03.990 --> 00:03:06.450 align:middle line:84% been numbered based on the original estimates of the size 00:03:06.450 --> 00:03:07.325 align:middle line:90% of these chromosomes. 00:03:07.325 --> 00:03:09.210 align:middle line:84% So chromosome 1 is thought to be the biggest. 00:03:09.210 --> 00:03:11.043 align:middle line:84% Chromosome 22 is thought to be the smallest, 00:03:11.043 --> 00:03:13.320 align:middle line:84% but it turns out that 21 is actually smaller than 22. 00:03:13.320 --> 00:03:16.230 align:middle line:84% And there's a couple other missed pairings like this now 00:03:16.230 --> 00:03:18.300 align:middle line:90% that we have the exact length. 00:03:18.300 --> 00:03:21.630 align:middle line:84% But this is how we've been able to make genetic diagnoses. 00:03:21.630 --> 00:03:25.080 align:middle line:84% Each chromosome, so each chromosome 00:03:25.080 --> 00:03:29.400 align:middle line:84% is a single double strand of DNA from end to end. 00:03:29.400 --> 00:03:31.980 align:middle line:84% And obviously, it's wound and coiled and rewound 00:03:31.980 --> 00:03:34.590 align:middle line:84% and recoiled in such a degree that this looks nothing 00:03:34.590 --> 00:03:37.260 align:middle line:84% like a double helix only because we're at a big picture here. 00:03:37.260 --> 00:03:39.630 align:middle line:84% We're at least five orders of magnitude 00:03:39.630 --> 00:03:41.700 align:middle line:84% up in terms of magnification from actually 00:03:41.700 --> 00:03:44.250 align:middle line:90% visualizing a double helix. 00:03:44.250 --> 00:03:47.940 align:middle line:84% Now, like I said, we've been able to make diagnoses 00:03:47.940 --> 00:03:49.620 align:middle line:90% for more than 50 years using-- 00:03:49.620 --> 00:03:52.020 align:middle line:90% come on in-- in genomics. 00:03:52.020 --> 00:03:56.310 align:middle line:84% Anyone want to guess what this diagnosis is? 00:03:56.310 --> 00:03:58.860 align:middle line:84% Let's keep this interactive here. 00:03:58.860 --> 00:03:59.760 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:03:59.760 --> 00:04:01.500 align:middle line:84% ATUL J. BUTTE: This is Down syndrome or trisomy 21. 00:04:01.500 --> 00:04:03.300 align:middle line:84% So you can see, obviously, it's highlight here 00:04:03.300 --> 00:04:04.200 align:middle line:90% with the little arrow. 00:04:04.200 --> 00:04:06.158 align:middle line:84% But there's three copies of chromosome 21 here. 00:04:06.158 --> 00:04:07.952 align:middle line:84% And that happens to be a viable phenotype. 00:04:07.952 --> 00:04:08.910 align:middle line:90% You can born with this. 00:04:08.910 --> 00:04:11.390 align:middle line:84% And you can actually live for quite some time, 00:04:11.390 --> 00:04:13.140 align:middle line:84% though you have marked phenotypic problems 00:04:13.140 --> 00:04:14.162 align:middle line:90% or phenotypic issues. 00:04:14.162 --> 00:04:15.120 align:middle line:90% You can live with that. 00:04:15.120 --> 00:04:16.237 align:middle line:90% This is trisomy 21. 00:04:16.237 --> 00:04:17.820 align:middle line:84% There's two other chromosomes that you 00:04:17.820 --> 00:04:20.310 align:middle line:84% can have an entire trisomy of and still survive 00:04:20.310 --> 00:04:21.519 align:middle line:90% one way or another. 00:04:21.519 --> 00:04:24.032 align:middle line:84% Now, we have 22 pairs of chromosomes here. 00:04:24.032 --> 00:04:24.990 align:middle line:90% AUDIENCE: Interruption. 00:04:24.990 --> 00:04:25.782 align:middle line:90% ATUL J. BUTTE: Yes. 00:04:25.782 --> 00:04:28.330 align:middle line:90% AUDIENCE: [INAUDIBLE] question. 00:04:28.330 --> 00:04:29.680 align:middle line:90% Explain the following fact. 00:04:29.680 --> 00:04:32.290 align:middle line:90% 00:04:32.290 --> 00:04:35.980 align:middle line:90% So this [INAUDIBLE]. 00:04:35.980 --> 00:04:38.590 align:middle line:84% How come with exception of testicular 00:04:38.590 --> 00:04:45.370 align:middle line:84% cancer Down syndrome people don't get cancer? 00:04:45.370 --> 00:04:46.120 align:middle line:90% ATUL J. BUTTE: OK. 00:04:46.120 --> 00:04:49.048 align:middle line:90% 00:04:49.048 --> 00:04:50.840 align:middle line:84% AUDIENCE: Look at the genes, figure it out. 00:04:50.840 --> 00:04:54.320 align:middle line:84% [INAUDIBLE] who don't get cancer. 00:04:54.320 --> 00:04:58.530 align:middle line:90% 00:04:58.530 --> 00:05:00.030 align:middle line:84% ATUL J. BUTTE: So like I was saying, 00:05:00.030 --> 00:05:03.260 align:middle line:84% we have 22 pairs of chromosomes that are ideally paired, 00:05:03.260 --> 00:05:05.720 align:middle line:84% although we can get trisomies rarely. 00:05:05.720 --> 00:05:07.880 align:middle line:84% We also have a pair of sex chromosomes. 00:05:07.880 --> 00:05:10.460 align:middle line:84% Either you have two Xs or you have an X and a Y. 00:05:10.460 --> 00:05:12.170 align:middle line:84% And those are the normal conditions. 00:05:12.170 --> 00:05:14.210 align:middle line:84% There are abnormal conditions as well, 00:05:14.210 --> 00:05:16.160 align:middle line:84% but these are the sex chromosomes here. 00:05:16.160 --> 00:05:19.520 align:middle line:84% Most of the Y pairs with a piece of the X, 00:05:19.520 --> 00:05:21.030 align:middle line:84% but the Y has some unique material. 00:05:21.030 --> 00:05:23.030 align:middle line:84% And the X certainly has a lot of unique material 00:05:23.030 --> 00:05:24.197 align:middle line:90% that doesn't pair at all. 00:05:24.197 --> 00:05:26.030 align:middle line:84% There's some very important things on the Y, 00:05:26.030 --> 00:05:28.095 align:middle line:84% for example, a piece of the androgen receptor 00:05:28.095 --> 00:05:28.970 align:middle line:90% and things like that. 00:05:28.970 --> 00:05:32.550 align:middle line:84% The coding for that is on the Y chromosome. 00:05:32.550 --> 00:05:34.740 align:middle line:84% So this is what an actual photograph looks like. 00:05:34.740 --> 00:05:37.850 align:middle line:84% And this is the idealized version of this ideogram. 00:05:37.850 --> 00:05:40.460 align:middle line:84% The bands essentially are actually 00:05:40.460 --> 00:05:42.182 align:middle line:84% quite specific with a particular stain. 00:05:42.182 --> 00:05:43.640 align:middle line:84% When a particular stain is applied, 00:05:43.640 --> 00:05:44.660 align:middle line:84% when you apply different stains, you 00:05:44.660 --> 00:05:45.920 align:middle line:90% get different banding patterns. 00:05:45.920 --> 00:05:47.462 align:middle line:84% And it turns out these bands actually 00:05:47.462 --> 00:05:50.030 align:middle line:84% have a lot to do with the base pair 00:05:50.030 --> 00:05:51.800 align:middle line:90% composition of those regions. 00:05:51.800 --> 00:05:54.680 align:middle line:84% For example, GC-rich, or regions that have a lot of Gs and Cs, 00:05:54.680 --> 00:05:58.718 align:middle line:84% will turn out to be one color versus the As and the Ts. 00:05:58.718 --> 00:06:01.010 align:middle line:84% So now, we've seen the big picture from the chromosome. 00:06:01.010 --> 00:06:02.843 align:middle line:84% We started with the DNA and the nucleotides. 00:06:02.843 --> 00:06:05.010 align:middle line:84% And we've already covered the double helices. 00:06:05.010 --> 00:06:07.090 align:middle line:84% So just as an introduction, I got only 10 slides 00:06:07.090 --> 00:06:08.840 align:middle line:84% of basic biology refresher, and then we'll 00:06:08.840 --> 00:06:10.560 align:middle line:90% get to some interesting stuff. 00:06:10.560 --> 00:06:13.640 align:middle line:84% So how do we get from the DNA to the genes, right? 00:06:13.640 --> 00:06:18.470 align:middle line:84% Because even though we have 3.5 billion base pairs, only 1.5% 00:06:18.470 --> 00:06:21.020 align:middle line:90% of that codes for the proteins. 00:06:21.020 --> 00:06:23.090 align:middle line:84% We said the proteins were the final goal. 00:06:23.090 --> 00:06:25.520 align:middle line:84% And out of the 3.5 billion base pairs, only 1 and 1/2% 00:06:25.520 --> 00:06:28.700 align:middle line:84% actually are in a coding region per se. 00:06:28.700 --> 00:06:32.210 align:middle line:84% So let's talk about what we call a coding region exactly. 00:06:32.210 --> 00:06:34.130 align:middle line:84% This line here, this red line here, 00:06:34.130 --> 00:06:36.500 align:middle line:90% represents the double helix. 00:06:36.500 --> 00:06:39.740 align:middle line:84% That represents the As, Cs, Ts, and Gs. 00:06:39.740 --> 00:06:43.860 align:middle line:84% And the structure of a gene is such that there's a piece here, 00:06:43.860 --> 00:06:46.362 align:middle line:84% which is going to serve as our blueprint for the protein 00:06:46.362 --> 00:06:47.570 align:middle line:90% called an open reading frame. 00:06:47.570 --> 00:06:49.940 align:middle line:90% It's got a start initiation. 00:06:49.940 --> 00:06:51.850 align:middle line:84% It's got an end or a termination. 00:06:51.850 --> 00:06:54.350 align:middle line:84% And this piece of that is going to be spliced in and spliced 00:06:54.350 --> 00:06:55.600 align:middle line:90% out as part of this blueprint. 00:06:55.600 --> 00:06:57.770 align:middle line:84% And we'll talk about the splicing in a second. 00:06:57.770 --> 00:07:01.880 align:middle line:84% The piece of the double helix that's upstream or before this, 00:07:01.880 --> 00:07:03.860 align:middle line:84% towards the 5 prime end, it's called upstream. 00:07:03.860 --> 00:07:05.235 align:middle line:84% And the piece of the double helix 00:07:05.235 --> 00:07:07.130 align:middle line:84% that's after this termination codon, 00:07:07.130 --> 00:07:09.770 align:middle line:84% it's called downstream when it's on 3 prime end. 00:07:09.770 --> 00:07:11.600 align:middle line:84% Now, here's the most crucial point. 00:07:11.600 --> 00:07:15.110 align:middle line:84% We all love diagramming these things like this, 00:07:15.110 --> 00:07:18.410 align:middle line:84% but there is no punctuation in the genome. 00:07:18.410 --> 00:07:19.880 align:middle line:90% There's no highlighter. 00:07:19.880 --> 00:07:21.680 align:middle line:90% There's no caps lowercase. 00:07:21.680 --> 00:07:23.630 align:middle line:90% There's no italics bold. 00:07:23.630 --> 00:07:26.690 align:middle line:84% All you have are the letters, right? 00:07:26.690 --> 00:07:28.370 align:middle line:84% So one of the hardest problems is 00:07:28.370 --> 00:07:31.670 align:middle line:84% trying to figure out this structure from the letters 00:07:31.670 --> 00:07:33.890 align:middle line:84% because there's no flashing arrow that says, here is 00:07:33.890 --> 00:07:35.150 align:middle line:90% where the gene begins. 00:07:35.150 --> 00:07:37.220 align:middle line:84% All you have are the As, Ts, Gs, and Cs. 00:07:37.220 --> 00:07:41.300 align:middle line:84% And it's only informatics that actually deciphers this, OK? 00:07:41.300 --> 00:07:44.930 align:middle line:84% So we have algorithms that people use commonly now that 00:07:44.930 --> 00:07:48.380 align:middle line:84% have been trained on manual experiments and the results 00:07:48.380 --> 00:07:50.500 align:middle line:84% of where this gene began and that gene began. 00:07:50.500 --> 00:07:51.500 align:middle line:90% They learn the patterns. 00:07:51.500 --> 00:07:53.500 align:middle line:84% They code it into algorithms, and the algorithms 00:07:53.500 --> 00:07:55.973 align:middle line:84% run on the entire genome to try to make predictions. 00:07:55.973 --> 00:07:57.890 align:middle line:84% But in the end, they're still only predictions 00:07:57.890 --> 00:08:01.880 align:middle line:84% of where these genes are, and they're still commonly wrong. 00:08:01.880 --> 00:08:04.190 align:middle line:84% So each gene, a gene, encodes instructions 00:08:04.190 --> 00:08:07.520 align:middle line:90% to make a single protein. 00:08:07.520 --> 00:08:09.800 align:middle line:84% The DNA is called, before, is upstream. 00:08:09.800 --> 00:08:13.460 align:middle line:84% And a lot of the regulatory elements for this gene 00:08:13.460 --> 00:08:15.390 align:middle line:84% are actually in the upstream component. 00:08:15.390 --> 00:08:18.890 align:middle line:84% So let's say more than they are in the downstream component. 00:08:18.890 --> 00:08:21.530 align:middle line:84% Regarding the splicing, imagine if an architect made 00:08:21.530 --> 00:08:24.637 align:middle line:84% a series of blueprints, but the architect said, 00:08:24.637 --> 00:08:26.720 align:middle line:84% well, if you don't want the Jacuzzi in your house, 00:08:26.720 --> 00:08:29.030 align:middle line:90% you can take this page out. 00:08:29.030 --> 00:08:31.280 align:middle line:84% If you only want a one car garage instead of a two car 00:08:31.280 --> 00:08:33.200 align:middle line:84% garage, you could take this page out. 00:08:33.200 --> 00:08:35.179 align:middle line:90% That's what the splicing is. 00:08:35.179 --> 00:08:37.730 align:middle line:84% You start with a whole set of what 00:08:37.730 --> 00:08:40.640 align:middle line:84% could be in the plans and pieces could be taken out 00:08:40.640 --> 00:08:42.409 align:middle line:90% or pieces could be left in. 00:08:42.409 --> 00:08:45.890 align:middle line:84% That's what the exons and introns can serve as. 00:08:45.890 --> 00:08:47.540 align:middle line:90% The introns get spliced out. 00:08:47.540 --> 00:08:48.680 align:middle line:90% They can be thrown out. 00:08:48.680 --> 00:08:50.883 align:middle line:84% And exons remain to actually be part 00:08:50.883 --> 00:08:52.550 align:middle line:84% of the blueprint, which we'll talk about 00:08:52.550 --> 00:08:54.170 align:middle line:90% in the next few slides. 00:08:54.170 --> 00:08:58.790 align:middle line:84% Different tissues may splice in and out different components 00:08:58.790 --> 00:08:59.460 align:middle line:90% here. 00:08:59.460 --> 00:09:00.700 align:middle line:90% It's not always the case. 00:09:00.700 --> 00:09:02.700 align:middle line:84% Even though I'm drawing it here as exon, intron, 00:09:02.700 --> 00:09:05.600 align:middle line:84% exon, in a different part of the same organism 00:09:05.600 --> 00:09:08.540 align:middle line:90% that intron might be left in. 00:09:08.540 --> 00:09:10.550 align:middle line:84% That's still a mystery as to exactly why 00:09:10.550 --> 00:09:13.790 align:middle line:84% that happens in some tissues, not in other tissues. 00:09:13.790 --> 00:09:15.788 align:middle line:84% And even the mechanism, the proteins 00:09:15.788 --> 00:09:18.080 align:middle line:84% that are involved in cutting and splicing these things, 00:09:18.080 --> 00:09:20.040 align:middle line:90% are reasonably well-defined. 00:09:20.040 --> 00:09:22.010 align:middle line:84% It's not clear why is it sometimes 00:09:22.010 --> 00:09:25.020 align:middle line:84% specific to one tissue versus another. 00:09:25.020 --> 00:09:27.720 align:middle line:84% These are called alternative spliced products. 00:09:27.720 --> 00:09:31.697 align:middle line:84% If I have the same open reading frame from start to end-- 00:09:31.697 --> 00:09:33.530 align:middle line:84% that's how we define an open reading frame-- 00:09:33.530 --> 00:09:35.540 align:middle line:84% if I have the exon intron exon in one blueprint, 00:09:35.540 --> 00:09:36.740 align:middle line:84% but just the two exons in the other, 00:09:36.740 --> 00:09:38.510 align:middle line:84% those are alternative splice products. 00:09:38.510 --> 00:09:40.130 align:middle line:84% And because they're two different blueprints, 00:09:40.130 --> 00:09:41.963 align:middle line:84% you're going get two different houses or two 00:09:41.963 --> 00:09:43.970 align:middle line:90% different proteins at the end. 00:09:43.970 --> 00:09:46.920 align:middle line:84% So we talked about the code at the start, a code at the end. 00:09:46.920 --> 00:09:49.310 align:middle line:84% And again, the biological system can figure out 00:09:49.310 --> 00:09:52.160 align:middle line:84% the promoter regions, where is the boundary between the intron 00:09:52.160 --> 00:09:54.230 align:middle line:84% and exon, where the start, where is the end using 00:09:54.230 --> 00:09:55.910 align:middle line:90% just the sequence syntax. 00:09:55.910 --> 00:09:56.960 align:middle line:90% There's no highlighter. 00:09:56.960 --> 00:10:00.620 align:middle line:84% There's no bold or uppercase, lowercase. 00:10:00.620 --> 00:10:03.210 align:middle line:90% 00:10:03.210 --> 00:10:07.030 align:middle line:84% So we have really 3.5 billion base pairs, 00:10:07.030 --> 00:10:09.180 align:middle line:90% but only 35,000 genes. 00:10:09.180 --> 00:10:10.680 align:middle line:90% And the reason why-- 00:10:10.680 --> 00:10:12.210 align:middle line:84% this was an early estimate of 3%. 00:10:12.210 --> 00:10:17.220 align:middle line:84% It's actually only 1.5% of the genome, so only around 60 or 70 00:10:17.220 --> 00:10:21.030 align:middle line:84% or 80 megabases or million bases out of the 3.5 billion base 00:10:21.030 --> 00:10:21.880 align:middle line:90% pairs. 00:10:21.880 --> 00:10:25.060 align:middle line:90% So what makes up the rest? 00:10:25.060 --> 00:10:27.060 align:middle line:84% If I just showed you the structure of the gene-- 00:10:27.060 --> 00:10:30.000 align:middle line:84% here is the gene that I just showed you, let's say. 00:10:30.000 --> 00:10:32.670 align:middle line:84% What makes up the space between genes? 00:10:32.670 --> 00:10:34.710 align:middle line:84% Well, we're not going to call it junk here, 00:10:34.710 --> 00:10:38.100 align:middle line:84% but the 50% of the regions between genes 00:10:38.100 --> 00:10:40.920 align:middle line:90% are repeated elements. 00:10:40.920 --> 00:10:44.190 align:middle line:84% So in other words, there are at least four different types 00:10:44.190 --> 00:10:45.300 align:middle line:90% of repeated elements. 00:10:45.300 --> 00:10:49.140 align:middle line:84% And those four are repeated many times in the genome. 00:10:49.140 --> 00:10:53.260 align:middle line:84% That makes up 50% of the genome itself. 00:10:53.260 --> 00:10:55.090 align:middle line:90% Now, these repeated elements-- 00:10:55.090 --> 00:10:57.180 align:middle line:84% so the repeated sequences are interesting. 00:10:57.180 --> 00:11:00.210 align:middle line:84% Because early on, when we had to actually-- when people were 00:11:00.210 --> 00:11:03.540 align:middle line:84% using DNA in a forensic type of way, the number of repeats 00:11:03.540 --> 00:11:05.830 align:middle line:84% can be different between individuals. 00:11:05.830 --> 00:11:07.030 align:middle line:90% So what do I mean by that? 00:11:07.030 --> 00:11:09.150 align:middle line:84% So here is one type of repeated element. 00:11:09.150 --> 00:11:11.560 align:middle line:84% This is a LINE, or Long Interspersed Nuclear Element. 00:11:11.560 --> 00:11:13.380 align:middle line:84% It doesn't matter what it is, but here you 00:11:13.380 --> 00:11:18.570 align:middle line:84% see 1, 2, 3, 4, 5, maybe 6, 7, 8 copies 00:11:18.570 --> 00:11:21.520 align:middle line:84% of this line between this gene and this gene. 00:11:21.520 --> 00:11:25.650 align:middle line:84% So you might have 998 repeats between these two parts, 00:11:25.650 --> 00:11:28.480 align:middle line:90% but I might have 997 of them. 00:11:28.480 --> 00:11:30.600 align:middle line:84% And if you do that between a whole bunch 00:11:30.600 --> 00:11:32.043 align:middle line:84% of different areas of the genome, 00:11:32.043 --> 00:11:33.210 align:middle line:90% you can actually measure it. 00:11:33.210 --> 00:11:35.220 align:middle line:84% You can actually find differences. 00:11:35.220 --> 00:11:37.740 align:middle line:84% You can do simple things, like figure out is a child 00:11:37.740 --> 00:11:40.200 align:middle line:84% a descendent of one or both parents. 00:11:40.200 --> 00:11:43.620 align:middle line:84% You can certainly tell, to some certain degree of confidence, 00:11:43.620 --> 00:11:46.020 align:middle line:84% whether a sample left at a crime scene 00:11:46.020 --> 00:11:48.940 align:middle line:84% is from this individual versus the random population. 00:11:48.940 --> 00:11:51.900 align:middle line:84% So just counting these repeats is actually pretty useful. 00:11:51.900 --> 00:11:54.690 align:middle line:84% Even though today we have much more accurate ways 00:11:54.690 --> 00:11:56.310 align:middle line:84% of figuring out whether a sample came 00:11:56.310 --> 00:11:57.990 align:middle line:84% from one person versus another, this 00:11:57.990 --> 00:12:00.570 align:middle line:84% is actually what was first used almost 20 years ago 00:12:00.570 --> 00:12:02.890 align:middle line:90% in the forensic view of DNA. 00:12:02.890 --> 00:12:04.690 align:middle line:84% So these are the repeated elements here. 00:12:04.690 --> 00:12:08.487 align:middle line:84% And the number of repeats can be different between individuals. 00:12:08.487 --> 00:12:09.195 align:middle line:90% Now, the repeat-- 00:12:09.195 --> 00:12:09.695 align:middle line:90% 00:12:09.695 --> 00:12:12.180 align:middle line:84% AUDIENCE: Another [INAUDIBLE] idea. 00:12:12.180 --> 00:12:16.780 align:middle line:84% Look at the number and placement of these repeats. 00:12:16.780 --> 00:12:18.480 align:middle line:84% And as you learn about [INAUDIBLE],, 00:12:18.480 --> 00:12:22.530 align:middle line:84% you can measure comprehensively especially these genomes. 00:12:22.530 --> 00:12:24.660 align:middle line:84% We have thousands of [INAUDIBLE] in the public. 00:12:24.660 --> 00:12:27.060 align:middle line:84% And we actually determine what is it 00:12:27.060 --> 00:12:31.260 align:middle line:84% about the spacing or number of these repeats that does or does 00:12:31.260 --> 00:12:33.690 align:middle line:90% not influence the expression. 00:12:33.690 --> 00:12:36.090 align:middle line:84% Do they have a purpose given that they're 00:12:36.090 --> 00:12:38.530 align:middle line:84% [? matured ?] quite widely throughout the species? 00:12:38.530 --> 00:12:40.572 align:middle line:84% ATUL J. BUTTE: So the repeats themselves are also 00:12:40.572 --> 00:12:42.120 align:middle line:90% interesting for one other thing. 00:12:42.120 --> 00:12:44.130 align:middle line:84% They repeat, so it makes you think that these 00:12:44.130 --> 00:12:45.750 align:middle line:90% are copies of an original. 00:12:45.750 --> 00:12:48.420 align:middle line:84% And in fact, when we look at the sequence of these repeated 00:12:48.420 --> 00:12:52.770 align:middle line:84% elements, we can tell what the original looked like 00:12:52.770 --> 00:12:56.370 align:middle line:84% and how deviant is this sequence from the original. 00:12:56.370 --> 00:12:57.670 align:middle line:90% So how can we tell that? 00:12:57.670 --> 00:13:03.720 align:middle line:84% So what a repeat exactly is a repeat codes for machinery 00:13:03.720 --> 00:13:05.760 align:middle line:90% to make a copy of itself. 00:13:05.760 --> 00:13:08.487 align:middle line:90% That's all a repeat is. 00:13:08.487 --> 00:13:09.570 align:middle line:90% So let me be really clear. 00:13:09.570 --> 00:13:14.160 align:middle line:84% The repeated elements making up 50% of the genome codes 00:13:14.160 --> 00:13:18.600 align:middle line:84% for the machinery that goes back and makes a copy of itself. 00:13:18.600 --> 00:13:21.515 align:middle line:84% That's why it's going to persist in the genome, right? 00:13:21.515 --> 00:13:23.640 align:middle line:84% If all you need is one or a couple of these things, 00:13:23.640 --> 00:13:25.557 align:middle line:84% eventually they're going to come back and make 00:13:25.557 --> 00:13:27.810 align:middle line:84% copies of themselves as the genome goes on 00:13:27.810 --> 00:13:29.730 align:middle line:90% through evolution. 00:13:29.730 --> 00:13:33.600 align:middle line:84% Now, it turns out all of these repeated elements 00:13:33.600 --> 00:13:35.790 align:middle line:90% are broken today in humans. 00:13:35.790 --> 00:13:38.010 align:middle line:84% None of them work today in humans. 00:13:38.010 --> 00:13:39.540 align:middle line:84% That's not true in other organisms. 00:13:39.540 --> 00:13:42.630 align:middle line:84% Repeats are still alive and well in the mouse, 00:13:42.630 --> 00:13:43.900 align:middle line:90% but not in humans. 00:13:43.900 --> 00:13:46.780 align:middle line:84% In fact, they're all dead in humans. 00:13:46.780 --> 00:13:47.305 align:middle line:90% Why? 00:13:47.305 --> 00:13:48.930 align:middle line:84% AUDIENCE: You mean they don't work in-- 00:13:48.930 --> 00:13:49.770 align:middle line:90% ATUL J. BUTTE: They can't work. 00:13:49.770 --> 00:13:50.812 align:middle line:90% None of them are working. 00:13:50.812 --> 00:13:53.110 align:middle line:84% They cannot go back and make copies anymore. 00:13:53.110 --> 00:13:54.808 align:middle line:84% None of them code for that anymore. 00:13:54.808 --> 00:13:55.350 align:middle line:90% AUDIENCE: OK. 00:13:55.350 --> 00:13:56.490 align:middle line:84% Generally, evolutionary what we see 00:13:56.490 --> 00:13:57.910 align:middle line:84% is the first of these classes are 00:13:57.910 --> 00:14:02.752 align:middle line:84% repeats that populate a certain branch in a tree. 00:14:02.752 --> 00:14:04.710 align:middle line:84% ATUL J. BUTTE: Absolutely, I have a whole thing 00:14:04.710 --> 00:14:05.970 align:middle line:90% on this if we get to it, OK? 00:14:05.970 --> 00:14:06.390 align:middle line:90% AUDIENCE: Oh, all right. 00:14:06.390 --> 00:14:07.682 align:middle line:90% ATUL J. BUTTE: If we get to it. 00:14:07.682 --> 00:14:09.352 align:middle line:90% 00:14:09.352 --> 00:14:10.560 align:middle line:90% So just think about that, OK? 00:14:10.560 --> 00:14:12.960 align:middle line:84% So these different types of nuclear-- these different 00:14:12.960 --> 00:14:14.925 align:middle line:84% repeated sequences are all different ways 00:14:14.925 --> 00:14:16.050 align:middle line:90% to code for that machinery. 00:14:16.050 --> 00:14:19.170 align:middle line:90% That's all it is. 00:14:19.170 --> 00:14:21.300 align:middle line:84% We're the smartest organisms on this planet 00:14:21.300 --> 00:14:24.330 align:middle line:84% or so we like to think, so we must have the largest genome, 00:14:24.330 --> 00:14:25.620 align:middle line:90% right? 00:14:25.620 --> 00:14:26.550 align:middle line:90% Absolutely wrong. 00:14:26.550 --> 00:14:29.460 align:middle line:90% So we have 3.5 million bases. 00:14:29.460 --> 00:14:33.390 align:middle line:84% So it turns out with 3.5 million bases and only four letters 00:14:33.390 --> 00:14:39.210 align:middle line:84% A, G, T, C, you can fit your entire genome on a CD-ROM. 00:14:39.210 --> 00:14:41.790 align:middle line:84% You don't even need a DVD for this. 00:14:41.790 --> 00:14:44.700 align:middle line:84% You can easily fit-- with no additional compression, 00:14:44.700 --> 00:14:48.000 align:middle line:84% you can fit your entire genome in 750 megabytes, which is 00:14:48.000 --> 00:14:50.550 align:middle line:90% what a typical CD-ROM can hold. 00:14:50.550 --> 00:14:52.870 align:middle line:84% So just think about that for a second. 00:14:52.870 --> 00:14:55.590 align:middle line:84% So look at these other organisms and what their genome sizes 00:14:55.590 --> 00:14:56.508 align:middle line:90% are, the E. coli. 00:14:56.508 --> 00:14:58.800 align:middle line:84% So E. coli is a common bacteria that lives in your gut. 00:14:58.800 --> 00:14:59.592 align:middle line:90% It can be friendly. 00:14:59.592 --> 00:15:02.870 align:middle line:84% It can also be pathogenic-- has 4 million basis. 00:15:02.870 --> 00:15:05.780 align:middle line:84% Yeast, which is used to make bread-- 00:15:05.780 --> 00:15:07.130 align:middle line:90% 12 million bases. 00:15:07.130 --> 00:15:10.930 align:middle line:84% But even the pea, the garden pea-- 00:15:10.930 --> 00:15:13.310 align:middle line:84% you got peas in your salad maybe for lunch-- 00:15:13.310 --> 00:15:17.510 align:middle line:84% has more of a genome than we do, 4.8 billion. 00:15:17.510 --> 00:15:21.890 align:middle line:84% Maize and wheat-- wheat has 17 billion base pairs. 00:15:21.890 --> 00:15:25.898 align:middle line:84% And we have 3 billion base pairs. 00:15:25.898 --> 00:15:27.440 align:middle line:84% So the size of the genome has nothing 00:15:27.440 --> 00:15:29.232 align:middle line:84% to do with the intelligence of the organism 00:15:29.232 --> 00:15:30.500 align:middle line:90% at all it turns out. 00:15:30.500 --> 00:15:33.170 align:middle line:84% So where are these genomes different? 00:15:33.170 --> 00:15:36.420 align:middle line:84% A lot of the difference is in the space between the genes. 00:15:36.420 --> 00:15:38.120 align:middle line:90% So here are four organisms. 00:15:38.120 --> 00:15:41.060 align:middle line:84% Here's human, the same 50,000 base pairs I just showed you 00:15:41.060 --> 00:15:42.330 align:middle line:90% on the previous slide. 00:15:42.330 --> 00:15:44.720 align:middle line:84% And here are four genes within 50,000 base pairs 00:15:44.720 --> 00:15:46.620 align:middle line:84% and a whole bunch of repeats in the middle. 00:15:46.620 --> 00:15:47.960 align:middle line:90% Here's yeast. 00:15:47.960 --> 00:15:51.200 align:middle line:84% And you can see how many genes there are and very few spaces. 00:15:51.200 --> 00:15:53.570 align:middle line:90% So it's much more compact. 00:15:53.570 --> 00:15:54.470 align:middle line:90% Here's maize. 00:15:54.470 --> 00:15:55.890 align:middle line:84% And wheat would be the same thing. 00:15:55.890 --> 00:15:58.970 align:middle line:84% And 50,000 base pairs, there's only one gene there. 00:15:58.970 --> 00:16:00.260 align:middle line:90% And here's E. coli. 00:16:00.260 --> 00:16:03.800 align:middle line:84% In 50,000 bases, not only is it so compact, 00:16:03.800 --> 00:16:05.930 align:middle line:84% but there's actually genes on both sides 00:16:05.930 --> 00:16:07.760 align:middle line:90% of the genome, both strands. 00:16:07.760 --> 00:16:10.820 align:middle line:84% And they can even overlap it turns out. 00:16:10.820 --> 00:16:13.370 align:middle line:84% Then we talk about that more specifically here. 00:16:13.370 --> 00:16:16.490 align:middle line:84% This is a genome for plasmodium falciparum, which 00:16:16.490 --> 00:16:19.130 align:middle line:90% is the organism that causes-- 00:16:19.130 --> 00:16:20.030 align:middle line:90% AUDIENCE: Malaria? 00:16:20.030 --> 00:16:22.160 align:middle line:84% ATUL J. BUTTE: --malaria, still the number one 00:16:22.160 --> 00:16:24.710 align:middle line:84% infectious disease killer in the world. 00:16:24.710 --> 00:16:25.790 align:middle line:90% This is chromosome 2. 00:16:25.790 --> 00:16:28.130 align:middle line:90% And you see the arrows here. 00:16:28.130 --> 00:16:30.020 align:middle line:84% The arrows represent which strand 00:16:30.020 --> 00:16:31.760 align:middle line:90% of the DNA the gene is on. 00:16:31.760 --> 00:16:34.670 align:middle line:84% They don't all have to be on the same strand, 00:16:34.670 --> 00:16:36.170 align:middle line:90% making the problem even harder. 00:16:36.170 --> 00:16:38.030 align:middle line:90% And they can overlap. 00:16:38.030 --> 00:16:41.660 align:middle line:84% There's many instances of a gene where another gene exactly 00:16:41.660 --> 00:16:42.890 align:middle line:90% overlaps. 00:16:42.890 --> 00:16:47.700 align:middle line:84% It makes a hard to measure one versus the other it turns out. 00:16:47.700 --> 00:16:50.127 align:middle line:90% So this is the central dogma. 00:16:50.127 --> 00:16:51.960 align:middle line:84% We'll talk about central dogmas in a second, 00:16:51.960 --> 00:16:54.180 align:middle line:84% but this is basically the meat of how this happens. 00:16:54.180 --> 00:16:55.222 align:middle line:90% It serves as a blueprint. 00:16:55.222 --> 00:16:57.270 align:middle line:90% As we all know, the DNA-- 00:16:57.270 --> 00:17:01.590 align:middle line:84% when the cell decides to make this protein, 00:17:01.590 --> 00:17:04.349 align:middle line:84% somehow the DNA unravels in such a way 00:17:04.349 --> 00:17:06.869 align:middle line:84% to expose just the smallest piece 00:17:06.869 --> 00:17:10.050 align:middle line:84% of the 1.5% of the genome that's going to code for the protein 00:17:10.050 --> 00:17:11.369 align:middle line:90% it wants to make. 00:17:11.369 --> 00:17:14.293 align:middle line:84% And so the DNA is kept locked up and nice and safe 00:17:14.293 --> 00:17:15.960 align:middle line:84% in the nucleus, the center of the cells. 00:17:15.960 --> 00:17:18.690 align:middle line:84% And the proteins have to be made in exterior of the cells. 00:17:18.690 --> 00:17:22.415 align:middle line:84% So a temporary copy is made of the DNA that goes out. 00:17:22.415 --> 00:17:24.540 align:middle line:84% And it's going to be worked on to make the protein. 00:17:24.540 --> 00:17:27.000 align:middle line:84% And the temporary copy is called RNA or, specifically, 00:17:27.000 --> 00:17:29.220 align:middle line:90% messenger RNA. 00:17:29.220 --> 00:17:31.830 align:middle line:84% And so when you unravel, a messenger RNA 00:17:31.830 --> 00:17:35.370 align:middle line:84% is actually made in an opposite sequence, 00:17:35.370 --> 00:17:39.330 align:middle line:84% or this is actually built in the opposite complementarity 00:17:39.330 --> 00:17:40.150 align:middle line:90% of the DNA. 00:17:40.150 --> 00:17:42.720 align:middle line:84% So if there's an A there, there's a T here, et cetera. 00:17:42.720 --> 00:17:45.630 align:middle line:84% That gets exported out actively out of the nucleus 00:17:45.630 --> 00:17:50.550 align:middle line:84% and goes into the cytoplasm where it's made into a protein. 00:17:50.550 --> 00:17:52.530 align:middle line:84% And to me, this is the most interesting part 00:17:52.530 --> 00:17:54.120 align:middle line:90% of the whole thing. 00:17:54.120 --> 00:17:58.410 align:middle line:84% Somehow nature realized that you can't just make proteins 00:17:58.410 --> 00:17:59.640 align:middle line:90% with the four letters. 00:17:59.640 --> 00:18:01.950 align:middle line:90% You have As, Ts, Cs, and Gs. 00:18:01.950 --> 00:18:05.790 align:middle line:84% And proteins are made up of 21 amino acids-- 00:18:05.790 --> 00:18:08.400 align:middle line:90% not 20, 21 amino acids. 00:18:08.400 --> 00:18:11.190 align:middle line:84% The 21st is actually one that was just discovered about 10 00:18:11.190 --> 00:18:13.220 align:middle line:90% years or so ago it turns out. 00:18:13.220 --> 00:18:15.640 align:middle line:84% And the coding for that one breaks all the rules. 00:18:15.640 --> 00:18:17.890 align:middle line:84% And we could talk about that if people are interested. 00:18:17.890 --> 00:18:20.640 align:middle line:84% But there's 20 amino acids, let's say. 00:18:20.640 --> 00:18:21.690 align:middle line:90% And I have four letters. 00:18:21.690 --> 00:18:24.672 align:middle line:84% Obviously, if I made one amino acid for one letter, 00:18:24.672 --> 00:18:25.380 align:middle line:90% I couldn't do it. 00:18:25.380 --> 00:18:28.398 align:middle line:84% If I had two letters, I got four things in this one and four 00:18:28.398 --> 00:18:29.190 align:middle line:90% things in this one. 00:18:29.190 --> 00:18:30.960 align:middle line:84% That would code for 16 amino acids, 00:18:30.960 --> 00:18:33.580 align:middle line:90% but I have 20 to choose from. 00:18:33.580 --> 00:18:37.180 align:middle line:84% So I need three letters to code for at least 20 amino acids. 00:18:37.180 --> 00:18:42.810 align:middle line:84% So that's what's known as the genetic code. 00:18:42.810 --> 00:18:45.203 align:middle line:90% I'm blanking on the term. 00:18:45.203 --> 00:18:47.370 align:middle line:84% So each position can be one of the four nucleotides. 00:18:47.370 --> 00:18:50.040 align:middle line:84% And nature evolved into using three nucleotides to code 00:18:50.040 --> 00:18:51.460 align:middle line:90% for a single amino acid. 00:18:51.460 --> 00:18:53.820 align:middle line:84% So that messenger RNA comes in here. 00:18:53.820 --> 00:18:55.920 align:middle line:84% And one of the most amazing pieces 00:18:55.920 --> 00:19:00.390 align:middle line:84% of nanotechnology, this ribosome, looks at the strand 00:19:00.390 --> 00:19:01.800 align:middle line:90% as it's coming in. 00:19:01.800 --> 00:19:05.310 align:middle line:84% And it's able to link that strand with other amino acids 00:19:05.310 --> 00:19:09.480 align:middle line:84% that are held in place by these transfer RNAs that recognize 00:19:09.480 --> 00:19:11.260 align:middle line:90% these triplet sequences. 00:19:11.260 --> 00:19:14.130 align:middle line:84% So if there's a particular A, A, T, 00:19:14.130 --> 00:19:17.610 align:middle line:84% the tRNA that recognizes A, A, T, or rather 00:19:17.610 --> 00:19:20.550 align:middle line:84% the opposite of that, which is holding the right amino acid, 00:19:20.550 --> 00:19:21.630 align:middle line:90% comes into place. 00:19:21.630 --> 00:19:27.180 align:middle line:84% And the ribosome attaches it to build a growing protein chain. 00:19:27.180 --> 00:19:29.970 align:middle line:84% So in comes the RNA, out comes the protein. 00:19:29.970 --> 00:19:32.580 align:middle line:84% The micro machinery for this is astounding 00:19:32.580 --> 00:19:34.090 align:middle line:90% when you think about it. 00:19:34.090 --> 00:19:38.160 align:middle line:84% Remember, this is happening many, many times per second 00:19:38.160 --> 00:19:39.090 align:middle line:90% in all of your cells. 00:19:39.090 --> 00:19:40.480 align:middle line:84% You don't have to think about it. 00:19:40.480 --> 00:19:41.850 align:middle line:90% It hardly ever fails. 00:19:41.850 --> 00:19:45.220 align:middle line:84% If it failed, we just couldn't do anything. 00:19:45.220 --> 00:19:46.630 align:middle line:90% So here is this genetic code. 00:19:46.630 --> 00:19:49.060 align:middle line:84% So there's 64 different positions here. 00:19:49.060 --> 00:19:51.480 align:middle line:84% And so you see something like UUU. 00:19:51.480 --> 00:19:54.120 align:middle line:84% So Us are used in the RNA instead of Ts. 00:19:54.120 --> 00:19:55.620 align:middle line:84% It's just a slight difference there. 00:19:55.620 --> 00:19:57.780 align:middle line:90% UUU is code for phenylalanine. 00:19:57.780 --> 00:19:59.520 align:middle line:90% UCC codes for serine. 00:19:59.520 --> 00:20:00.930 align:middle line:84% And that's the code, essentially. 00:20:00.930 --> 00:20:03.870 align:middle line:84% Now, obviously, these four different codes 00:20:03.870 --> 00:20:05.160 align:middle line:90% code for serine. 00:20:05.160 --> 00:20:08.472 align:middle line:84% It doesn't actually matter what the third base pair is. 00:20:08.472 --> 00:20:10.180 align:middle line:84% It's the first two are what are specific. 00:20:10.180 --> 00:20:11.680 align:middle line:84% So there's degeneracies in this code 00:20:11.680 --> 00:20:13.620 align:middle line:84% because we don't have to choose from 64. 00:20:13.620 --> 00:20:18.210 align:middle line:84% We only have 20 or 21 amino acids to choose from. 00:20:18.210 --> 00:20:20.293 align:middle line:84% So this is actually the central dogma here. 00:20:20.293 --> 00:20:21.960 align:middle line:84% And these are all the different concepts 00:20:21.960 --> 00:20:23.160 align:middle line:90% we've talked about so far. 00:20:23.160 --> 00:20:24.630 align:middle line:84% We talked about nucleotides, which 00:20:24.630 --> 00:20:26.700 align:middle line:90% are held in a double helix. 00:20:26.700 --> 00:20:29.385 align:middle line:84% One single double helix makes up a chromosome. 00:20:29.385 --> 00:20:31.950 align:middle line:84% A chromosome holds the genes or DNA. 00:20:31.950 --> 00:20:35.370 align:middle line:84% And the entire set of all these genes is called the genome. 00:20:35.370 --> 00:20:37.920 align:middle line:84% The genes code for messenger RNA. 00:20:37.920 --> 00:20:40.680 align:middle line:84% The transfer RNAs bring the amino acids, 00:20:40.680 --> 00:20:43.230 align:middle line:84% and the ribosome operates on that to make the proteins. 00:20:43.230 --> 00:20:45.120 align:middle line:84% We'll talk about signal sequence in a second. 00:20:45.120 --> 00:20:47.340 align:middle line:84% The amino acids are joined together to make proteins. 00:20:47.340 --> 00:20:50.910 align:middle line:84% And this sequence of arrows here is called a central dogma. 00:20:50.910 --> 00:20:54.630 align:middle line:84% Even though the arrows work bidirectionally now, 00:20:54.630 --> 00:20:58.290 align:middle line:84% this is the original central dogma. 00:20:58.290 --> 00:21:00.840 align:middle line:84% Now, another fascinating part is this. 00:21:00.840 --> 00:21:04.410 align:middle line:84% Sometimes the cell needs to make something 00:21:04.410 --> 00:21:08.530 align:middle line:84% that would be quite detrimental to the cell. 00:21:08.530 --> 00:21:10.950 align:middle line:84% So for example, there are cells in your pancreas 00:21:10.950 --> 00:21:15.390 align:middle line:84% that need to make things to digest the food we eat. 00:21:15.390 --> 00:21:16.890 align:middle line:90% I just went to the truck. 00:21:16.890 --> 00:21:18.810 align:middle line:90% I had a little dumpling. 00:21:18.810 --> 00:21:21.930 align:middle line:84% Now, I need to break down that protein. 00:21:21.930 --> 00:21:23.790 align:middle line:84% And it turns out I can break down protein 00:21:23.790 --> 00:21:25.840 align:middle line:90% with something called trypsin. 00:21:25.840 --> 00:21:27.840 align:middle line:84% Trypsin will help me break down proteins as well 00:21:27.840 --> 00:21:32.940 align:middle line:84% as many other things, but those things are proteins themselves. 00:21:32.940 --> 00:21:35.460 align:middle line:90% So here's the quandary here. 00:21:35.460 --> 00:21:37.080 align:middle line:84% My pancreatic cells, how are they 00:21:37.080 --> 00:21:41.250 align:middle line:84% going to make this trypsin without digesting itself? 00:21:41.250 --> 00:21:44.010 align:middle line:84% How do I make proteins that are going to kill me 00:21:44.010 --> 00:21:45.600 align:middle line:90% if I need to make them? 00:21:45.600 --> 00:21:49.080 align:middle line:84% Well, so what happens is the first few base 00:21:49.080 --> 00:21:52.590 align:middle line:84% pairs of the sequence, in fact, the first few amino acids that 00:21:52.590 --> 00:21:55.050 align:middle line:84% get spit out by the ribosome, might 00:21:55.050 --> 00:21:57.360 align:middle line:84% code for what's called a signal peptide. 00:21:57.360 --> 00:22:01.020 align:middle line:84% And the ribosome might see this and say, hold on. 00:22:01.020 --> 00:22:04.600 align:middle line:84% The thing I'm about to make could kill me. 00:22:04.600 --> 00:22:06.610 align:middle line:90% I need to stop here. 00:22:06.610 --> 00:22:08.640 align:middle line:84% And what happens is somehow the ribosome 00:22:08.640 --> 00:22:11.160 align:middle line:84% makes its way to a safer place to make 00:22:11.160 --> 00:22:13.770 align:middle line:84% the proteins called the Endoplasmic Reticulum, 00:22:13.770 --> 00:22:15.180 align:middle line:90% or the ER. 00:22:15.180 --> 00:22:16.920 align:middle line:90% So this is the ribosome. 00:22:16.920 --> 00:22:18.250 align:middle line:90% Here's the RNA coming in. 00:22:18.250 --> 00:22:20.063 align:middle line:90% Here's a growing protein chain. 00:22:20.063 --> 00:22:21.480 align:middle line:84% And all of a sudden this red piece 00:22:21.480 --> 00:22:24.688 align:middle line:84% here gets recognized by the accessory part of the ribosome, 00:22:24.688 --> 00:22:25.980 align:middle line:90% the signal recognition peptide. 00:22:25.980 --> 00:22:26.940 align:middle line:90% It says, hold on. 00:22:26.940 --> 00:22:28.528 align:middle line:90% Don't keep going here. 00:22:28.528 --> 00:22:30.570 align:middle line:84% Because if I even have just one molecule of this, 00:22:30.570 --> 00:22:33.442 align:middle line:84% I'm going to start degrading the proteins within the cell. 00:22:33.442 --> 00:22:34.650 align:middle line:90% So the whole thing goes over. 00:22:34.650 --> 00:22:38.370 align:middle line:84% It starts to spit the protein chain into this hole 00:22:38.370 --> 00:22:41.820 align:middle line:84% into the endoplasmic reticulum where that can actually then 00:22:41.820 --> 00:22:43.595 align:middle line:90% be built safely. 00:22:43.595 --> 00:22:44.970 align:middle line:84% This is in a separate compartment 00:22:44.970 --> 00:22:46.710 align:middle line:90% that can be exported. 00:22:46.710 --> 00:22:48.600 align:middle line:84% It's a pretty sophisticated piece 00:22:48.600 --> 00:22:50.220 align:middle line:90% of nanotechnology, this thing. 00:22:50.220 --> 00:22:53.040 align:middle line:84% It knows, because it's given the right codes, to not 00:22:53.040 --> 00:22:55.380 align:middle line:84% digest itself because it might need to make something 00:22:55.380 --> 00:22:57.660 align:middle line:90% quite toxic to the cell. 00:22:57.660 --> 00:22:59.370 align:middle line:84% There are signal recognition peptides 00:22:59.370 --> 00:23:01.650 align:middle line:84% like this likely for a lot of different compartments 00:23:01.650 --> 00:23:03.210 align:middle line:90% in the cell. 00:23:03.210 --> 00:23:04.680 align:middle line:90% Don't build this into the ER. 00:23:04.680 --> 00:23:06.510 align:middle line:90% Build it somewhere else. 00:23:06.510 --> 00:23:09.030 align:middle line:84% The complete catalog of those is still unknown, 00:23:09.030 --> 00:23:10.770 align:middle line:84% but would be quite valuable to know. 00:23:10.770 --> 00:23:13.390 align:middle line:90% 00:23:13.390 --> 00:23:18.340 align:middle line:84% So how did the cell decide to make this gene? 00:23:18.340 --> 00:23:21.340 align:middle line:84% This area of research is called transcriptional regulation. 00:23:21.340 --> 00:23:24.640 align:middle line:84% So why does this cell decide to make that messenger RNA? 00:23:24.640 --> 00:23:27.140 align:middle line:84% No one knows for sure in any one case 00:23:27.140 --> 00:23:29.800 align:middle line:84% how we actually get the DNA to unfold 00:23:29.800 --> 00:23:32.830 align:middle line:84% in just the right way for these things to actually be built, 00:23:32.830 --> 00:23:34.150 align:middle line:90% but we love to theorize. 00:23:34.150 --> 00:23:36.500 align:middle line:84% And we love to draw pictures like this. 00:23:36.500 --> 00:23:38.980 align:middle line:84% And you'll even see pictures like this in The New England 00:23:38.980 --> 00:23:40.180 align:middle line:90% Journal of Medicine now. 00:23:40.180 --> 00:23:42.860 align:middle line:84% This is a commonly accepted way to draw a gene. 00:23:42.860 --> 00:23:44.720 align:middle line:90% So here's the start of a gene. 00:23:44.720 --> 00:23:47.860 align:middle line:84% The arrow here means that this is where the gene, a copy, 00:23:47.860 --> 00:23:49.240 align:middle line:90% is going to be taken off. 00:23:49.240 --> 00:23:50.830 align:middle line:90% And this is the upstream region. 00:23:50.830 --> 00:23:52.690 align:middle line:84% And these boxes represent things that 00:23:52.690 --> 00:23:56.230 align:middle line:84% can bind to that upstream region to turn on or turn off 00:23:56.230 --> 00:23:57.560 align:middle line:90% this whole process. 00:23:57.560 --> 00:24:00.620 align:middle line:84% So the GR here represents the Glucocorticoid Receptor. 00:24:00.620 --> 00:24:03.625 align:middle line:84% So when this cell sees steroids, the steroids 00:24:03.625 --> 00:24:05.500 align:middle line:84% are going to bind to glucocorticoid receptor, 00:24:05.500 --> 00:24:08.200 align:middle line:84% and the glucocorticoid receptor is going to bind here. 00:24:08.200 --> 00:24:10.390 align:middle line:84% And this picture here shows that with the hormone, 00:24:10.390 --> 00:24:13.210 align:middle line:84% with steroids, this gets turned off. 00:24:13.210 --> 00:24:15.520 align:middle line:90% They put a little X here. 00:24:15.520 --> 00:24:17.560 align:middle line:84% And here, this is saying without, 00:24:17.560 --> 00:24:21.370 align:middle line:84% so with no steroids present, these two are actually 00:24:21.370 --> 00:24:22.540 align:middle line:90% bound in the natural state. 00:24:22.540 --> 00:24:24.790 align:middle line:84% But when you have that plus a glucocorticoid receptor, 00:24:24.790 --> 00:24:25.840 align:middle line:90% the thing gets turned on. 00:24:25.840 --> 00:24:29.170 align:middle line:84% So you can imagine things, in fact other proteins-- 00:24:29.170 --> 00:24:32.290 align:middle line:84% the glucocorticoid receptor is another protein-- 00:24:32.290 --> 00:24:36.130 align:middle line:84% come back to the DNA and bind to the regions 00:24:36.130 --> 00:24:39.800 align:middle line:84% upstream of the genes to turn on and turn off other genes. 00:24:39.800 --> 00:24:43.690 align:middle line:84% So the whole thing is one amazing repetitive cascade. 00:24:43.690 --> 00:24:46.150 align:middle line:84% Genes code for proteins, some of which 00:24:46.150 --> 00:24:48.880 align:middle line:84% do their job, some of which come back to the DNA 00:24:48.880 --> 00:24:51.400 align:middle line:84% and turn on other genes or turn off other genes. 00:24:51.400 --> 00:24:54.140 align:middle line:84% And this cycle goes on endlessly. 00:24:54.140 --> 00:24:59.470 align:middle line:84% And we just know the smallest fraction of these cycles. 00:24:59.470 --> 00:25:02.620 align:middle line:84% Remember, we're drawing nice pretty boxes like this here. 00:25:02.620 --> 00:25:05.230 align:middle line:84% There is no boldface, no highlighter, 00:25:05.230 --> 00:25:07.300 align:middle line:84% nothing upstream that tells us this is 00:25:07.300 --> 00:25:08.650 align:middle line:90% where it's going to bind. 00:25:08.650 --> 00:25:10.720 align:middle line:90% All we have are the letters. 00:25:10.720 --> 00:25:13.570 align:middle line:84% And a lot of slow motion biology goes into figuring out 00:25:13.570 --> 00:25:15.257 align:middle line:90% where these binding sites are. 00:25:15.257 --> 00:25:17.590 align:middle line:84% And we love to think that we have algorithms to do this. 00:25:17.590 --> 00:25:19.900 align:middle line:90% Some work and some don't. 00:25:19.900 --> 00:25:22.150 align:middle line:84% Huge area of research here if we could figure out 00:25:22.150 --> 00:25:25.450 align:middle line:84% this immense network of transcription. 00:25:25.450 --> 00:25:27.550 align:middle line:84% What can start the process even though these 00:25:27.550 --> 00:25:29.320 align:middle line:90% are going on endlessly? 00:25:29.320 --> 00:25:32.080 align:middle line:84% Whole bunches, whole different actions, on cells 00:25:32.080 --> 00:25:34.730 align:middle line:84% can actually trigger specific responses. 00:25:34.730 --> 00:25:37.300 align:middle line:84% So for example, hormone action on receptors-- 00:25:37.300 --> 00:25:38.500 align:middle line:90% I just had lunch. 00:25:38.500 --> 00:25:40.630 align:middle line:84% My pancreas is now making insulin. 00:25:40.630 --> 00:25:42.370 align:middle line:90% That insulin is a small protein. 00:25:42.370 --> 00:25:44.710 align:middle line:84% It's getting exported out of my pancreatic cells 00:25:44.710 --> 00:25:46.150 align:middle line:90% into my bloodstream. 00:25:46.150 --> 00:25:48.970 align:middle line:84% That insulin is going to now go to the muscle and fat 00:25:48.970 --> 00:25:51.280 align:middle line:84% because my stomach is taking in all this sugar 00:25:51.280 --> 00:25:53.650 align:middle line:90% from what I just ate. 00:25:53.650 --> 00:25:56.120 align:middle line:84% I need to store that sugar somewhere. 00:25:56.120 --> 00:25:58.840 align:middle line:84% So the insulin is going from the pancreas into the blood 00:25:58.840 --> 00:26:02.410 align:middle line:84% and triggering a response as I speak in my liver 00:26:02.410 --> 00:26:03.920 align:middle line:90% fat and muscle cells. 00:26:03.920 --> 00:26:07.210 align:middle line:84% So there might be a receptor up here, another protein. 00:26:07.210 --> 00:26:09.460 align:middle line:84% The insulin comes and binds and triggers a whole bunch 00:26:09.460 --> 00:26:11.800 align:middle line:84% of things to turn on and turn off different genes being 00:26:11.800 --> 00:26:13.450 align:middle line:90% expressed. 00:26:13.450 --> 00:26:15.250 align:middle line:90% Shock or stress to the cell-- 00:26:15.250 --> 00:26:17.230 align:middle line:84% I just walked over here from Children's. 00:26:17.230 --> 00:26:20.050 align:middle line:90% It is frickin' cold out there. 00:26:20.050 --> 00:26:21.550 align:middle line:84% A whole bunch of things just started 00:26:21.550 --> 00:26:25.750 align:middle line:84% to be transcribed in my genome, new source of or lack 00:26:25.750 --> 00:26:26.510 align:middle line:90% of nutrients. 00:26:26.510 --> 00:26:29.270 align:middle line:84% If you have yeast and you put in one type of sugar, 00:26:29.270 --> 00:26:31.270 align:middle line:84% all of a sudden it starts to make things to deal 00:26:31.270 --> 00:26:32.270 align:middle line:90% with that type of sugar. 00:26:32.270 --> 00:26:34.300 align:middle line:84% If you change it to a different source of sugar, 00:26:34.300 --> 00:26:35.410 align:middle line:90% those genes go down. 00:26:35.410 --> 00:26:36.710 align:middle line:90% Another set of genes go up. 00:26:36.710 --> 00:26:38.085 align:middle line:84% So that could start a whole bunch 00:26:38.085 --> 00:26:39.640 align:middle line:90% of genes being transcribed. 00:26:39.640 --> 00:26:42.887 align:middle line:84% Internal derangements of a cell, this 00:26:42.887 --> 00:26:44.470 align:middle line:84% is actually two different ways to make 00:26:44.470 --> 00:26:45.850 align:middle line:90% a transcriptional program. 00:26:45.850 --> 00:26:47.938 align:middle line:84% If I fly from here to San Francisco, 00:26:47.938 --> 00:26:49.480 align:middle line:84% I'm going to be hit with cosmic rays. 00:26:49.480 --> 00:26:50.855 align:middle line:84% And more than likely, one of them 00:26:50.855 --> 00:26:52.378 align:middle line:90% is going to hit one of my cells. 00:26:52.378 --> 00:26:54.920 align:middle line:84% And all of a sudden, if it hits something in the wrong place, 00:26:54.920 --> 00:26:57.970 align:middle line:84% I might start to make a gene that I shouldn't be making. 00:26:57.970 --> 00:27:00.207 align:middle line:84% But even better, if my cell detects 00:27:00.207 --> 00:27:02.290 align:middle line:84% that it's making something it shouldn't be making, 00:27:02.290 --> 00:27:05.620 align:middle line:84% it might just decide to kill itself, save me 00:27:05.620 --> 00:27:07.970 align:middle line:90% the problem of having cancer. 00:27:07.970 --> 00:27:10.360 align:middle line:84% So those are two different ways internal derangements can 00:27:10.360 --> 00:27:11.860 align:middle line:90% actually trigger these things. 00:27:11.860 --> 00:27:13.867 align:middle line:84% Sometimes cells can decide to kill themselves 00:27:13.867 --> 00:27:15.200 align:middle line:90% for the benefit of the organism. 00:27:15.200 --> 00:27:17.620 align:middle line:84% And that's another common way to start these programs. 00:27:17.620 --> 00:27:19.090 align:middle line:84% And of course, the list of things 00:27:19.090 --> 00:27:21.310 align:middle line:84% that can trigger a transcriptional program 00:27:21.310 --> 00:27:21.920 align:middle line:90% is endless. 00:27:21.920 --> 00:27:23.770 align:middle line:90% It's infinite, right? 00:27:23.770 --> 00:27:26.830 align:middle line:84% I can go like this, and I've just triggered some genes 00:27:26.830 --> 00:27:28.358 align:middle line:90% in my hand, for example. 00:27:28.358 --> 00:27:30.025 align:middle line:84% Anything could start these things going. 00:27:30.025 --> 00:27:33.490 align:middle line:90% 00:27:33.490 --> 00:27:35.230 align:middle line:90% Any questions so far? 00:27:35.230 --> 00:27:39.550 align:middle line:84% I have about three or four slides left on biology. 00:27:39.550 --> 00:27:41.260 align:middle line:84% Is this stuff you all already knew? 00:27:41.260 --> 00:27:43.150 align:middle line:84% Are you learning anything at all so far? 00:27:43.150 --> 00:27:46.480 align:middle line:90% Perhaps, OK. 00:27:46.480 --> 00:27:49.330 align:middle line:84% So this is a little bit more esoteric now, 00:27:49.330 --> 00:27:51.320 align:middle line:90% this idea of temporal programs. 00:27:51.320 --> 00:27:53.560 align:middle line:84% So one of the more important places 00:27:53.560 --> 00:27:56.020 align:middle line:84% where genes are getting turned on and turned off 00:27:56.020 --> 00:27:59.170 align:middle line:84% is during the development of an organism. 00:27:59.170 --> 00:28:01.840 align:middle line:84% So in our 10 months of gestation when 00:28:01.840 --> 00:28:05.200 align:middle line:84% we go from fertilized egg to actually leaving the womb, 00:28:05.200 --> 00:28:06.670 align:middle line:84% whole bunches of genes are getting 00:28:06.670 --> 00:28:10.570 align:middle line:84% turned on and turned off in a pretty well-defined program. 00:28:10.570 --> 00:28:12.490 align:middle line:84% And the picture I'm showing here is 00:28:12.490 --> 00:28:19.210 align:middle line:84% the difference between development and-- 00:28:19.210 --> 00:28:20.710 align:middle line:84% I guess-- so the terms they use here 00:28:20.710 --> 00:28:22.422 align:middle line:84% are segmentation versus homeosis. 00:28:22.422 --> 00:28:24.130 align:middle line:84% Take a look at this picture, for example. 00:28:24.130 --> 00:28:28.720 align:middle line:84% These are two houses built in San Francisco in 1857. 00:28:28.720 --> 00:28:31.262 align:middle line:84% And they both started with the same blueprint. 00:28:31.262 --> 00:28:33.220 align:middle line:84% And they both started out-- they're essentially 00:28:33.220 --> 00:28:35.500 align:middle line:84% the same house right next to each other. 00:28:35.500 --> 00:28:38.290 align:middle line:84% But after more than 100 years, the two houses 00:28:38.290 --> 00:28:40.510 align:middle line:84% look very different from each other. 00:28:40.510 --> 00:28:42.880 align:middle line:84% So this is given as an example on this cell paper 00:28:42.880 --> 00:28:44.590 align:middle line:84% as basically saying, the genes that 00:28:44.590 --> 00:28:47.830 align:middle line:84% are involved in the development of a particular tissue 00:28:47.830 --> 00:28:51.730 align:middle line:84% might not be the same ones that refine the tissue after it's 00:28:51.730 --> 00:28:54.670 align:middle line:84% been built. So the same tissue can 00:28:54.670 --> 00:28:58.180 align:middle line:84% have two different programs on two different time scales 00:28:58.180 --> 00:29:01.143 align:middle line:90% in terms of gene expression. 00:29:01.143 --> 00:29:03.560 align:middle line:84% Now, let's talk about this process of making the messenger 00:29:03.560 --> 00:29:05.060 align:middle line:90% RNA because this is important. 00:29:05.060 --> 00:29:09.170 align:middle line:84% Most common, our most impressive way to measure genes 00:29:09.170 --> 00:29:12.590 align:middle line:84% now happens to do a lot with messenger RNA. 00:29:12.590 --> 00:29:16.010 align:middle line:84% mRNA can be transcribed at several hundred nucleotides 00:29:16.010 --> 00:29:17.490 align:middle line:90% per minute. 00:29:17.490 --> 00:29:20.540 align:middle line:90% That's not that fast, actually. 00:29:20.540 --> 00:29:24.140 align:middle line:84% So if you have a gene that's 10,000 base pairs long, 00:29:24.140 --> 00:29:27.620 align:middle line:84% I'm only doing 100 a minute, maybe a couple 100 minute. 00:29:27.620 --> 00:29:32.330 align:middle line:84% It could take me hours to transcribe one gene. 00:29:32.330 --> 00:29:34.340 align:middle line:84% So in thinking about this process, 00:29:34.340 --> 00:29:36.860 align:middle line:84% this actually puts it into a real time scale 00:29:36.860 --> 00:29:39.560 align:middle line:84% that we can know and love here in terms of minutes. 00:29:39.560 --> 00:29:42.080 align:middle line:84% So the gene dystrophin, dystrophin 00:29:42.080 --> 00:29:43.145 align:middle line:90% is involved in muscle. 00:29:43.145 --> 00:29:45.410 align:middle line:84% It turns out dystrophin has defects that 00:29:45.410 --> 00:29:47.287 align:middle line:90% can lead to muscular dystrophy. 00:29:47.287 --> 00:29:49.370 align:middle line:84% Dystrophin is involved with how muscles are formed 00:29:49.370 --> 00:29:50.360 align:middle line:90% and contract. 00:29:50.360 --> 00:29:55.732 align:middle line:84% One dystrophin gene can take 20 hours to transcribe. 00:29:55.732 --> 00:29:57.440 align:middle line:84% So how are we able to make enough of this 00:29:57.440 --> 00:30:02.690 align:middle line:84% if it takes 20 hours to make, any ideas? 00:30:02.690 --> 00:30:04.430 align:middle line:84% If it takes 20 hours for me to make 00:30:04.430 --> 00:30:08.210 align:middle line:84% one transcript of dystrophin and I need to make a bunch, 00:30:08.210 --> 00:30:10.180 align:middle line:90% how am I going to do that? 00:30:10.180 --> 00:30:11.692 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:30:11.692 --> 00:30:14.150 align:middle line:84% ATUL J. BUTTE: You have to have a whole bunch of ribosomes, 00:30:14.150 --> 00:30:14.720 align:middle line:90% but what? 00:30:14.720 --> 00:30:16.730 align:middle line:84% AUDIENCE: Multiple copies of the gene. 00:30:16.730 --> 00:30:18.688 align:middle line:84% ATUL J. BUTTE: OK, you can have multiple copies 00:30:18.688 --> 00:30:20.990 align:middle line:84% of the gene in the genome, but there's one other way 00:30:20.990 --> 00:30:24.200 align:middle line:90% that you haven't thought of. 00:30:24.200 --> 00:30:27.110 align:middle line:84% Look, I'm telling you it takes 20 hours 00:30:27.110 --> 00:30:29.690 align:middle line:84% at the start of the pipe to the end of the pipe, right? 00:30:29.690 --> 00:30:31.400 align:middle line:84% Now, I'm not telling you how fast I'm 00:30:31.400 --> 00:30:32.690 align:middle line:90% shoving things into the pipe. 00:30:32.690 --> 00:30:36.517 align:middle line:84% I can have one of the things making the [INAUDIBLE] right 00:30:36.517 --> 00:30:38.600 align:middle line:84% behind the next, behind the next, behind the next. 00:30:38.600 --> 00:30:42.830 align:middle line:84% So I could just stream off thousands of copies at a time, 00:30:42.830 --> 00:30:43.760 align:middle line:90% right? 00:30:43.760 --> 00:30:46.640 align:middle line:84% I don't have to wait for one to be done to start the next. 00:30:46.640 --> 00:30:48.440 align:middle line:84% I can have them all go in one base 00:30:48.440 --> 00:30:49.880 align:middle line:84% pair off of each other, a couple of base pairs 00:30:49.880 --> 00:30:50.640 align:middle line:90% off from each other. 00:30:50.640 --> 00:30:52.170 align:middle line:84% That's how I can make more or fewer. 00:30:52.170 --> 00:30:53.705 align:middle line:84% But I still need all the ribosomes 00:30:53.705 --> 00:30:54.830 align:middle line:90% to make them into proteins. 00:30:54.830 --> 00:30:57.410 align:middle line:90% You're absolutely right. 00:30:57.410 --> 00:30:59.780 align:middle line:90% So that's an important concept. 00:30:59.780 --> 00:31:02.570 align:middle line:84% Even though they are long, it doesn't tell you really 00:31:02.570 --> 00:31:06.168 align:middle line:84% how many we can make given any unit time. 00:31:06.168 --> 00:31:08.210 align:middle line:84% If the cell really wants to make a bunch of them, 00:31:08.210 --> 00:31:09.627 align:middle line:84% it could probably figure out a way 00:31:09.627 --> 00:31:12.440 align:middle line:84% to do it by just lining up the polymerases 00:31:12.440 --> 00:31:14.820 align:middle line:90% to make copies of the gene. 00:31:14.820 --> 00:31:17.060 align:middle line:84% Now, it turns out a lot of messenger 00:31:17.060 --> 00:31:20.390 align:middle line:90% RNA happens to end with poly(A). 00:31:20.390 --> 00:31:21.780 align:middle line:90% So what do I mean by this? 00:31:21.780 --> 00:31:25.400 align:middle line:84% So as that little polymerase is working on the DNA streaming 00:31:25.400 --> 00:31:32.090 align:middle line:84% off a copy of the mRNA, when it's done it seems to stutter. 00:31:32.090 --> 00:31:36.560 align:middle line:84% So it adds a whole bunch of A, A, A, A, As at the end. 00:31:36.560 --> 00:31:38.900 align:middle line:90% Lucky for us-- 00:31:38.900 --> 00:31:41.758 align:middle line:84% I'm sure there's some biological reason for this, 00:31:41.758 --> 00:31:43.550 align:middle line:84% but it happens to be very fortunate for us. 00:31:43.550 --> 00:31:47.600 align:middle line:84% Because if we have something that sticks to this wall that's 00:31:47.600 --> 00:31:52.280 align:middle line:84% just made of T, T, T, T, T, it's going to bind to the A, A, A, 00:31:52.280 --> 00:31:54.080 align:middle line:84% A, A. And all of a sudden, we can now 00:31:54.080 --> 00:31:57.500 align:middle line:84% have a big filter for all the messenger RNAs. 00:31:57.500 --> 00:32:00.650 align:middle line:84% All we had to do is stream the messenger RNAs past my little 00:32:00.650 --> 00:32:03.140 align:middle line:84% thing stuck to the wall with the poly(T)s, 00:32:03.140 --> 00:32:05.000 align:middle line:84% and it's going to bind to the poly(A)s. 00:32:05.000 --> 00:32:06.980 align:middle line:84% That's how we can fish out all the messenger 00:32:06.980 --> 00:32:09.560 align:middle line:84% RNAs from the general pool of RNAs 00:32:09.560 --> 00:32:12.080 align:middle line:84% that might be out there because not all RNAs are actually 00:32:12.080 --> 00:32:14.180 align:middle line:90% coding for proteins. 00:32:14.180 --> 00:32:17.330 align:middle line:84% And in general, this is how we detect RNAs. 00:32:17.330 --> 00:32:20.000 align:middle line:84% You can take the sequence of RNAs 00:32:20.000 --> 00:32:22.920 align:middle line:84% and build the reverse complement of it. 00:32:22.920 --> 00:32:25.400 align:middle line:84% So everywhere there's a G, put a C in my detection. 00:32:25.400 --> 00:32:28.010 align:middle line:84% And if there's an A, put a T. If there's a C, put a G. 00:32:28.010 --> 00:32:30.140 align:middle line:84% So we can use the reverse complement 00:32:30.140 --> 00:32:32.120 align:middle line:84% to help us tell whether this sequence is 00:32:32.120 --> 00:32:34.370 align:middle line:84% there or not depending on how we designed the probe. 00:32:34.370 --> 00:32:38.030 align:middle line:84% And we're going to talk about that. 00:32:38.030 --> 00:32:39.530 align:middle line:84% So this is, I think, the last slide. 00:32:39.530 --> 00:32:41.650 align:middle line:84% So why are we trying to do this whole Genome Project? 00:32:41.650 --> 00:32:44.140 align:middle line:84% Why was it important to actually complete the Genome Project? 00:32:44.140 --> 00:32:46.060 align:middle line:84% [? Eric ?] [? Weiner ?] basically uses this analogy 00:32:46.060 --> 00:32:48.185 align:middle line:84% that knowing all the genes is equivalent to knowing 00:32:48.185 --> 00:32:50.800 align:middle line:84% the periodic table of the elements for biology. 00:32:50.800 --> 00:32:54.880 align:middle line:84% So around 1850s, 1860s, 1870s, Mendeleev 00:32:54.880 --> 00:32:57.648 align:middle line:84% came up with the way of lining up all the elements in such 00:32:57.648 --> 00:32:59.440 align:middle line:84% a way to be able to predict characteristics 00:32:59.440 --> 00:33:00.770 align:middle line:90% of those elements. 00:33:00.770 --> 00:33:03.480 align:middle line:84% Now, back then, most of this was empty spaces 00:33:03.480 --> 00:33:05.980 align:middle line:84% because they hadn't seen some of these rare elements before. 00:33:05.980 --> 00:33:09.460 align:middle line:84% But they knew, by golly, there has to be an element right here 00:33:09.460 --> 00:33:11.050 align:middle line:90% that's inert, for example. 00:33:11.050 --> 00:33:13.482 align:middle line:84% Or there's got to be an element here that's magnetic. 00:33:13.482 --> 00:33:15.940 align:middle line:84% They can make these predictions even though they didn't see 00:33:15.940 --> 00:33:17.107 align:middle line:90% what was actually there yet. 00:33:17.107 --> 00:33:18.220 align:middle line:90% They hadn't discovered it. 00:33:18.220 --> 00:33:21.307 align:middle line:84% The same way, the periodic table for biology for genomics 00:33:21.307 --> 00:33:22.390 align:middle line:90% isn't going to be a table. 00:33:22.390 --> 00:33:23.980 align:middle line:84% It might be a tree because that's 00:33:23.980 --> 00:33:25.480 align:middle line:84% how these genes are actually formed, 00:33:25.480 --> 00:33:27.380 align:middle line:90% through duplication events. 00:33:27.380 --> 00:33:29.800 align:middle line:84% And so the idea is maybe we can predict 00:33:29.800 --> 00:33:31.930 align:middle line:84% that there must be something in the genome that's 00:33:31.930 --> 00:33:32.597 align:middle line:90% binding to this. 00:33:32.597 --> 00:33:34.555 align:middle line:84% So there must be something in the genome that's 00:33:34.555 --> 00:33:35.710 align:middle line:90% responsible for that. 00:33:35.710 --> 00:33:37.210 align:middle line:84% Now that we have the entire catalog, 00:33:37.210 --> 00:33:39.340 align:middle line:84% we can start to fish those things out. 00:33:39.340 --> 00:33:40.840 align:middle line:84% And what this whole course is is how 00:33:40.840 --> 00:33:43.360 align:middle line:90% this is relevant to medicine. 00:33:43.360 --> 00:33:45.220 align:middle line:84% So most of the figures I've shown you here 00:33:45.220 --> 00:33:46.502 align:middle line:90% are from this book Genomes. 00:33:46.502 --> 00:33:48.085 align:middle line:84% In fact, there's a second edition now. 00:33:48.085 --> 00:33:49.630 align:middle line:84% It might be one of the required reading 00:33:49.630 --> 00:33:50.797 align:middle line:90% or something for the course. 00:33:50.797 --> 00:33:51.730 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:33:51.730 --> 00:33:53.772 align:middle line:84% ATUL J. BUTTE: But it's freely available at NCBI. 00:33:53.772 --> 00:33:55.720 align:middle line:90% Definitely know this. 00:33:55.720 --> 00:33:58.300 align:middle line:84% So I would-- and in fact, all the figures are there, too, 00:33:58.300 --> 00:34:00.092 align:middle line:84% if you're interested in adding these things 00:34:00.092 --> 00:34:01.210 align:middle line:90% to reports and stuff. 00:34:01.210 --> 00:34:03.460 align:middle line:84% I highly recommend this more than anything 00:34:03.460 --> 00:34:06.400 align:middle line:84% else only because it's very readable, lots of great figures 00:34:06.400 --> 00:34:08.920 align:middle line:84% to help explain these processes, and it's very cutting edge, 00:34:08.920 --> 00:34:10.510 align:middle line:84% starts with proteomics, mass spec, 00:34:10.510 --> 00:34:13.570 align:middle line:84% microarrays, the whole works, much better than other books, 00:34:13.570 --> 00:34:15.400 align:middle line:90% relevant books in this field. 00:34:15.400 --> 00:34:17.320 align:middle line:84% And the Department of Medicine at this website 00:34:17.320 --> 00:34:18.280 align:middle line:90% has a primer on this. 00:34:18.280 --> 00:34:19.659 align:middle line:90% It's starting to get old now. 00:34:19.659 --> 00:34:20.679 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:34:20.679 --> 00:34:23.492 align:middle line:84% You guys received another email OK? 00:34:23.492 --> 00:34:25.659 align:middle line:84% ATUL J. BUTTE: This is starting to get a little old, 00:34:25.659 --> 00:34:26.909 align:middle line:90% but it's still quite relevant. 00:34:26.909 --> 00:34:30.130 align:middle line:84% Some of that stuff will never go out of date. 00:34:30.130 --> 00:34:33.627 align:middle line:84% OK, so I got another, let's say, 45 minutes or so? 00:34:33.627 --> 00:34:34.210 align:middle line:90% AUDIENCE: Yup. 00:34:34.210 --> 00:34:34.960 align:middle line:90% ATUL J. BUTTE: OK. 00:34:34.960 --> 00:34:37.969 align:middle line:90% 00:34:37.969 --> 00:34:41.840 align:middle line:84% Let's talk about gene measurement techniques. 00:34:41.840 --> 00:34:45.900 align:middle line:84% I will come up with something different for Tuesday. 00:34:45.900 --> 00:34:48.030 align:middle line:84% So what I'm going to talk about now-- 00:34:48.030 --> 00:34:49.940 align:middle line:90% so any questions so far? 00:34:49.940 --> 00:34:50.440 align:middle line:90% Yes. 00:34:50.440 --> 00:34:54.830 align:middle line:84% AUDIENCE: [INAUDIBLE] time scale for the transcription? 00:34:54.830 --> 00:34:56.530 align:middle line:90% ATUL J. BUTTE: Sure. 00:34:56.530 --> 00:35:01.730 align:middle line:84% AUDIENCE: Does anybody know like said 00:35:01.730 --> 00:35:05.330 align:middle line:84% you can line up multiple transcription [INAUDIBLE] 00:35:05.330 --> 00:35:07.940 align:middle line:84% to combat these slower [INAUDIBLE],, right? 00:35:07.940 --> 00:35:09.750 align:middle line:90% ATUL J. BUTTE: Yup. 00:35:09.750 --> 00:35:14.220 align:middle line:84% AUDIENCE: And I guess assuming that transcription complex 00:35:14.220 --> 00:35:16.832 align:middle line:90% binds to the start-- 00:35:16.832 --> 00:35:17.790 align:middle line:90% ATUL J. BUTTE: Exactly. 00:35:17.790 --> 00:35:20.690 align:middle line:84% AUDIENCE: --and then faster than it 00:35:20.690 --> 00:35:25.430 align:middle line:84% moves along the strain, what's the lower limit? 00:35:25.430 --> 00:35:28.630 align:middle line:84% I guess the [INAUDIBLE] what's the fastest these things move? 00:35:28.630 --> 00:35:30.130 align:middle line:84% ATUL J. BUTTE: So what's the fastest 00:35:30.130 --> 00:35:32.140 align:middle line:90% that gene can be transcribed? 00:35:32.140 --> 00:35:32.950 align:middle line:90% AUDIENCE: Yeah. 00:35:32.950 --> 00:35:36.370 align:middle line:84% I mean because you've got the first transcription complex 00:35:36.370 --> 00:35:37.823 align:middle line:90% just move out of the way. 00:35:37.823 --> 00:35:38.906 align:middle line:90% ATUL J. BUTTE: Absolutely. 00:35:38.906 --> 00:35:39.610 align:middle line:90% Yeah, exactly. 00:35:39.610 --> 00:35:42.287 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:35:42.287 --> 00:35:44.620 align:middle line:84% ATUL J. BUTTE: You could probably do the math and figure 00:35:44.620 --> 00:35:45.490 align:middle line:90% this out, OK? 00:35:45.490 --> 00:35:45.580 align:middle line:90% 00:35:45.580 --> 00:35:46.760 align:middle line:84% AUDIENCE: I think that's a great project, actually. 00:35:46.760 --> 00:35:47.170 align:middle line:90% ATUL J. BUTTE: Right. 00:35:47.170 --> 00:35:47.670 align:middle line:90% So-- 00:35:47.670 --> 00:35:48.687 align:middle line:90% AUDIENCE: Helps with a-- 00:35:48.687 --> 00:35:49.270 align:middle line:90% no, seriously. 00:35:49.270 --> 00:35:49.660 align:middle line:90% ATUL J. BUTTE: Absolutely. 00:35:49.660 --> 00:35:50.743 align:middle line:90% You can model this, right? 00:35:50.743 --> 00:35:50.842 align:middle line:90% 00:35:50.842 --> 00:35:53.217 align:middle line:84% AUDIENCE: [INAUDIBLE] a sound basis understand everything 00:35:53.217 --> 00:35:57.280 align:middle line:84% we know about the initiation of the [? transcriptual ?] 00:35:57.280 --> 00:35:58.810 align:middle line:84% site and moving on and the letting 00:35:58.810 --> 00:36:00.970 align:middle line:84% go, all the things that have to happen. 00:36:00.970 --> 00:36:04.060 align:middle line:84% And there are quite detailed documentations 00:36:04.060 --> 00:36:06.530 align:middle line:84% of the [? characteristics ?] and processes. 00:36:06.530 --> 00:36:10.002 align:middle line:84% What is the upper bound for transcription and translation? 00:36:10.002 --> 00:36:12.460 align:middle line:84% ATUL J. BUTTE: Well, basically, the model, the whole thing, 00:36:12.460 --> 00:36:14.050 align:middle line:90% is a pipe or a channel. 00:36:14.050 --> 00:36:16.870 align:middle line:84% And I already told you the length of this thing, 00:36:16.870 --> 00:36:19.120 align:middle line:84% how long it takes to make dystrophin. 00:36:19.120 --> 00:36:22.265 align:middle line:84% You just have to model how fast I can put the next one in. 00:36:22.265 --> 00:36:23.860 align:middle line:84% You might be able to model it based 00:36:23.860 --> 00:36:27.430 align:middle line:84% on just looking at the size of the polymerase and maybe 00:36:27.430 --> 00:36:29.248 align:middle line:84% how many base pairs it's covering up. 00:36:29.248 --> 00:36:30.790 align:middle line:84% Maybe it's got to move out of the way 00:36:30.790 --> 00:36:32.123 align:middle line:90% before the next one can come in. 00:36:32.123 --> 00:36:34.040 align:middle line:84% You can model it just like that, for example. 00:36:34.040 --> 00:36:35.320 align:middle line:84% It would be quite interesting to see 00:36:35.320 --> 00:36:37.270 align:middle line:84% how that matches up with the actual published 00:36:37.270 --> 00:36:39.340 align:middle line:90% numbers for it. 00:36:39.340 --> 00:36:41.600 align:middle line:84% My bet is that's probably still not the rate 00:36:41.600 --> 00:36:42.850 align:middle line:90% limiting step in this process. 00:36:42.850 --> 00:36:44.350 align:middle line:84% That's another interesting question. 00:36:44.350 --> 00:36:46.763 align:middle line:90% What's the slow part here? 00:36:46.763 --> 00:36:47.680 align:middle line:90% I'm not actually sure. 00:36:47.680 --> 00:36:49.650 align:middle line:84% It probably depends on the genes, actually. 00:36:49.650 --> 00:36:51.400 align:middle line:84% There are some genes that are only-- there 00:36:51.400 --> 00:36:55.660 align:middle line:84% are some important proteins that are only 10 amino acids long. 00:36:55.660 --> 00:36:59.380 align:middle line:84% Insulin, for example, is one with 50 amino acids. 00:36:59.380 --> 00:37:02.990 align:middle line:84% So some of those can be spit out pretty quick. 00:37:02.990 --> 00:37:06.010 align:middle line:84% And then you got some that are huge like dystrophin and titin, 00:37:06.010 --> 00:37:08.020 align:middle line:90% some of these others. 00:37:08.020 --> 00:37:10.480 align:middle line:90% Any other questions so far? 00:37:10.480 --> 00:37:15.000 align:middle line:84% AUDIENCE: [INAUDIBLE] in terms of [INAUDIBLE],, what 00:37:15.000 --> 00:37:16.780 align:middle line:90% in terms of bind [INAUDIBLE]? 00:37:16.780 --> 00:37:18.530 align:middle line:84% ATUL J. BUTTE: So again, remember, there's 00:37:18.530 --> 00:37:20.120 align:middle line:90% no nice line there. 00:37:20.120 --> 00:37:21.510 align:middle line:90% It's got to be in the sequence. 00:37:21.510 --> 00:37:23.120 align:middle line:84% So there are specific sequences that 00:37:23.120 --> 00:37:26.030 align:middle line:84% are known to be intron-exon boundaries, 00:37:26.030 --> 00:37:28.130 align:middle line:90% but clearly they're not 100%. 00:37:28.130 --> 00:37:30.380 align:middle line:84% Because in other cells, that same sequence 00:37:30.380 --> 00:37:32.600 align:middle line:84% might result in another transcript. 00:37:32.600 --> 00:37:34.910 align:middle line:84% So there are intron-exon boundaries 00:37:34.910 --> 00:37:38.120 align:middle line:84% about the start of the intron and the end of the intron. 00:37:38.120 --> 00:37:40.737 align:middle line:84% And it turns out one end of it complements the other side. 00:37:40.737 --> 00:37:42.320 align:middle line:84% So these things are actually taken out 00:37:42.320 --> 00:37:44.935 align:middle line:90% in a loop type of form. 00:37:44.935 --> 00:37:46.310 align:middle line:84% So the thing that cuts this thing 00:37:46.310 --> 00:37:48.653 align:middle line:84% out lines up the mRNA in such a way 00:37:48.653 --> 00:37:51.320 align:middle line:84% that's able to put the start and end together and just basically 00:37:51.320 --> 00:37:53.660 align:middle line:84% snip the whole thing out like that. 00:37:53.660 --> 00:37:57.300 align:middle line:84% I can refer you to pictures in the genome exactly on this. 00:37:57.300 --> 00:37:58.460 align:middle line:90% But again, it's not 100%. 00:37:58.460 --> 00:38:00.800 align:middle line:84% AUDIENCE: [INAUDIBLE] like GA, for instance, 00:38:00.800 --> 00:38:03.230 align:middle line:84% is one of the dinucleotides that sometimes 00:38:03.230 --> 00:38:08.840 align:middle line:90% initiates the splicing complex. 00:38:08.840 --> 00:38:11.045 align:middle line:84% But often it doesn't, meaning it's 00:38:11.045 --> 00:38:13.310 align:middle line:90% a very common dinucleotide. 00:38:13.310 --> 00:38:19.880 align:middle line:84% So [INAUDIBLE] the syntax that determines the splicing is not 00:38:19.880 --> 00:38:20.760 align:middle line:90% [INAUDIBLE]. 00:38:20.760 --> 00:38:25.160 align:middle line:84% So we have some heuristics which work some of the time. 00:38:25.160 --> 00:38:28.500 align:middle line:84% We have not reverse engineered [INAUDIBLE] 00:38:28.500 --> 00:38:30.380 align:middle line:90% the alternative splicing code. 00:38:30.380 --> 00:38:31.880 align:middle line:84% ATUL J. BUTTE: You think about this. 00:38:31.880 --> 00:38:35.000 align:middle line:84% This is a massive decryption effort in the end. 00:38:35.000 --> 00:38:37.160 align:middle line:84% All we have are the As, Ts, Cs, and Gs. 00:38:37.160 --> 00:38:39.890 align:middle line:84% One level of coding is the triplet codon 00:38:39.890 --> 00:38:42.440 align:middle line:84% for the amino acids, but another more interesting one 00:38:42.440 --> 00:38:43.968 align:middle line:90% is the grammar here. 00:38:43.968 --> 00:38:45.260 align:middle line:90% Where is the start of the gene? 00:38:45.260 --> 00:38:45.980 align:middle line:90% Where is the end of the gene? 00:38:45.980 --> 00:38:46.580 align:middle line:90% When do I splice it? 00:38:46.580 --> 00:38:47.663 align:middle line:90% And when do I splice that? 00:38:47.663 --> 00:38:50.750 align:middle line:84% What are all the transcriptional regulatory elements? 00:38:50.750 --> 00:38:52.602 align:middle line:90% That's the Holy Grail here. 00:38:52.602 --> 00:38:53.560 align:middle line:90% That is the Holy Grail. 00:38:53.560 --> 00:38:56.630 align:middle line:90% AUDIENCE: [INAUDIBLE]? 00:38:56.630 --> 00:38:58.157 align:middle line:90% ATUL J. BUTTE: We know those. 00:38:58.157 --> 00:38:59.990 align:middle line:84% AUDIENCE: But it doesn't always work either. 00:38:59.990 --> 00:39:00.920 align:middle line:90% ATUL J. BUTTE: Absolutely. 00:39:00.920 --> 00:39:01.430 align:middle line:90% OK. 00:39:01.430 --> 00:39:02.180 align:middle line:90% So I'll give you-- 00:39:02.180 --> 00:39:04.130 align:middle line:90% I'll tell you exactly why. 00:39:04.130 --> 00:39:07.320 align:middle line:84% The code for the start, there's only one code for start, 00:39:07.320 --> 00:39:10.880 align:middle line:84% AUG, but there's three codes to stop. 00:39:10.880 --> 00:39:14.360 align:middle line:84% It turns out one of the codes to stop, 00:39:14.360 --> 00:39:18.560 align:middle line:84% if there's another code 100 base pairs later, 00:39:18.560 --> 00:39:22.500 align:middle line:90% that makes the 21st amino acid. 00:39:22.500 --> 00:39:25.000 align:middle line:84% So in other words, you can't even just look at the triplets. 00:39:25.000 --> 00:39:27.525 align:middle line:84% It's in the context of other things, too. 00:39:27.525 --> 00:39:30.040 align:middle line:90% It makes things messy again. 00:39:30.040 --> 00:39:32.290 align:middle line:84% It's so nice and neat with just triplet codons. 00:39:32.290 --> 00:39:35.590 align:middle line:84% But that stop plus something else 100 base pairs later 00:39:35.590 --> 00:39:41.000 align:middle line:84% gives, actually, the 21st amino acid, selenocysteine. 00:39:41.000 --> 00:39:42.990 align:middle line:90% That's life, messy. 00:39:42.990 --> 00:39:45.110 align:middle line:84% AUDIENCE: So people discovered the selenocysteine. 00:39:45.110 --> 00:39:46.860 align:middle line:84% ATUL J. BUTTE: People for the life of them 00:39:46.860 --> 00:39:47.860 align:middle line:90% couldn't figure out why. 00:39:47.860 --> 00:39:50.300 align:middle line:84% How are you making selenocysteine proteins? 00:39:50.300 --> 00:39:52.760 align:middle line:84% There's only four in humans, one of which 00:39:52.760 --> 00:39:54.650 align:middle line:84% is involved in the thyroid, which 00:39:54.650 --> 00:39:56.900 align:middle line:84% is why a whole bunch of endocrinologists look at this. 00:39:56.900 --> 00:40:01.340 align:middle line:84% It metabolizes thyroid hormone into one form to another, 00:40:01.340 --> 00:40:05.810 align:middle line:90% thyroxine to triiodothyronine. 00:40:05.810 --> 00:40:07.553 align:middle line:84% That enzyme has a selenocysteine. 00:40:07.553 --> 00:40:08.970 align:middle line:84% And people couldn't figure out why 00:40:08.970 --> 00:40:11.000 align:middle line:84% it was coding for selenocysteine because it 00:40:11.000 --> 00:40:12.870 align:middle line:90% looked like a stop codon their. 00:40:12.870 --> 00:40:14.870 align:middle line:84% Well, you could start to figure these things out 00:40:14.870 --> 00:40:16.345 align:middle line:84% once you have the whole genome now. 00:40:16.345 --> 00:40:17.720 align:middle line:84% You can look at these exceptions. 00:40:17.720 --> 00:40:25.070 align:middle line:84% AUDIENCE: [INAUDIBLE] the function 00:40:25.070 --> 00:40:26.925 align:middle line:90% of [INAUDIBLE] mice and humans. 00:40:26.925 --> 00:40:28.400 align:middle line:90% What [INAUDIBLE]? 00:40:28.400 --> 00:40:30.650 align:middle line:84% ATUL J. BUTTE: OK, let's be very specific. 00:40:30.650 --> 00:40:35.900 align:middle line:84% The repeats code for a protein that comes back, 00:40:35.900 --> 00:40:40.130 align:middle line:84% looks for the sequence, and actually integrates 00:40:40.130 --> 00:40:44.390 align:middle line:84% the sequence somewhere else in the genome. 00:40:44.390 --> 00:40:48.200 align:middle line:84% It repeats code for protein like any other gene 00:40:48.200 --> 00:40:50.480 align:middle line:84% that finds its sequence and puts it 00:40:50.480 --> 00:40:52.850 align:middle line:90% in somewhere else in the genome. 00:40:52.850 --> 00:40:57.110 align:middle line:84% But for that to happen the code for the protein has to work. 00:40:57.110 --> 00:41:02.087 align:middle line:84% If it's damaged beyond repair, the thing can't work. 00:41:02.087 --> 00:41:03.420 align:middle line:90% And sometimes it makes mistakes. 00:41:03.420 --> 00:41:04.872 align:middle line:90% So the thing's coming back. 00:41:04.872 --> 00:41:05.580 align:middle line:90% I made a protein. 00:41:05.580 --> 00:41:06.498 align:middle line:90% It's coming back. 00:41:06.498 --> 00:41:08.040 align:middle line:84% It's trying to make a copy of itself. 00:41:08.040 --> 00:41:10.248 align:middle line:84% Sometimes it grabs the wrong thing to make a copy of. 00:41:10.248 --> 00:41:12.790 align:middle line:84% So that's why we have copies of lots of things in the genome. 00:41:12.790 --> 00:41:14.010 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:41:14.010 --> 00:41:16.972 align:middle line:90% 00:41:16.972 --> 00:41:18.180 align:middle line:90% ATUL J. BUTTE: Good question. 00:41:18.180 --> 00:41:20.280 align:middle line:84% It's able to recognize its own sequence, 00:41:20.280 --> 00:41:23.535 align:middle line:90% but it's not at all perfect. 00:41:23.535 --> 00:41:25.785 align:middle line:84% But what we can say about the difference between human 00:41:25.785 --> 00:41:29.160 align:middle line:84% and mouse is that the mouse has those sequences 00:41:29.160 --> 00:41:31.540 align:middle line:84% for that protein, a reverse transcriptase. 00:41:31.540 --> 00:41:35.010 align:middle line:84% It has codes for it that seem to be fully functional. 00:41:35.010 --> 00:41:38.550 align:middle line:84% And in a human, there are none that are still functional. 00:41:38.550 --> 00:41:40.920 align:middle line:90% That's what we're saying. 00:41:40.920 --> 00:41:45.390 align:middle line:84% It's very curious why that is, this whole area 00:41:45.390 --> 00:41:49.578 align:middle line:84% of genomic archaeology that you can do with the genome now. 00:41:49.578 --> 00:41:51.120 align:middle line:84% In fact, what Zach was saying earlier 00:41:51.120 --> 00:41:54.750 align:middle line:84% is that we have so many millions of repeats 00:41:54.750 --> 00:41:58.260 align:middle line:84% that we can look to see how deviant they are from a working 00:41:58.260 --> 00:41:59.430 align:middle line:90% copy. 00:41:59.430 --> 00:42:02.160 align:middle line:84% And so we can age each of the repeats. 00:42:02.160 --> 00:42:03.870 align:middle line:90% This one's 40 million years old. 00:42:03.870 --> 00:42:05.350 align:middle line:90% This one's 50 million years old. 00:42:05.350 --> 00:42:07.870 align:middle line:84% Because if it's got four places where it's different 00:42:07.870 --> 00:42:10.170 align:middle line:84% and this one only has two, the one with four 00:42:10.170 --> 00:42:12.960 align:middle line:84% is probably older than the one with two, right? 00:42:12.960 --> 00:42:16.417 align:middle line:84% So like that, we can age all of these repeats. 00:42:16.417 --> 00:42:18.000 align:middle line:84% And it turns out a whole bunch of them 00:42:18.000 --> 00:42:20.810 align:middle line:84% were made at a certain point in time and then they dropped off. 00:42:20.810 --> 00:42:21.810 align:middle line:90% Again, no one knows why. 00:42:21.810 --> 00:42:23.280 align:middle line:84% I'll show you the graph in the second here. 00:42:23.280 --> 00:42:24.600 align:middle line:84% AUDIENCE: So we can assume that the genome is 00:42:24.600 --> 00:42:25.600 align:middle line:90% growing all the time. 00:42:25.600 --> 00:42:29.128 align:middle line:84% So the [INAUDIBLE] much bigger than that. 00:42:29.128 --> 00:42:30.420 align:middle line:90% ATUL J. BUTTE: Or smaller, too. 00:42:30.420 --> 00:42:33.210 align:middle line:84% There's ways to get rid of these things, too. 00:42:33.210 --> 00:42:36.100 align:middle line:84% If you do too much damage, the thing can't reproduce. 00:42:36.100 --> 00:42:36.940 align:middle line:90% And that's it. 00:42:36.940 --> 00:42:38.590 align:middle line:90% It's the end of the line. 00:42:38.590 --> 00:42:41.400 align:middle line:84% The most mind blowing thing that the Genome Project has 00:42:41.400 --> 00:42:44.250 align:middle line:84% done for me personally is that, now, I 00:42:44.250 --> 00:42:48.180 align:middle line:84% realize there's only one life form on this planet. 00:42:48.180 --> 00:42:49.680 align:middle line:90% It's DNA. 00:42:49.680 --> 00:42:52.913 align:middle line:84% Everything else is a side effect. 00:42:52.913 --> 00:42:55.080 align:middle line:84% That DNA is doing everything it can to actually keep 00:42:55.080 --> 00:42:56.850 align:middle line:90% its code going. 00:42:56.850 --> 00:42:57.870 align:middle line:90% That's all there's is. 00:42:57.870 --> 00:42:59.078 align:middle line:90% There's a great book called-- 00:42:59.078 --> 00:43:00.210 align:middle line:90% AUDIENCE: The Self Machine. 00:43:00.210 --> 00:43:01.140 align:middle line:84% ATUL J. BUTTE: The Self Machine, of course, 00:43:01.140 --> 00:43:03.640 align:middle line:84% but there's also a more recent one called-- it's on genomes. 00:43:03.640 --> 00:43:07.260 align:middle line:90% It's a paperback. 00:43:07.260 --> 00:43:09.780 align:middle line:84% The Autobiography of a Genome I think it's called. 00:43:09.780 --> 00:43:10.807 align:middle line:90% It's like 23 chapters. 00:43:10.807 --> 00:43:13.140 align:middle line:84% And basically, they just pick one gene from each chapter 00:43:13.140 --> 00:43:13.785 align:middle line:90% and make a story. 00:43:13.785 --> 00:43:14.130 align:middle line:90% AUDIENCE: Oh, yes. 00:43:14.130 --> 00:43:14.910 align:middle line:84% ATUL J. BUTTE: It's a great book. 00:43:14.910 --> 00:43:15.952 align:middle line:90% I recommend reading that. 00:43:15.952 --> 00:43:17.613 align:middle line:90% It's a lay kind of reading. 00:43:17.613 --> 00:43:18.780 align:middle line:90% AUDIENCE: Yeah, [INAUDIBLE]. 00:43:18.780 --> 00:43:20.160 align:middle line:84% ATUL J. BUTTE: It was on a bestseller list for a while. 00:43:20.160 --> 00:43:21.118 align:middle line:90% AUDIENCE: That's right. 00:43:21.118 --> 00:43:25.610 align:middle line:84% ATUL J. BUTTE: Autobiography of a Genome I think it's called, 00:43:25.610 --> 00:43:26.854 align:middle line:90% something like that. 00:43:26.854 --> 00:43:28.813 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:43:28.813 --> 00:43:30.480 align:middle line:84% ATUL J. BUTTE: Some chapters were harder 00:43:30.480 --> 00:43:32.313 align:middle line:84% to write than others, I think, for this guy. 00:43:32.313 --> 00:43:34.245 align:middle line:90% [INTERPOSING VOICES] 00:43:34.245 --> 00:43:38.325 align:middle line:84% AUDIENCE: It's just a question about [INAUDIBLE] age, 00:43:38.325 --> 00:43:42.090 align:middle line:90% you know, and aging [INAUDIBLE]. 00:43:42.090 --> 00:43:48.450 align:middle line:84% Given that you can pretty much get rid of or add to or-- 00:43:48.450 --> 00:43:51.090 align:middle line:84% it seems like dating genes according 00:43:51.090 --> 00:43:54.810 align:middle line:84% to [INAUDIBLE] number of mutations 00:43:54.810 --> 00:43:56.190 align:middle line:90% against mutation frequency. 00:43:56.190 --> 00:44:02.890 align:middle line:84% And it seems kind of invalid now that we know that there are 00:44:02.890 --> 00:44:03.513 align:middle line:90% [INAUDIBLE]. 00:44:03.513 --> 00:44:05.680 align:middle line:84% AUDIENCE: Are you saying that because of these kinds 00:44:05.680 --> 00:44:11.380 align:middle line:84% of influences just looking at, let's say, 00:44:11.380 --> 00:44:13.690 align:middle line:84% the refresh rate of synonymous substitutions 00:44:13.690 --> 00:44:16.090 align:middle line:84% versus non-synonymous substitutions, 00:44:16.090 --> 00:44:19.610 align:middle line:90% it's not a good clock anymore? 00:44:19.610 --> 00:44:21.070 align:middle line:90% AUDIENCE: I guess. 00:44:21.070 --> 00:44:22.480 align:middle line:90% I guess I'm doing-- 00:44:22.480 --> 00:44:25.735 align:middle line:84% what I'm driving at is it seems like one 00:44:25.735 --> 00:44:29.485 align:middle line:84% of the main-- one of the underlying points that you've 00:44:29.485 --> 00:44:32.890 align:middle line:84% been making is that there's a lot of change 00:44:32.890 --> 00:44:35.170 align:middle line:84% and a lot of mistakes that can be introduced 00:44:35.170 --> 00:44:37.690 align:middle line:84% from a variety of sources, none of which 00:44:37.690 --> 00:44:41.200 align:middle line:84% have to be all that magnificent or catastrophic. 00:44:41.200 --> 00:44:43.810 align:middle line:90% It's every day types of things. 00:44:43.810 --> 00:44:48.730 align:middle line:84% And yet, on the other hand, when we're talking about, 00:44:48.730 --> 00:44:50.605 align:middle line:84% well, this gene's probably over because it 00:44:50.605 --> 00:44:51.995 align:middle line:90% has four mutations, not two-- 00:44:51.995 --> 00:44:52.870 align:middle line:90% ATUL J. BUTTE: Right. 00:44:52.870 --> 00:44:54.170 align:middle line:90% It's all based on-- 00:44:54.170 --> 00:44:55.280 align:middle line:90% that's a great point. 00:44:55.280 --> 00:44:58.570 align:middle line:84% We can age these things because we think we know models. 00:44:58.570 --> 00:45:03.850 align:middle line:84% So given a fixed rate of nucleotide substitution 00:45:03.850 --> 00:45:06.220 align:middle line:84% where people have tried to calculate numbers for that, 00:45:06.220 --> 00:45:07.300 align:middle line:90% we can make a premise. 00:45:07.300 --> 00:45:09.970 align:middle line:84% But if that's not true, then it's not going to hold. 00:45:09.970 --> 00:45:11.800 align:middle line:84% AUDIENCE: So first, if you have a transpose 00:45:11.800 --> 00:45:14.530 align:middle line:84% that on the [? box ?] right in the middle of the gene, then 00:45:14.530 --> 00:45:16.900 align:middle line:84% that would be invalid when you look 00:45:16.900 --> 00:45:19.130 align:middle line:84% at the rate of the mutation processing [INAUDIBLE].. 00:45:19.130 --> 00:45:21.100 align:middle line:90% It got a new puzzle. 00:45:21.100 --> 00:45:23.740 align:middle line:84% [INAUDIBLE] like a new to maybe kind of compare. 00:45:23.740 --> 00:45:26.515 align:middle line:84% The model [? is wrong. ?] But if you have a stretch, I mean-- 00:45:26.515 --> 00:45:27.390 align:middle line:90% ATUL J. BUTTE: Right. 00:45:27.390 --> 00:45:29.723 align:middle line:84% AUDIENCE: --a stretch, it's more or less the same except 00:45:29.723 --> 00:45:32.650 align:middle line:84% for some [INAUDIBLE],, then a lot of these-- 00:45:32.650 --> 00:45:36.850 align:middle line:84% not catastrophic, but more violent changes has to happen. 00:45:36.850 --> 00:45:39.220 align:middle line:84% ATUL J. BUTTE: It's all a matter of time scales. 00:45:39.220 --> 00:45:41.320 align:middle line:84% These things are making copies of themselves 00:45:41.320 --> 00:45:44.090 align:middle line:90% over millions of years. 00:45:44.090 --> 00:45:46.420 align:middle line:84% It's not a day to day kind of thing at all. 00:45:46.420 --> 00:45:49.930 align:middle line:84% The cell in my lung can make a copy of itself 00:45:49.930 --> 00:45:51.700 align:middle line:84% and move it somewhere in the genome. 00:45:51.700 --> 00:45:53.660 align:middle line:84% On the grand scheme of things, who cares? 00:45:53.660 --> 00:45:55.180 align:middle line:90% It's not in my sperm cells. 00:45:55.180 --> 00:45:57.160 align:middle line:90% It's not going to go on. 00:45:57.160 --> 00:45:59.650 align:middle line:84% So it's got to happen in the right cell at the right time. 00:45:59.650 --> 00:46:01.192 align:middle line:84% Across millions of years these things 00:46:01.192 --> 00:46:02.980 align:middle line:84% are happening when you think about it. 00:46:02.980 --> 00:46:04.565 align:middle line:84% It's got to be in the right cell, too. 00:46:04.565 --> 00:46:06.940 align:middle line:84% AUDIENCE: When you look at the [INAUDIBLE] between genes, 00:46:06.940 --> 00:46:10.360 align:middle line:84% is there anything to say that the rate of the mutations 00:46:10.360 --> 00:46:13.487 align:middle line:84% would be about the same all regions of [INAUDIBLE]?? 00:46:13.487 --> 00:46:14.320 align:middle line:90% ATUL J. BUTTE: Yeah. 00:46:14.320 --> 00:46:16.195 align:middle line:84% AUDIENCE: Actually, one of [INAUDIBLE] study. 00:46:16.195 --> 00:46:17.080 align:middle line:90% [INAUDIBLE] 00:46:17.080 --> 00:46:19.705 align:middle line:84% ATUL J. BUTTE: Yeah, I think the rates are definitely different 00:46:19.705 --> 00:46:22.900 align:middle line:84% between different regions only because some of these processes 00:46:22.900 --> 00:46:25.060 align:middle line:90% have different rates. 00:46:25.060 --> 00:46:27.670 align:middle line:84% Remember the banding pattern I showed you in the AT-rich? 00:46:27.670 --> 00:46:30.970 align:middle line:84% That corresponds to what regions are rich in AT and rich in GC. 00:46:30.970 --> 00:46:33.700 align:middle line:84% Some processes occur at different rates between them 00:46:33.700 --> 00:46:35.935 align:middle line:84% because there are tighter bonds there. 00:46:35.935 --> 00:46:37.288 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:46:37.288 --> 00:46:38.830 align:middle line:84% ATUL J. BUTTE: It might, but we don't 00:46:38.830 --> 00:46:39.580 align:middle line:90% have anywhere to model it. 00:46:39.580 --> 00:46:40.090 align:middle line:90% AUDIENCE: That we don't know. 00:46:40.090 --> 00:46:41.410 align:middle line:84% ATUL J. BUTTE: We don't have any way to model it. 00:46:41.410 --> 00:46:44.080 align:middle line:84% AUDIENCE: [INAUDIBLE] when the genome was first 00:46:44.080 --> 00:46:48.580 align:middle line:84% presented in 2000 with [INAUDIBLE],, people [INAUDIBLE] 00:46:48.580 --> 00:46:50.920 align:middle line:84% stopped seeing the genes in junk DNA. 00:46:50.920 --> 00:46:53.950 align:middle line:84% But if you look at the plate of sequences 00:46:53.950 --> 00:46:59.800 align:middle line:84% from [INAUDIBLE] of DNA from [INAUDIBLE] phylogenetic tree, 00:46:59.800 --> 00:47:02.170 align:middle line:84% the percentages are quite similar 00:47:02.170 --> 00:47:04.090 align:middle line:90% for the intergenic region. 00:47:04.090 --> 00:47:06.820 align:middle line:84% That's what the [INAUDIBLE] region tells us. 00:47:06.820 --> 00:47:11.260 align:middle line:84% [INAUDIBLE] speaking, this stuff is important 00:47:11.260 --> 00:47:16.720 align:middle line:84% whether it is to keep selfishly some [INAUDIBLE] to survive 00:47:16.720 --> 00:47:19.000 align:middle line:84% or because it's actually necessary to the organism's 00:47:19.000 --> 00:47:19.973 align:middle line:90% function. 00:47:19.973 --> 00:47:21.890 align:middle line:84% ATUL J. BUTTE: Let me just show this one slide 00:47:21.890 --> 00:47:24.140 align:middle line:84% because you're asking so many questions about repeats. 00:47:24.140 --> 00:47:25.580 align:middle line:84% And this sort of summarizes a lot. 00:47:25.580 --> 00:47:29.238 align:middle line:84% So 3 million repeat copies started as working elements. 00:47:29.238 --> 00:47:30.530 align:middle line:90% That's how they can model this. 00:47:30.530 --> 00:47:33.650 align:middle line:84% There's 3 million repeats in the genome. 00:47:33.650 --> 00:47:39.590 align:middle line:84% So measure how far are they away from a working copy of them. 00:47:39.590 --> 00:47:41.480 align:middle line:84% Because through time, these mutations just 00:47:41.480 --> 00:47:43.910 align:middle line:84% happen randomly presumably at a fixed rate, 00:47:43.910 --> 00:47:46.070 align:middle line:90% but not necessarily so. 00:47:46.070 --> 00:47:49.360 align:middle line:84% Most of the repeats predate the mammalian radiations, 00:47:49.360 --> 00:47:50.990 align:middle line:84% so before mammals were even formed. 00:47:50.990 --> 00:47:52.407 align:middle line:84% Most of these repeats are actually 00:47:52.407 --> 00:47:54.500 align:middle line:90% older than that in the genomes. 00:47:54.500 --> 00:47:56.750 align:middle line:84% But there's no evidence for transposon activity 00:47:56.750 --> 00:47:58.910 align:middle line:84% in the last 50 million years since we've 00:47:58.910 --> 00:48:01.070 align:middle line:90% diverged from monkeys. 00:48:01.070 --> 00:48:03.110 align:middle line:84% None of these things have been alive. 00:48:03.110 --> 00:48:07.010 align:middle line:84% But they're still alive and well in the mouse and the rat. 00:48:07.010 --> 00:48:07.895 align:middle line:90% So not-- go ahead. 00:48:07.895 --> 00:48:08.770 align:middle line:90% AUDIENCE: Is that it? 00:48:08.770 --> 00:48:10.710 align:middle line:84% I mean, what's sort of the breakdown 00:48:10.710 --> 00:48:12.800 align:middle line:90% of how many species are still-- 00:48:12.800 --> 00:48:15.763 align:middle line:84% I mean are monkeys still doing this or they're [INAUDIBLE]?? 00:48:15.763 --> 00:48:17.430 align:middle line:84% ATUL J. BUTTE: I don't think we know yet 00:48:17.430 --> 00:48:19.880 align:middle line:84% because the monkey genomes are just being finished now. 00:48:19.880 --> 00:48:23.515 align:middle line:84% So we can better answer this question every month. 00:48:23.515 --> 00:48:24.890 align:middle line:84% By the end of this course, you'll 00:48:24.890 --> 00:48:27.170 align:middle line:84% be able to answer it better than today because these genomes are 00:48:27.170 --> 00:48:27.773 align:middle line:90% coming online. 00:48:27.773 --> 00:48:30.065 align:middle line:84% Literally, this is just only two years old information. 00:48:30.065 --> 00:48:31.085 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:48:31.085 --> 00:48:32.210 align:middle line:90% ATUL J. BUTTE: What's that? 00:48:32.210 --> 00:48:34.345 align:middle line:90% AUDIENCE: Severely [INAUDIBLE]. 00:48:34.345 --> 00:48:35.720 align:middle line:84% ATUL J. BUTTE: You're never going 00:48:35.720 --> 00:48:37.700 align:middle line:84% to be done sequencing every organism on the planet, 00:48:37.700 --> 00:48:40.033 align:middle line:84% but I think we will have a better estimate now than when 00:48:40.033 --> 00:48:41.960 align:middle line:84% this paper came out if you choose to do this. 00:48:41.960 --> 00:48:44.690 align:middle line:84% Here's the graph that I like to think of. 00:48:44.690 --> 00:48:46.940 align:middle line:84% So the different bars here, the different colors 00:48:46.940 --> 00:48:49.220 align:middle line:84% represent the different repeat elements. 00:48:49.220 --> 00:48:52.520 align:middle line:84% The lines, the signs, they're just different codes. 00:48:52.520 --> 00:48:55.310 align:middle line:84% And here are less than 1% substitution 00:48:55.310 --> 00:48:56.780 align:middle line:90% from a working copy. 00:48:56.780 --> 00:48:59.720 align:middle line:84% And here's 34% substitution from working copy. 00:48:59.720 --> 00:49:01.790 align:middle line:84% And you can read that as millions of years 00:49:01.790 --> 00:49:05.100 align:middle line:84% basically, going back hundreds of millions of years. 00:49:05.100 --> 00:49:08.810 align:middle line:84% And you can see at 7% substitution there's a peak. 00:49:08.810 --> 00:49:11.420 align:middle line:84% And then it drops and then a slight peak, and then it drops. 00:49:11.420 --> 00:49:13.070 align:middle line:90% So these aren't fixed at all. 00:49:13.070 --> 00:49:16.130 align:middle line:84% It makes you wonder what happened on the planet 00:49:16.130 --> 00:49:18.920 align:middle line:90% at this point and in evolution. 00:49:18.920 --> 00:49:20.660 align:middle line:84% Why were all of these things alive, 00:49:20.660 --> 00:49:22.340 align:middle line:84% for example, then all of a sudden 00:49:22.340 --> 00:49:24.740 align:middle line:90% they died off down to zero? 00:49:24.740 --> 00:49:27.650 align:middle line:84% So it's amazing when you think about that only because they 00:49:27.650 --> 00:49:28.970 align:middle line:90% model these as working copies. 00:49:28.970 --> 00:49:31.910 align:middle line:84% And how do they diverge from those working copies? 00:49:31.910 --> 00:49:32.660 align:middle line:90% It's fascinating. 00:49:32.660 --> 00:49:36.530 align:middle line:84% Genomic archaeology, this was not a field three years ago. 00:49:36.530 --> 00:49:39.872 align:middle line:90% 00:49:39.872 --> 00:49:41.330 align:middle line:84% The relevance in medicine still has 00:49:41.330 --> 00:49:43.913 align:middle line:84% to be determined, though, which is why we're going to move on. 00:49:43.913 --> 00:49:49.400 align:middle line:90% 00:49:49.400 --> 00:49:51.318 align:middle line:84% OK, so I've got about 20 minutes left. 00:49:51.318 --> 00:49:53.360 align:middle line:84% Let's see how much of gene measurement techniques 00:49:53.360 --> 00:49:54.100 align:middle line:90% I can cover here. 00:49:54.100 --> 00:49:55.190 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:49:55.190 --> 00:49:56.940 align:middle line:84% ATUL J. BUTTE: I'm going till 2:00, right? 00:49:56.940 --> 00:49:57.650 align:middle line:90% OK. 00:49:57.650 --> 00:49:59.610 align:middle line:90% Let's say 30 minutes. 00:49:59.610 --> 00:50:01.880 align:middle line:84% So the gene measurement techniques I'm going to talk 00:50:01.880 --> 00:50:05.040 align:middle line:84% about today are-- well, one's how do we measure DNA? 00:50:05.040 --> 00:50:06.110 align:middle line:90% How do we measure RNA? 00:50:06.110 --> 00:50:08.455 align:middle line:90% And how do we measure proteins? 00:50:08.455 --> 00:50:09.830 align:middle line:84% If you can get this, you're going 00:50:09.830 --> 00:50:13.560 align:middle line:84% to know 95% of what you need to in this area. 00:50:13.560 --> 00:50:14.700 align:middle line:90% Let's start with DNA. 00:50:14.700 --> 00:50:18.140 align:middle line:84% So DNA sequencing was invented back in the mid-'70s. 00:50:18.140 --> 00:50:21.180 align:middle line:84% And when we measure DNA, we're sequencing it. 00:50:21.180 --> 00:50:24.170 align:middle line:84% So we want to know the sequence of As, Gs, Ts, and Cs. 00:50:24.170 --> 00:50:26.540 align:middle line:84% And so how did we get all 3 and 1/2 billion base pairs-- 00:50:26.540 --> 00:50:29.450 align:middle line:84% basically, just slightly more automated ways than what 00:50:29.450 --> 00:50:31.260 align:middle line:90% I'm going to show you here. 00:50:31.260 --> 00:50:34.430 align:middle line:84% So what you can do is, if you have a DNA sample-- 00:50:34.430 --> 00:50:36.770 align:middle line:84% remember, it's As, Gs, Ts, and Cs-- 00:50:36.770 --> 00:50:40.020 align:middle line:84% we have tools to make copies of them. 00:50:40.020 --> 00:50:42.860 align:middle line:84% In fact, we can borrow the enzymes that the cell uses. 00:50:42.860 --> 00:50:45.530 align:middle line:84% And in a test tube, I can take my As, Gs, Ts, and Cs 00:50:45.530 --> 00:50:47.480 align:middle line:90% and make copies of them. 00:50:47.480 --> 00:50:49.700 align:middle line:84% The process of making copies is a technique 00:50:49.700 --> 00:50:52.160 align:middle line:84% called Polymerase Chain Reaction, or PCR. 00:50:52.160 --> 00:50:55.887 align:middle line:84% All PCR does is make copies of things and amplify them. 00:50:55.887 --> 00:50:57.470 align:middle line:84% If you have one, you make it into two. 00:50:57.470 --> 00:50:58.580 align:middle line:90% Then those two become four. 00:50:58.580 --> 00:50:59.660 align:middle line:90% Four become eight, et cetera. 00:50:59.660 --> 00:51:02.243 align:middle line:84% And it grows to be pretty big, something that you can actually 00:51:02.243 --> 00:51:03.470 align:middle line:90% visualize. 00:51:03.470 --> 00:51:07.550 align:middle line:84% So now, basically the way this technique works is this. 00:51:07.550 --> 00:51:13.550 align:middle line:84% As I'm making a copy of one strand of DNA, 00:51:13.550 --> 00:51:16.200 align:middle line:90% let's stop the reaction. 00:51:16.200 --> 00:51:19.410 align:middle line:84% So as I'm making a copy-- let's say I have 100 base pairs. 00:51:19.410 --> 00:51:20.870 align:middle line:90% Let's stop the reaction. 00:51:20.870 --> 00:51:24.020 align:middle line:84% And I stopped it after I have a whole bunch of copies going. 00:51:24.020 --> 00:51:27.320 align:middle line:84% So some of the copies are going to be 50 when I stopped it. 00:51:27.320 --> 00:51:28.600 align:middle line:90% Some are going to be 99. 00:51:28.600 --> 00:51:29.600 align:middle line:90% Some are going to be 98. 00:51:29.600 --> 00:51:30.600 align:middle line:90% Some are going to be 97. 00:51:30.600 --> 00:51:33.770 align:middle line:84% So I'm going to have a whole variety of lengths of copies 00:51:33.770 --> 00:51:35.830 align:middle line:90% of that gene of that sequence. 00:51:35.830 --> 00:51:37.910 align:middle line:90% It might not even be a gene. 00:51:37.910 --> 00:51:42.295 align:middle line:84% So then let's start it up again, but the last base pair 00:51:42.295 --> 00:51:43.670 align:middle line:84% I'm going to add to that copy I'm 00:51:43.670 --> 00:51:46.430 align:middle line:90% going to make a fluorescent. 00:51:46.430 --> 00:51:49.310 align:middle line:84% And we'll make it fluorescent in such a way that I'm 00:51:49.310 --> 00:51:52.700 align:middle line:84% going to make four colors, one color for A, one for T, 00:51:52.700 --> 00:51:56.540 align:middle line:84% one for C, one for G. So I have four 00:51:56.540 --> 00:51:59.420 align:middle line:84% different fluorescent colors that I could incorporate 00:51:59.420 --> 00:52:02.360 align:middle line:84% into the nucleotide, the last nucleotide I'm going to put 00:52:02.360 --> 00:52:05.440 align:middle line:90% on this particular sequence. 00:52:05.440 --> 00:52:07.190 align:middle line:84% So now, I have a whole variety of lengths, 00:52:07.190 --> 00:52:10.730 align:middle line:84% and the last base pair might be a fluorescent color. 00:52:10.730 --> 00:52:14.210 align:middle line:84% Then what I can do is I can put that whole mix of length 00:52:14.210 --> 00:52:16.460 align:middle line:90% of DNAs and put them in a gel. 00:52:16.460 --> 00:52:19.640 align:middle line:84% I want to actually activate an electric current. 00:52:19.640 --> 00:52:22.460 align:middle line:84% The DNA starts to fall because DNA gets 00:52:22.460 --> 00:52:24.680 align:middle line:90% pulled based on its charge. 00:52:24.680 --> 00:52:27.740 align:middle line:84% But as you would guess, the smaller sequences 00:52:27.740 --> 00:52:29.960 align:middle line:84% get pulled faster because it's easier 00:52:29.960 --> 00:52:31.670 align:middle line:90% to move than the larger ones. 00:52:31.670 --> 00:52:35.570 align:middle line:84% Even by one base pair you can separate them this way. 00:52:35.570 --> 00:52:36.950 align:middle line:90% It's called gel electrophoresis. 00:52:36.950 --> 00:52:39.770 align:middle line:84% That's why a one base pair difference we can see on a gel 00:52:39.770 --> 00:52:42.148 align:middle line:84% depending on the charge and all of this kind of thing. 00:52:42.148 --> 00:52:43.940 align:middle line:84% So then what you do is you put a laser down 00:52:43.940 --> 00:52:45.120 align:middle line:90% at the bottom of the gel. 00:52:45.120 --> 00:52:47.480 align:middle line:84% So it used to be we'd run the gel out. 00:52:47.480 --> 00:52:49.910 align:middle line:84% And if it went off past the end, you'd 00:52:49.910 --> 00:52:52.610 align:middle line:84% swear because you just ruined your sequence. 00:52:52.610 --> 00:52:53.180 align:middle line:90% You lost it. 00:52:53.180 --> 00:52:55.070 align:middle line:84% It's in the puddle at the bottom of the gel. 00:52:55.070 --> 00:52:58.400 align:middle line:84% But now, we want it to happen this way because we set up 00:52:58.400 --> 00:53:01.850 align:middle line:84% a laser at the end that excites whatever fluorescent 00:53:01.850 --> 00:53:03.740 align:middle line:90% is coming down there. 00:53:03.740 --> 00:53:06.980 align:middle line:84% And depending on the color, we can just read it off. 00:53:06.980 --> 00:53:08.090 align:middle line:90% Oh, here's a red peak. 00:53:08.090 --> 00:53:08.990 align:middle line:90% Here's a blue peak. 00:53:08.990 --> 00:53:12.230 align:middle line:90% Here's a green peak. 00:53:12.230 --> 00:53:14.090 align:middle line:84% So we have a whole bunch of these peaks, 00:53:14.090 --> 00:53:16.100 align:middle line:90% and then software looks at this. 00:53:16.100 --> 00:53:18.560 align:middle line:84% In fact, the most commonly used software for this 00:53:18.560 --> 00:53:21.410 align:middle line:84% is something called phrap and phred, which is an open source, 00:53:21.410 --> 00:53:22.520 align:middle line:90% freely available program. 00:53:22.520 --> 00:53:23.937 align:middle line:84% People don't even use the software 00:53:23.937 --> 00:53:24.980 align:middle line:90% that came with the thing. 00:53:24.980 --> 00:53:27.350 align:middle line:84% They wrote their own, and it's open source. 00:53:27.350 --> 00:53:29.990 align:middle line:84% It looks at the peaks, and then makes the call. 00:53:29.990 --> 00:53:31.100 align:middle line:90% I see this peak here. 00:53:31.100 --> 00:53:32.780 align:middle line:84% I think it's a G. I see this peak here. 00:53:32.780 --> 00:53:35.810 align:middle line:84% I think it's a C. There's going to be 00:53:35.810 --> 00:53:37.610 align:middle line:90% immediate problems with this. 00:53:37.610 --> 00:53:41.330 align:middle line:84% First, despite how fancy this looks, 00:53:41.330 --> 00:53:45.300 align:middle line:84% we still can't do more than 300 base pairs with this. 00:53:45.300 --> 00:53:48.170 align:middle line:84% AUDIENCE: [INAUDIBLE] some flaws related 00:53:48.170 --> 00:53:50.132 align:middle line:90% to the size of the [INAUDIBLE]. 00:53:50.132 --> 00:53:51.590 align:middle line:84% ATUL J. BUTTE: They all, in theory, 00:53:51.590 --> 00:53:55.280 align:middle line:84% should be just one base pair apart, 00:53:55.280 --> 00:53:57.240 align:middle line:84% but there's all sorts of reasons. 00:53:57.240 --> 00:53:58.700 align:middle line:84% So why are they some jumbled here 00:53:58.700 --> 00:54:00.170 align:middle line:84% and some they're spaced out here? 00:54:00.170 --> 00:54:02.360 align:middle line:90% As you let the thing go-- 00:54:02.360 --> 00:54:04.247 align:middle line:84% it's not a linear thing, the spacing. 00:54:04.247 --> 00:54:06.080 align:middle line:84% So as you let it go, it gets wider and wider 00:54:06.080 --> 00:54:07.260 align:middle line:90% and wider in general. 00:54:07.260 --> 00:54:11.780 align:middle line:84% But sometimes the copying process itself can have errors. 00:54:11.780 --> 00:54:12.690 align:middle line:90% Let me be clear. 00:54:12.690 --> 00:54:14.150 align:middle line:84% If the sequence I'm starting with 00:54:14.150 --> 00:54:16.220 align:middle line:84% has a whole bunch of Gs in a row, 00:54:16.220 --> 00:54:17.840 align:middle line:90% sometimes it gets confused. 00:54:17.840 --> 00:54:20.840 align:middle line:84% And it just happens to make one the wrong length. 00:54:20.840 --> 00:54:23.750 align:middle line:84% So if I have a lot of repeats, in the sequence 00:54:23.750 --> 00:54:26.960 align:middle line:84% I'm trying to sequence, I'm going to have problems. 00:54:26.960 --> 00:54:30.260 align:middle line:84% And I just told you 50% of the genome is repeats. 00:54:30.260 --> 00:54:31.265 align:middle line:90% It's not easy. 00:54:31.265 --> 00:54:32.507 align:middle line:90% 00:54:32.507 --> 00:54:34.340 align:middle line:84% AUDIENCE: Different sequences might actually 00:54:34.340 --> 00:54:36.410 align:middle line:84% have different proficiencies in these reactions. 00:54:36.410 --> 00:54:37.243 align:middle line:90% ATUL J. BUTTE: Yeah. 00:54:37.243 --> 00:54:39.807 align:middle line:84% AUDIENCE: And so that might cause some slight shifts. 00:54:39.807 --> 00:54:41.390 align:middle line:84% ATUL J. BUTTE: But basically, we can't 00:54:41.390 --> 00:54:45.352 align:middle line:84% do more than 300 to 400 base pairs than this, beyond this. 00:54:45.352 --> 00:54:46.560 align:middle line:90% And this is state of the art. 00:54:46.560 --> 00:54:48.490 align:middle line:90% This is the state of the art. 00:54:48.490 --> 00:54:52.870 align:middle line:84% So to go from this to 3 and 1/2 billion base pairs, 00:54:52.870 --> 00:54:55.250 align:middle line:90% you do it 300 at a time. 00:54:55.250 --> 00:54:55.750 align:middle line:90% That's it. 00:54:55.750 --> 00:54:58.030 align:middle line:90% There's no more magic than this. 00:54:58.030 --> 00:54:59.270 align:middle line:90% There's also another problem. 00:54:59.270 --> 00:55:01.660 align:middle line:84% What if I have two peaks in the same spot? 00:55:01.660 --> 00:55:04.690 align:middle line:90% 00:55:04.690 --> 00:55:05.190 align:middle line:90% Remember? 00:55:05.190 --> 00:55:06.930 align:middle line:90% I have two chromosomes. 00:55:06.930 --> 00:55:08.680 align:middle line:84% On one chromosome, it might be one letter. 00:55:08.680 --> 00:55:10.388 align:middle line:84% On the other, it might be another letter. 00:55:10.388 --> 00:55:11.910 align:middle line:84% Happens all the time, 1 in 1,000 it 00:55:11.910 --> 00:55:13.500 align:middle line:84% turns out, which we'll talk about. 00:55:13.500 --> 00:55:14.610 align:middle line:90% That's a polymorphism. 00:55:14.610 --> 00:55:16.182 align:middle line:90% It could be this or that. 00:55:16.182 --> 00:55:18.765 align:middle line:84% So that's how you can read it by looking at the peaks as well. 00:55:18.765 --> 00:55:22.310 align:middle line:90% 00:55:22.310 --> 00:55:23.290 align:middle line:90% That's basically it. 00:55:23.290 --> 00:55:27.725 align:middle line:84% You terminate the chain with the fluorescent nucleotide. 00:55:27.725 --> 00:55:29.350 align:middle line:84% And you just line them up, and you just 00:55:29.350 --> 00:55:32.110 align:middle line:90% read them off at the bottom. 00:55:32.110 --> 00:55:34.032 align:middle line:84% And phred basically makes the calls here. 00:55:34.032 --> 00:55:36.490 align:middle line:84% The software looks at the peaks and makes it the best guess 00:55:36.490 --> 00:55:38.510 align:middle line:90% here. 00:55:38.510 --> 00:55:41.320 align:middle line:84% So all of these peaks, not just the letters-- 00:55:41.320 --> 00:55:45.370 align:middle line:84% I told you the letters would fit on a CD-ROM. 00:55:45.370 --> 00:55:47.230 align:middle line:84% If you wanted to keep the tracings, 00:55:47.230 --> 00:55:50.410 align:middle line:84% you're going to need a hell of a lot of storage, 00:55:50.410 --> 00:55:53.665 align:middle line:84% on the order of hundreds of terabytes for just one human. 00:55:53.665 --> 00:55:55.540 align:middle line:84% But they're all stored, and they're all saved 00:55:55.540 --> 00:55:57.115 align:middle line:90% for the Human Genome Project. 00:55:57.115 --> 00:55:59.740 align:middle line:84% Because people go back and say, well, this peak wasn't so high. 00:55:59.740 --> 00:56:01.540 align:middle line:84% I don't think this is the right letter. 00:56:01.540 --> 00:56:02.950 align:middle line:84% But for the Human Genome Project, 00:56:02.950 --> 00:56:05.158 align:middle line:84% not only do we have all 3 and 1/2 billion base pairs, 00:56:05.158 --> 00:56:07.480 align:middle line:84% we also have an assessment of how good of a call 00:56:07.480 --> 00:56:09.985 align:middle line:84% it was for each letter based on these peaks. 00:56:09.985 --> 00:56:13.600 align:middle line:90% 00:56:13.600 --> 00:56:16.270 align:middle line:84% So how do we get to the entire Human Genome Project? 00:56:16.270 --> 00:56:17.450 align:middle line:90% You automate the process. 00:56:17.450 --> 00:56:19.240 align:middle line:84% So this is a picture of the Genome Center, the Whitehead 00:56:19.240 --> 00:56:20.140 align:middle line:90% Institute. 00:56:20.140 --> 00:56:23.500 align:middle line:84% And basically, it's a whole bunch of robots just going off 00:56:23.500 --> 00:56:25.690 align:middle line:90% setting these things going. 00:56:25.690 --> 00:56:28.390 align:middle line:84% Most of the genome was sequence in just the 12 months 00:56:28.390 --> 00:56:30.400 align:middle line:84% prior to the finish of the genome. 00:56:30.400 --> 00:56:32.710 align:middle line:84% Even though the process had been going on for 10 years, 00:56:32.710 --> 00:56:34.690 align:middle line:84% it takes time to develop these machines, 00:56:34.690 --> 00:56:36.730 align:middle line:84% develop the technology, and more importantly 00:56:36.730 --> 00:56:39.255 align:middle line:84% to develop the strategy, which we'll talk about in a second. 00:56:39.255 --> 00:56:40.630 align:middle line:84% And then the machines were built. 00:56:40.630 --> 00:56:41.830 align:middle line:84% And basically, they could do the whole genome 00:56:41.830 --> 00:56:43.480 align:middle line:90% in 12 months at the end. 00:56:43.480 --> 00:56:46.810 align:middle line:84% You could see the number of genes that were in draft form 00:56:46.810 --> 00:56:50.410 align:middle line:84% grew exponentially until the end even though it started back 00:56:50.410 --> 00:56:52.750 align:middle line:90% in the mid-'90s. 00:56:52.750 --> 00:56:57.220 align:middle line:84% The Whitehead can run 100,000 sequencing reactions 00:56:57.220 --> 00:56:59.290 align:middle line:90% every 12 hours. 00:56:59.290 --> 00:57:01.870 align:middle line:84% Multiply that by 300 base pairs for each, 00:57:01.870 --> 00:57:04.160 align:middle line:84% and you can get on the order of 3 million base pairs. 00:57:04.160 --> 00:57:08.470 align:middle line:84% Let's say if I do 300, that's going to be three-- 00:57:08.470 --> 00:57:12.520 align:middle line:84% 30 million base pairs, let's say, every 12 hours now. 00:57:12.520 --> 00:57:14.770 align:middle line:84% Robots pick all the colonies because these things 00:57:14.770 --> 00:57:16.300 align:middle line:90% are grown up in bacteria. 00:57:16.300 --> 00:57:19.040 align:middle line:84% OK, here we go, 60 million nucleotides per day 00:57:19.040 --> 00:57:21.220 align:middle line:84% is the current estimate of what they can sequence. 00:57:21.220 --> 00:57:24.770 align:middle line:84% And we couldn't do it even five years ago. 00:57:24.770 --> 00:57:27.820 align:middle line:84% So then you have all of these little, little, little pieces 00:57:27.820 --> 00:57:31.090 align:middle line:90% of 300 base pairs. 00:57:31.090 --> 00:57:33.640 align:middle line:84% And somehow, you need to put this entire jigsaw 00:57:33.640 --> 00:57:35.980 align:middle line:90% puzzle back together. 00:57:35.980 --> 00:57:38.690 align:middle line:84% So there's essentially two ways you can do this. 00:57:38.690 --> 00:57:42.220 align:middle line:84% You can start with some strategy and say, well, I'm 00:57:42.220 --> 00:57:44.590 align:middle line:84% just going to take this piece of this chromosome 00:57:44.590 --> 00:57:45.430 align:middle line:90% and sequence that. 00:57:45.430 --> 00:57:46.805 align:middle line:84% Then let's move to the next piece 00:57:46.805 --> 00:57:49.390 align:middle line:84% and sequence that and take that and take that and just 00:57:49.390 --> 00:57:52.030 align:middle line:84% have some structure in mind as to the order you're 00:57:52.030 --> 00:57:54.100 align:middle line:84% going to do it and keep tract of the pieces. 00:57:54.100 --> 00:57:56.530 align:middle line:84% And then how they're going to go back. 00:57:56.530 --> 00:57:58.990 align:middle line:84% The opposite approach is called shotgun sequencing. 00:57:58.990 --> 00:58:01.810 align:middle line:84% And that's basically saying, take the entire genome, 00:58:01.810 --> 00:58:05.530 align:middle line:84% split it up into these 300 base pair fragments, 00:58:05.530 --> 00:58:07.090 align:middle line:90% and split it up again. 00:58:07.090 --> 00:58:08.770 align:middle line:84% Take the whole genome and split up again 00:58:08.770 --> 00:58:12.670 align:middle line:84% into 300 base pair fragments using some other way. 00:58:12.670 --> 00:58:15.230 align:middle line:84% So the place where I'm going to cut it for this method 00:58:15.230 --> 00:58:17.230 align:middle line:84% are going to be different than the ways I cut it 00:58:17.230 --> 00:58:18.130 align:middle line:90% for this method. 00:58:18.130 --> 00:58:21.220 align:middle line:84% Sequence the whole mess and just see what overlaps. 00:58:21.220 --> 00:58:22.930 align:middle line:84% Basically, this piece of this puzzle 00:58:22.930 --> 00:58:24.972 align:middle line:84% seems to go on this piece of that kind of puzzle. 00:58:24.972 --> 00:58:27.670 align:middle line:84% And you basically put the whole pieces together like that. 00:58:27.670 --> 00:58:31.638 align:middle line:84% Now, that assumes a lot of computational power. 00:58:31.638 --> 00:58:33.430 align:middle line:84% You got to keep all these things in memory. 00:58:33.430 --> 00:58:35.860 align:middle line:84% You got to line them up and see which ones work best. 00:58:35.860 --> 00:58:38.688 align:middle line:84% So the best approach in the end was a hybrid. 00:58:38.688 --> 00:58:40.480 align:middle line:84% You don't just put all the pieces together, 00:58:40.480 --> 00:58:43.030 align:middle line:84% but you also keep some idea of the scaffolding, where 00:58:43.030 --> 00:58:45.940 align:middle line:84% you got the pieces from, instead of just doing it randomly 00:58:45.940 --> 00:58:46.720 align:middle line:90% like this. 00:58:46.720 --> 00:58:49.480 align:middle line:84% So like most things in life, the hybrid worked the best. 00:58:49.480 --> 00:58:52.000 align:middle line:90% 00:58:52.000 --> 00:58:53.663 align:middle line:84% So here are all the pieces cut one way. 00:58:53.663 --> 00:58:55.330 align:middle line:84% Here are all the pieces cut another way. 00:58:55.330 --> 00:58:58.085 align:middle line:84% And this is basically saying, well, A1 to A2 has no overlap, 00:58:58.085 --> 00:58:58.960 align:middle line:90% so we're stuck there. 00:58:58.960 --> 00:59:00.500 align:middle line:90% We don't know how they join. 00:59:00.500 --> 00:59:03.240 align:middle line:84% But A2 and A3 might have some overlap. 00:59:03.240 --> 00:59:05.740 align:middle line:84% And there's B here, and then you basically put them together 00:59:05.740 --> 00:59:07.390 align:middle line:90% based on the pieces. 00:59:07.390 --> 00:59:10.420 align:middle line:84% And hopefully, you have some overlaps between them. 00:59:10.420 --> 00:59:13.970 align:middle line:84% This A goes with this A, but we also see B and B here. 00:59:13.970 --> 00:59:15.460 align:middle line:90% Here's B1 fits right here. 00:59:15.460 --> 00:59:17.350 align:middle line:84% And B3 fits like there, et cetera, et cetera. 00:59:17.350 --> 00:59:19.240 align:middle line:84% That's how these things are joined. 00:59:19.240 --> 00:59:23.418 align:middle line:84% AUDIENCE: [INAUDIBLE] you look for overlap [INAUDIBLE]?? 00:59:23.418 --> 00:59:25.960 align:middle line:84% ATUL J. BUTTE: Obviously, you're going to need a whole bunch. 00:59:25.960 --> 00:59:26.700 align:middle line:84% You're going to need a whole bunch. 00:59:26.700 --> 00:59:28.120 align:middle line:90% And so there's a lot of holes. 00:59:28.120 --> 00:59:30.200 align:middle line:84% AUDIENCE: So there's only 300 base pairs [? information. ?] 00:59:30.200 --> 00:59:30.790 align:middle line:90% ATUL J. BUTTE: And again-- 00:59:30.790 --> 00:59:31.707 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:59:31.707 --> 00:59:34.700 align:middle line:84% ATUL J. BUTTE: Again, I told you 50% of the genome is repeats. 00:59:34.700 --> 00:59:36.220 align:middle line:84% So statistically speaking, you're 00:59:36.220 --> 00:59:39.190 align:middle line:84% going to get a whole bunch that look like they fit together. 00:59:39.190 --> 00:59:41.770 align:middle line:84% That's why this doesn't work that easily right 00:59:41.770 --> 00:59:44.800 align:middle line:84% out of the box because you have three small pieces that 00:59:44.800 --> 00:59:46.700 align:middle line:90% look so much like each other. 00:59:46.700 --> 00:59:48.460 align:middle line:84% So that's why hybrid approach works 00:59:48.460 --> 00:59:51.460 align:middle line:84% best if you remember that this piece came from chromosome 1 00:59:51.460 --> 00:59:54.310 align:middle line:84% on this band on this thing, actually you then 00:59:54.310 --> 00:59:57.680 align:middle line:84% can put it somewhere physically in the genome. 00:59:57.680 --> 01:00:00.310 align:middle line:84% So for a human they had 10 4-processor machines, 01:00:00.310 --> 01:00:02.980 align:middle line:84% so 4 gigabytes of RAM each, and a 16-processor 01:00:02.980 --> 01:00:04.210 align:middle line:90% with 64 gigabytes. 01:00:04.210 --> 01:00:07.210 align:middle line:90% That took 10,000 CPU hours. 01:00:07.210 --> 01:00:09.190 align:middle line:84% That's a huge amount of computational power 01:00:09.190 --> 01:00:11.680 align:middle line:84% just to put all the pieces together for the human. 01:00:11.680 --> 01:00:16.020 align:middle line:84% The overlaps take memory, 600 gigabytes of RAM. 01:00:16.020 --> 01:00:18.552 align:middle line:84% And in the end, hybrid of putting it on a physical map 01:00:18.552 --> 01:00:19.260 align:middle line:90% is the way to go. 01:00:19.260 --> 01:00:20.440 align:middle line:90% Let's see if I have that there. 01:00:20.440 --> 01:00:22.398 align:middle line:84% So now, we're at a state where you can actually 01:00:22.398 --> 01:00:25.060 align:middle line:84% go to websites, like this one at the University of California 01:00:25.060 --> 01:00:27.787 align:middle line:84% Santa Cruz, which is probably one of the more popular ones, 01:00:27.787 --> 01:00:29.620 align:middle line:84% and you can just start to browse the genome. 01:00:29.620 --> 01:00:30.720 align:middle line:84% So if you just pick-- you can even 01:00:30.720 --> 01:00:33.053 align:middle line:84% hit a button to take you to a random spot on the genome. 01:00:33.053 --> 01:00:35.310 align:middle line:84% Here's an arbitrary region of chromosome 19. 01:00:35.310 --> 01:00:36.780 align:middle line:84% It shows you where the bands are, 01:00:36.780 --> 01:00:38.640 align:middle line:84% where the GC-rich regions are, what 01:00:38.640 --> 01:00:41.970 align:middle line:84% you would see on a karyotype, where the predicted genes are, 01:00:41.970 --> 01:00:45.060 align:middle line:84% where actually people have seen different pieces of gene 01:00:45.060 --> 01:00:46.150 align:middle line:90% in messenger RNA. 01:00:46.150 --> 01:00:49.350 align:middle line:84% So you can start to tell where the introns and exons are, 01:00:49.350 --> 01:00:51.600 align:middle line:84% where the polymorphisms are, all in one screen. 01:00:51.600 --> 01:00:53.600 align:middle line:84% And there's actually two different commonly used 01:00:53.600 --> 01:00:54.642 align:middle line:90% browsers for this. 01:00:54.642 --> 01:00:56.100 align:middle line:84% And I think the next screen is just 01:00:56.100 --> 01:01:00.690 align:middle line:90% a blow up of the UCSC browser. 01:01:00.690 --> 01:01:04.920 align:middle line:84% This is like a running thing of the GC content here. 01:01:04.920 --> 01:01:06.300 align:middle line:90% So that gets us to this. 01:01:06.300 --> 01:01:08.520 align:middle line:84% So now, I already started to hint at this. 01:01:08.520 --> 01:01:11.070 align:middle line:84% Let's say, off the automated sequencer, 01:01:11.070 --> 01:01:12.932 align:middle line:90% you see peaks like this. 01:01:12.932 --> 01:01:15.390 align:middle line:84% Or here, I see a little bit of a hump here and a little bit 01:01:15.390 --> 01:01:17.100 align:middle line:84% of a hump here or here, here exactly, 01:01:17.100 --> 01:01:19.860 align:middle line:84% two humps that are exactly the same size. 01:01:19.860 --> 01:01:20.910 align:middle line:90% That makes me think-- 01:01:20.910 --> 01:01:22.740 align:middle line:84% I mean, the program puts a little n here 01:01:22.740 --> 01:01:25.037 align:middle line:84% because I don't know which it is, an A or a T. 01:01:25.037 --> 01:01:27.120 align:middle line:84% But what that really could mean is that that could 01:01:27.120 --> 01:01:28.950 align:middle line:84% be a polymorphism, that that human actually 01:01:28.950 --> 01:01:31.290 align:middle line:84% has both letters, one on one chromosome, one 01:01:31.290 --> 01:01:34.360 align:middle line:90% on the other chromosome. 01:01:34.360 --> 01:01:36.400 align:middle line:84% And so if you have a chunk of DNA 01:01:36.400 --> 01:01:38.400 align:middle line:84% that you're interested in finding polymorphisms, 01:01:38.400 --> 01:01:41.520 align:middle line:84% you can get that gene or get that chunk of DNA, sequence 01:01:41.520 --> 01:01:44.370 align:middle line:84% in 100 humans, and just see what the different letters are 01:01:44.370 --> 01:01:45.270 align:middle line:90% in that sequence. 01:01:45.270 --> 01:01:46.445 align:middle line:90% That's one way to do it. 01:01:46.445 --> 01:01:48.570 align:middle line:84% But they found a whole bunch of these polymorphisms 01:01:48.570 --> 01:01:50.940 align:middle line:84% just by doing the Genome Project. 01:01:50.940 --> 01:01:53.880 align:middle line:84% The Genome Project was sequencing 30 or so humans. 01:01:53.880 --> 01:01:56.010 align:middle line:84% And there are going to be differences in 30 people. 01:01:56.010 --> 01:01:59.650 align:middle line:84% So a whole bunch of them were found that way as well. 01:01:59.650 --> 01:02:02.220 align:middle line:84% So you can either do this while sequencing the genome. 01:02:02.220 --> 01:02:04.200 align:middle line:84% You find arbitrary polymorphisms. 01:02:04.200 --> 01:02:05.700 align:middle line:84% Or if I'm interested in this gene 01:02:05.700 --> 01:02:07.200 align:middle line:84% because I think it's involved in asthma, 01:02:07.200 --> 01:02:08.908 align:middle line:84% I want to see if there are polymorphisms, 01:02:08.908 --> 01:02:11.220 align:middle line:84% I can go after that gene and sequence it the same way. 01:02:11.220 --> 01:02:13.887 align:middle line:84% Both ways are going to look like this on an automated sequencer, 01:02:13.887 --> 01:02:14.670 align:middle line:90% two peaks there. 01:02:14.670 --> 01:02:17.420 align:middle line:90% 01:02:17.420 --> 01:02:19.840 align:middle line:84% You can also use microwaves, which we'll skip this one. 01:02:19.840 --> 01:02:23.960 align:middle line:84% But you can use an array to help you find these SNPs as well. 01:02:23.960 --> 01:02:25.950 align:middle line:84% So I think you're going to have electron SNPs. 01:02:25.950 --> 01:02:26.590 align:middle line:90% If someone-- 01:02:26.590 --> 01:02:30.280 align:middle line:84% AUDIENCE: Joel Hirschhorn, who is an expert for Whitehead, 01:02:30.280 --> 01:02:33.830 align:middle line:84% told you about SNPs and problems of interpreting 01:02:33.830 --> 01:02:34.660 align:middle line:90% SNP [INAUDIBLE]. 01:02:34.660 --> 01:02:36.910 align:middle line:84% ATUL J. BUTTE: So I think this is my last slide on DNA 01:02:36.910 --> 01:02:37.810 align:middle line:90% measurements. 01:02:37.810 --> 01:02:39.940 align:middle line:84% And I'm just going to end on clinical uses 01:02:39.940 --> 01:02:42.490 align:middle line:90% since this is genomic medicine. 01:02:42.490 --> 01:02:45.250 align:middle line:84% So again, SNPs are Single Nucleotide Polymorphisms. 01:02:45.250 --> 01:02:48.400 align:middle line:84% At a particular arbitrary spot in that genome, 01:02:48.400 --> 01:02:50.800 align:middle line:84% you have an A. I have a T, let's say. 01:02:50.800 --> 01:02:53.050 align:middle line:84% And it's usually just one thing or another. 01:02:53.050 --> 01:02:55.540 align:middle line:84% Rarely, there's places where there's three combinations. 01:02:55.540 --> 01:02:57.700 align:middle line:90% Usually, it's one or the other. 01:02:57.700 --> 01:03:00.040 align:middle line:84% That happens, on average, rough estimate-- about 1 01:03:00.040 --> 01:03:02.563 align:middle line:84% in 1,000 base pairs can be a polymorphism. 01:03:02.563 --> 01:03:04.480 align:middle line:84% So if you do the math, that's maybe 3 million, 01:03:04.480 --> 01:03:06.700 align:middle line:84% 10 million different spots in the genome 01:03:06.700 --> 01:03:09.380 align:middle line:84% where you could have one letter or another. 01:03:09.380 --> 01:03:12.400 align:middle line:84% And figuring out the association of those letters 01:03:12.400 --> 01:03:15.280 align:middle line:84% to differences in susceptibility to disease 01:03:15.280 --> 01:03:18.010 align:middle line:90% is another Holy Grail now. 01:03:18.010 --> 01:03:19.970 align:middle line:84% If you have diabetes, I don't have diabetes. 01:03:19.970 --> 01:03:20.710 align:middle line:90% You have these letters. 01:03:20.710 --> 01:03:21.310 align:middle line:90% I have these letters. 01:03:21.310 --> 01:03:22.727 align:middle line:84% Maybe those letters have something 01:03:22.727 --> 01:03:24.555 align:middle line:90% to do with it is the basic idea. 01:03:24.555 --> 01:03:25.930 align:middle line:84% Now, here's one particular figure 01:03:25.930 --> 01:03:27.730 align:middle line:84% from an article in The New England Journal of Medicine 01:03:27.730 --> 01:03:29.110 align:middle line:90% just to give you an idea. 01:03:29.110 --> 01:03:31.420 align:middle line:84% There's one particular gene called HLA. 01:03:31.420 --> 01:03:36.670 align:middle line:84% And this is specifically HLA-B. It's a particular HLA gene. 01:03:36.670 --> 01:03:40.300 align:middle line:84% And you can have one particular polymorphism 01:03:40.300 --> 01:03:42.405 align:middle line:90% on both your chromosomes. 01:03:42.405 --> 01:03:44.030 align:middle line:84% You can have a mix of one or the other, 01:03:44.030 --> 01:03:46.280 align:middle line:84% or you can have the most common form of this. 01:03:46.280 --> 01:03:48.740 align:middle line:84% So there's a common form of the letter. 01:03:48.740 --> 01:03:51.920 align:middle line:84% This is a rare form, or we could have a mix of the two. 01:03:51.920 --> 01:03:55.930 align:middle line:84% And this one particular gene, if you're infected with HIV, 01:03:55.930 --> 01:04:00.430 align:middle line:84% that one spelling determines your conversion to AIDS. 01:04:00.430 --> 01:04:02.180 align:middle line:84% It can actually have a huge impact. 01:04:02.180 --> 01:04:06.610 align:middle line:84% So if you have this rare form of the polymorphism on both 01:04:06.610 --> 01:04:07.990 align:middle line:84% your chromosomes, homozygous, you 01:04:07.990 --> 01:04:12.850 align:middle line:84% can go from having HIV to AIDS very quickly within 10 years. 01:04:12.850 --> 01:04:15.020 align:middle line:84% If you don't have it, if both your chromosomes have 01:04:15.020 --> 01:04:17.780 align:middle line:84% the more common form, you can go on and on. 01:04:17.780 --> 01:04:21.710 align:middle line:84% And some people never get AIDS as far out as 18 years. 01:04:21.710 --> 01:04:25.040 align:middle line:84% And if you have the mix, you're somewhere in between here. 01:04:25.040 --> 01:04:27.760 align:middle line:84% So associations of SNPs with diseases 01:04:27.760 --> 01:04:30.247 align:middle line:84% are exceedingly common to find now. 01:04:30.247 --> 01:04:32.080 align:middle line:84% In fact, every other week in The New England 01:04:32.080 --> 01:04:34.122 align:middle line:84% Journal of Medicine, this is what you're reading. 01:04:34.122 --> 01:04:36.550 align:middle line:84% It's going to be some association with this spelling 01:04:36.550 --> 01:04:38.272 align:middle line:90% versus that disease. 01:04:38.272 --> 01:04:40.480 align:middle line:84% AUDIENCE: What's the most frequent form in this case? 01:04:40.480 --> 01:04:41.230 align:middle line:84% ATUL J. BUTTE: The most frequent is going 01:04:41.230 --> 01:04:43.090 align:middle line:90% to be no HLA, in the black one. 01:04:43.090 --> 01:04:44.080 align:middle line:90% Yup. 01:04:44.080 --> 01:04:45.580 align:middle line:90% So let me ask you. 01:04:45.580 --> 01:04:50.140 align:middle line:84% How much do you think it costs to measure a SNP? 01:04:50.140 --> 01:04:53.320 align:middle line:84% So if I have one arbitrary spot in all of our genomes 01:04:53.320 --> 01:04:55.780 align:middle line:84% that I want to measure, whether there's A, 01:04:55.780 --> 01:04:57.760 align:middle line:84% T, C, or G in any arbitrary genome, 01:04:57.760 --> 01:04:59.290 align:middle line:90% how much do you think it costs? 01:04:59.290 --> 01:05:00.880 align:middle line:84% Not how much the lab charges, how much 01:05:00.880 --> 01:05:03.430 align:middle line:84% do you think it costs to measure one base pair? 01:05:03.430 --> 01:05:04.750 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:05:04.750 --> 01:05:05.740 align:middle line:84% ATUL J. BUTTE: How much does it cost? 01:05:05.740 --> 01:05:06.400 align:middle line:90% AUDIENCE: I already [INAUDIBLE]. 01:05:06.400 --> 01:05:07.358 align:middle line:90% ATUL J. BUTTE: Guesses? 01:05:07.358 --> 01:05:08.620 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:05:08.620 --> 01:05:10.000 align:middle line:84% ATUL J. BUTTE: You already guess? 01:05:10.000 --> 01:05:10.500 align:middle line:90% Yeah. 01:05:10.500 --> 01:05:13.450 align:middle line:84% So we come up with the answer of $0.05 to $0.10. 01:05:13.450 --> 01:05:15.683 align:middle line:84% It's probably going to $0.01 in the next few years, 01:05:15.683 --> 01:05:16.225 align:middle line:90% pretty cheap. 01:05:16.225 --> 01:05:17.050 align:middle line:84% AUDIENCE: Stole your thunder, sorry. 01:05:17.050 --> 01:05:19.030 align:middle line:84% ATUL J. BUTTE: Stole my thunder, all right. 01:05:19.030 --> 01:05:21.070 align:middle line:90% Oh, yeah. 01:05:21.070 --> 01:05:23.290 align:middle line:84% Last thing on SNPs, you didn't take this, did you? 01:05:23.290 --> 01:05:23.950 align:middle line:90% AUDIENCE: I did, too. 01:05:23.950 --> 01:05:26.117 align:middle line:84% ATUL J. BUTTE: Damn, I'll stop giving you my slides. 01:05:26.117 --> 01:05:30.843 align:middle line:90% 01:05:30.843 --> 01:05:31.760 align:middle line:90% How much time we have? 01:05:31.760 --> 01:05:33.560 align:middle line:90% OK. 01:05:33.560 --> 01:05:37.108 align:middle line:84% Let's talk about gene expression. 01:05:37.108 --> 01:05:38.900 align:middle line:84% Let's talk about microarrays, fourth slide, 01:05:38.900 --> 01:05:40.700 align:middle line:84% and then we'll talk about protein. 01:05:40.700 --> 01:05:42.080 align:middle line:90% So we recovered DNA. 01:05:42.080 --> 01:05:44.270 align:middle line:84% Now, we're working our way out of the nucleus. 01:05:44.270 --> 01:05:46.950 align:middle line:84% Now, let's talk about the messenger RNA. 01:05:46.950 --> 01:05:48.650 align:middle line:84% So like we said, if the cell decides 01:05:48.650 --> 01:05:49.850 align:middle line:84% to make a particular protein, it's 01:05:49.850 --> 01:05:51.920 align:middle line:84% going to start making the blueprints for that protein, 01:05:51.920 --> 01:05:52.705 align:middle line:90% the messenger RNA. 01:05:52.705 --> 01:05:56.840 align:middle line:84% It turns out it's very easy to measure these now. 01:05:56.840 --> 01:05:59.570 align:middle line:84% The most common way to measure them is using a microarray. 01:05:59.570 --> 01:06:02.330 align:middle line:84% All a microarray is is a 2 centimeter 01:06:02.330 --> 01:06:04.670 align:middle line:84% by 2 centimeter, about the size of your thumbnail, 01:06:04.670 --> 01:06:08.120 align:middle line:84% it's a man-made grid of DNA where each spot in the grid 01:06:08.120 --> 01:06:10.910 align:middle line:84% is looking for one particular RNA. 01:06:10.910 --> 01:06:14.138 align:middle line:84% And so the density is about 500 by 500. 01:06:14.138 --> 01:06:16.430 align:middle line:84% So that's on order of a quarter million different spots 01:06:16.430 --> 01:06:17.457 align:middle line:90% on these arrays. 01:06:17.457 --> 01:06:19.040 align:middle line:84% And suffice it to say, you can measure 01:06:19.040 --> 01:06:23.720 align:middle line:84% the entire genome of transcripts with one array today. 01:06:23.720 --> 01:06:25.787 align:middle line:84% And the way this works is have to get the tissue 01:06:25.787 --> 01:06:26.870 align:middle line:90% that you're interested in. 01:06:26.870 --> 01:06:28.520 align:middle line:84% You can't just do this from blood 01:06:28.520 --> 01:06:30.253 align:middle line:84% unless you're interested in blood. 01:06:30.253 --> 01:06:31.670 align:middle line:84% If you're interested in the brain, 01:06:31.670 --> 01:06:33.050 align:middle line:84% you've got to get the brain tissue to figure out 01:06:33.050 --> 01:06:34.342 align:middle line:90% what genes are turned on there. 01:06:34.342 --> 01:06:36.380 align:middle line:84% It might expose it to certain influence. 01:06:36.380 --> 01:06:38.263 align:middle line:90% Get the RNA, so isolate it. 01:06:38.263 --> 01:06:40.430 align:middle line:84% Remember, most of them are sticky with the little A. 01:06:40.430 --> 01:06:42.830 align:middle line:84% We just have to get the Ts and pull them out. 01:06:42.830 --> 01:06:46.370 align:middle line:84% Make them fluorescent by making copies of them. 01:06:46.370 --> 01:06:48.140 align:middle line:84% Remember, I showed you that technique. 01:06:48.140 --> 01:06:50.082 align:middle line:84% You can make a copy of something using the PCR 01:06:50.082 --> 01:06:52.040 align:middle line:84% and just have it put fluorescent ones in there, 01:06:52.040 --> 01:06:53.248 align:middle line:90% and it'll become fluorescent. 01:06:53.248 --> 01:06:55.527 align:middle line:84% Let it sit on this man-made grid overnight. 01:06:55.527 --> 01:06:56.360 align:middle line:90% Come in the morning. 01:06:56.360 --> 01:06:57.140 align:middle line:90% Wash it off. 01:06:57.140 --> 01:06:58.478 align:middle line:90% Light it up with the laser. 01:06:58.478 --> 01:07:00.020 align:middle line:84% The laser lights up the fluorescence, 01:07:00.020 --> 01:07:01.710 align:middle line:84% and you basically take a picture. 01:07:01.710 --> 01:07:03.080 align:middle line:90% And you get a TIFF image. 01:07:03.080 --> 01:07:06.140 align:middle line:84% If you scan in pictures at home, you know what a TIFF image is. 01:07:06.140 --> 01:07:07.170 align:middle line:90% But these are huge. 01:07:07.170 --> 01:07:09.650 align:middle line:84% These are on order of 10 megabytes, these files. 01:07:09.650 --> 01:07:11.402 align:middle line:84% And then the biology takes a pause. 01:07:11.402 --> 01:07:12.860 align:middle line:84% The informatics comes in and starts 01:07:12.860 --> 01:07:14.510 align:middle line:90% to quantitate each of these. 01:07:14.510 --> 01:07:17.870 align:middle line:84% It's quantitative to 4 to 5 orders of magnitude, 01:07:17.870 --> 01:07:19.410 align:middle line:90% and it's not perfect. 01:07:19.410 --> 01:07:21.697 align:middle line:90% There's a lot of noise. 01:07:21.697 --> 01:07:23.030 align:middle line:90% Most people like this technique. 01:07:23.030 --> 01:07:25.070 align:middle line:84% Again, these are also commodity items today. 01:07:25.070 --> 01:07:27.800 align:middle line:84% You can get one of these for the entire genome for about $300, 01:07:27.800 --> 01:07:31.400 align:middle line:84% $350 bucks today on [? Longwood. ?] 01:07:31.400 --> 01:07:33.620 align:middle line:84% The quantitative, there's at least two different ways 01:07:33.620 --> 01:07:34.220 align:middle line:90% to do this. 01:07:34.220 --> 01:07:36.320 align:middle line:84% There's an absolute measurement technique 01:07:36.320 --> 01:07:37.560 align:middle line:90% and a relative technique. 01:07:37.560 --> 01:07:38.518 align:middle line:90% So what does that mean? 01:07:38.518 --> 01:07:41.390 align:middle line:84% How much of this gene is present is one technique. 01:07:41.390 --> 01:07:44.090 align:middle line:84% How much is this gene is present relative to this other tissue 01:07:44.090 --> 01:07:45.020 align:middle line:90% is another technique. 01:07:45.020 --> 01:07:47.657 align:middle line:84% And the relative one uses two fluorescent colors. 01:07:47.657 --> 01:07:49.490 align:middle line:84% The absolute one uses one fluorescent color. 01:07:49.490 --> 01:07:51.290 align:middle line:84% It depends on what's the nature of the spots are, 01:07:51.290 --> 01:07:53.123 align:middle line:84% but we're not going to go into those details 01:07:53.123 --> 01:07:54.800 align:middle line:90% unless you really want to. 01:07:54.800 --> 01:07:58.580 align:middle line:84% The important point is that the genes are chosen arbitrarily. 01:07:58.580 --> 01:08:00.560 align:middle line:84% Statisticians hate this aspect of it. 01:08:00.560 --> 01:08:03.500 align:middle line:84% These are not drawn from some master normal distribution 01:08:03.500 --> 01:08:04.310 align:middle line:90% here. 01:08:04.310 --> 01:08:07.010 align:middle line:84% This year's chips are different than last year's chips 01:08:07.010 --> 01:08:08.560 align:middle line:84% because we know more of the genes. 01:08:08.560 --> 01:08:11.060 align:middle line:84% Next year's chips are going to be different than this year's 01:08:11.060 --> 01:08:11.540 align:middle line:90% chips. 01:08:11.540 --> 01:08:13.248 align:middle line:84% So every year, these things are changing. 01:08:13.248 --> 01:08:15.620 align:middle line:84% There's not ever going to be one fixed genome here. 01:08:15.620 --> 01:08:17.870 align:middle line:84% And you need the functional tissue. 01:08:17.870 --> 01:08:20.180 align:middle line:84% There are some diseases that are easy to study 01:08:20.180 --> 01:08:21.920 align:middle line:90% with microarrays, like cancer. 01:08:21.920 --> 01:08:23.615 align:middle line:84% If you have a solid tumor, the patient 01:08:23.615 --> 01:08:24.740 align:middle line:90% goes to the operating room. 01:08:24.740 --> 01:08:25.790 align:middle line:90% The tumor gets taken out. 01:08:25.790 --> 01:08:26.873 align:middle line:90% You have a hunk of tissue. 01:08:26.873 --> 01:08:27.930 align:middle line:90% You can extract the RNA. 01:08:27.930 --> 01:08:29.502 align:middle line:84% You can find a lot of genes this way. 01:08:29.502 --> 01:08:31.460 align:middle line:84% And so that's why most of the microarray papers 01:08:31.460 --> 01:08:32.959 align:middle line:90% involve cancer. 01:08:32.959 --> 01:08:35.000 align:middle line:84% A slightly harder disease to study 01:08:35.000 --> 01:08:37.160 align:middle line:84% might be something like diabetes. 01:08:37.160 --> 01:08:40.880 align:middle line:84% Diabetes affects the liver, muscle, fat, pancreas, maybe 01:08:40.880 --> 01:08:42.178 align:middle line:90% brain simultaneously. 01:08:42.178 --> 01:08:43.970 align:middle line:84% You might want to get all of those tissues. 01:08:43.970 --> 01:08:45.740 align:middle line:84% And let me tell you, we don't normally 01:08:45.740 --> 01:08:48.590 align:middle line:84% do biopsies of these things when we diagnose diabetes. 01:08:48.590 --> 01:08:51.529 align:middle line:84% Well, we could use animal models, et cetera. 01:08:51.529 --> 01:08:53.330 align:middle line:84% Slightly harder disease to study would be 01:08:53.330 --> 01:08:54.680 align:middle line:90% something like schizophrenia. 01:08:54.680 --> 01:08:56.763 align:middle line:84% I know the functional tissue, but just 01:08:56.763 --> 01:08:59.180 align:middle line:84% which part of the brain am I going to actually get and put 01:08:59.180 --> 01:08:59.960 align:middle line:90% on this array? 01:08:59.960 --> 01:09:02.240 align:middle line:84% I can't take an entire human brain to do this. 01:09:02.240 --> 01:09:03.537 align:middle line:90% Is it this part or this part? 01:09:03.537 --> 01:09:05.120 align:middle line:84% So I even know the tissue, but I don't 01:09:05.120 --> 01:09:06.859 align:middle line:84% know where I'm going to put the array, what 01:09:06.859 --> 01:09:08.120 align:middle line:90% I'm going to put on the array. 01:09:08.120 --> 01:09:10.010 align:middle line:84% And then one harder disease to study 01:09:10.010 --> 01:09:11.870 align:middle line:84% would be something like hypertension. 01:09:11.870 --> 01:09:15.529 align:middle line:84% What's the functional tissue in high blood pressure? 01:09:15.529 --> 01:09:16.372 align:middle line:90% Is it the heart? 01:09:16.372 --> 01:09:17.330 align:middle line:90% Is it the blood vessel? 01:09:17.330 --> 01:09:19.620 align:middle line:84% Is it the smooth muscle of the blood vessel? 01:09:19.620 --> 01:09:21.979 align:middle line:84% What exactly am I going to put on the array 01:09:21.979 --> 01:09:23.819 align:middle line:84% to learn what genes are up and down? 01:09:23.819 --> 01:09:26.415 align:middle line:84% So this is not the panacea for all diseases. 01:09:26.415 --> 01:09:28.790 align:middle line:84% If you want to know what genes are functionally involved, 01:09:28.790 --> 01:09:31.130 align:middle line:84% you have to be able to think about the functional tissue 01:09:31.130 --> 01:09:32.300 align:middle line:90% here. 01:09:32.300 --> 01:09:34.100 align:middle line:90% Yes. 01:09:34.100 --> 01:09:36.920 align:middle line:84% AUDIENCE: How long after death can you do this [INAUDIBLE]?? 01:09:36.920 --> 01:09:37.760 align:middle line:84% ATUL J. BUTTE: The easy way to answer 01:09:37.760 --> 01:09:39.635 align:middle line:84% that is that people have been able to get RNA 01:09:39.635 --> 01:09:42.524 align:middle line:90% from paraffin-embedded tissues. 01:09:42.524 --> 01:09:44.149 align:middle line:84% And they think that they're successful, 01:09:44.149 --> 01:09:45.950 align:middle line:84% but most people flash freeze these things. 01:09:45.950 --> 01:09:47.810 align:middle line:84% If you're in the OR with the tumor getting taken out, 01:09:47.810 --> 01:09:49.760 align:middle line:84% most people just put it into [? monosodium ?] 01:09:49.760 --> 01:09:50.330 align:middle line:90% right then and there. 01:09:50.330 --> 01:09:51.830 align:middle line:84% AUDIENCE: The slightly longer answer 01:09:51.830 --> 01:09:54.680 align:middle line:84% is RNA does degrade as opposed to DNA, which can stick around. 01:09:54.680 --> 01:09:59.350 align:middle line:84% And that's why they can do Jurassic Park-like stunts. 01:09:59.350 --> 01:10:01.880 align:middle line:84% ATUL J. BUTTE: But there's a reason why RNA degrades, 01:10:01.880 --> 01:10:03.650 align:middle line:90% because we degrade it. 01:10:03.650 --> 01:10:06.090 align:middle line:84% A lot of viruses are just strands of RNA. 01:10:06.090 --> 01:10:07.760 align:middle line:84% So that's on our interest to degrade it. 01:10:07.760 --> 01:10:10.640 align:middle line:84% So we have RNA, so we exude RNAses. 01:10:10.640 --> 01:10:11.930 align:middle line:90% It's much harder to work-- 01:10:11.930 --> 01:10:13.983 align:middle line:84% I mean, high school students work with DNA. 01:10:13.983 --> 01:10:16.400 align:middle line:84% If you've ever done this in your high school where they go 01:10:16.400 --> 01:10:19.210 align:middle line:84% get blood from an abattoir-- and they just 01:10:19.210 --> 01:10:20.210 align:middle line:90% basically get the blood. 01:10:20.210 --> 01:10:21.835 align:middle line:84% And you can look at the strands of DNA. 01:10:21.835 --> 01:10:24.200 align:middle line:84% You can't do that with RNA because we all have RNAses. 01:10:24.200 --> 01:10:25.895 align:middle line:84% We're going to degrade it immediately. 01:10:25.895 --> 01:10:27.020 align:middle line:90% So you have to wear gloves. 01:10:27.020 --> 01:10:28.460 align:middle line:84% You have to use special techniques, et cetera. 01:10:28.460 --> 01:10:30.418 align:middle line:84% AUDIENCE: But sometimes studies [INAUDIBLE] you 01:10:30.418 --> 01:10:31.790 align:middle line:90% have hours after death. 01:10:31.790 --> 01:10:36.130 align:middle line:84% And definitely, a muscle tissue four hours after death 01:10:36.130 --> 01:10:40.950 align:middle line:84% has a different [INAUDIBLE] factor [INAUDIBLE],, 01:10:40.950 --> 01:10:44.140 align:middle line:84% for instance, [INAUDIBLE] in many other [INAUDIBLE].. 01:10:44.140 --> 01:10:47.490 align:middle line:84% And so if you really want to do it reasonably, 01:10:47.490 --> 01:10:50.040 align:middle line:84% then basically as close to get the moment of death 01:10:50.040 --> 01:10:51.558 align:middle line:90% is where you want to be. 01:10:51.558 --> 01:10:53.600 align:middle line:84% ATUL J. BUTTE: This is what these arrays actually 01:10:53.600 --> 01:10:54.030 align:middle line:90% look like. 01:10:54.030 --> 01:10:55.550 align:middle line:84% This is the whole human genome array 01:10:55.550 --> 01:10:57.980 align:middle line:84% from a company called Affymetrix that makes these things. 01:10:57.980 --> 01:11:00.540 align:middle line:84% And they actually make a 96-well format of this. 01:11:00.540 --> 01:11:03.287 align:middle line:84% So here are 96 microarrays in just one plate. 01:11:03.287 --> 01:11:05.120 align:middle line:84% So you can imagine how much data you get out 01:11:05.120 --> 01:11:08.180 align:middle line:84% of this with 10 megabytes of each for each of these images, 01:11:08.180 --> 01:11:11.080 align:middle line:84% huge amount of data being collected this way. 01:11:11.080 --> 01:11:13.490 align:middle line:84% And we're not going to talk about the differences here. 01:11:13.490 --> 01:11:15.960 align:middle line:90% And we'll not talk about that. 01:11:15.960 --> 01:11:18.535 align:middle line:84% So you can use these microarrays to tell which genes are up. 01:11:18.535 --> 01:11:20.660 align:middle line:84% So for example, if I have a whole bunch of patients 01:11:20.660 --> 01:11:23.660 align:middle line:84% and half of them had one type of leukemia 01:11:23.660 --> 01:11:25.668 align:middle line:84% and half had another type of leukemia, 01:11:25.668 --> 01:11:27.710 align:middle line:84% leukemia means a whole bunch of white blood cells 01:11:27.710 --> 01:11:28.680 align:middle line:90% go up in the blood. 01:11:28.680 --> 01:11:31.057 align:middle line:84% So you get easy access to that tissue. 01:11:31.057 --> 01:11:33.140 align:middle line:84% And you can find genes that are up in one and down 01:11:33.140 --> 01:11:36.020 align:middle line:84% and the other or down in one and up in the other, et cetera, et 01:11:36.020 --> 01:11:36.770 align:middle line:90% cetera. 01:11:36.770 --> 01:11:38.930 align:middle line:84% In the end, you can get lists of genes very easily. 01:11:38.930 --> 01:11:41.420 align:middle line:90% Now, we all commonly do this. 01:11:41.420 --> 01:11:43.790 align:middle line:84% And this core facility is basically in every hospital. 01:11:43.790 --> 01:11:46.070 align:middle line:84% If you're a clinician at the Brigham and Women's Hospital, 01:11:46.070 --> 01:11:48.020 align:middle line:84% you can take your sample down to the core facility, 01:11:48.020 --> 01:11:49.520 align:middle line:84% and they'll give you back the text 01:11:49.520 --> 01:11:52.080 align:middle line:84% file-- this gene, this amount, this gene, this amount. 01:11:52.080 --> 01:11:54.890 align:middle line:84% It's just a matter of asking the right questions now. 01:11:54.890 --> 01:11:56.900 align:middle line:84% But the list of genes isn't enough. 01:11:56.900 --> 01:11:58.730 align:middle line:84% Sometimes you want to go back and validate 01:11:58.730 --> 01:12:00.772 align:middle line:84% that a gene is actually where it's supposed to be 01:12:00.772 --> 01:12:02.610 align:middle line:84% or what you thought was actually happening. 01:12:02.610 --> 01:12:05.260 align:middle line:84% And there's two ways to do this, in situ hybridization 01:12:05.260 --> 01:12:07.400 align:middle line:90% and real-time PCR. 01:12:07.400 --> 01:12:09.500 align:middle line:84% So these are usually inserted after you've 01:12:09.500 --> 01:12:11.360 align:middle line:90% done a microarray analysis. 01:12:11.360 --> 01:12:13.310 align:middle line:84% In situ hybridization is very easy. 01:12:13.310 --> 01:12:15.770 align:middle line:84% I take the same RNA that I'm looking for, 01:12:15.770 --> 01:12:18.020 align:middle line:84% make the exact reverse complement of it, 01:12:18.020 --> 01:12:19.820 align:middle line:90% and make it fluorescent again. 01:12:19.820 --> 01:12:21.890 align:middle line:84% And I'll make enough of it that I can then 01:12:21.890 --> 01:12:25.350 align:middle line:84% stain a sample, a tissue sample, with that 01:12:25.350 --> 01:12:27.230 align:middle line:84% and then light it up to make a picture. 01:12:27.230 --> 01:12:29.300 align:middle line:84% So wherever that transcript was, this 01:12:29.300 --> 01:12:31.475 align:middle line:84% is going to stick to it after I've washed it off. 01:12:31.475 --> 01:12:32.600 align:middle line:90% And it's going to light up. 01:12:32.600 --> 01:12:34.642 align:middle line:84% And I can take glorious pictures like this. 01:12:34.642 --> 01:12:36.350 align:middle line:84% So I can use different colors to light up 01:12:36.350 --> 01:12:37.725 align:middle line:90% one RNA versus another. 01:12:37.725 --> 01:12:40.100 align:middle line:84% And that might help me tell that it's in the right tissue 01:12:40.100 --> 01:12:41.720 align:middle line:90% that I thought it was. 01:12:41.720 --> 01:12:45.715 align:middle line:90% Real-time PCR, it's even-- 01:12:45.715 --> 01:12:47.840 align:middle line:84% it's actually a little bit harder to conceptualize, 01:12:47.840 --> 01:12:52.010 align:middle line:84% but basically real-time PCR uses that polymerase chain reaction 01:12:52.010 --> 01:12:57.200 align:middle line:84% technique to actually detect one sequence as compared 01:12:57.200 --> 01:12:59.322 align:middle line:90% to a control sequence. 01:12:59.322 --> 01:13:01.280 align:middle line:84% So it gives me a more quantitative measurement. 01:13:01.280 --> 01:13:02.750 align:middle line:84% This gives me a nice picture that I 01:13:02.750 --> 01:13:05.930 align:middle line:84% can put on the cover of Nature, but RT PCR actually 01:13:05.930 --> 01:13:07.940 align:middle line:84% gives me a number that actually can then 01:13:07.940 --> 01:13:10.130 align:middle line:84% be used to validate the microarray findings. 01:13:10.130 --> 01:13:11.270 align:middle line:90% It was three times higher. 01:13:11.270 --> 01:13:12.950 align:middle line:90% It was four times higher. 01:13:12.950 --> 01:13:14.450 align:middle line:84% It's basically using a PCR technique 01:13:14.450 --> 01:13:17.660 align:middle line:84% to see how fast do I get a detectable amount using 01:13:17.660 --> 01:13:21.200 align:middle line:84% the repetitive the exponential growth aspects of PCR. 01:13:21.200 --> 01:13:24.460 align:middle line:90% 01:13:24.460 --> 01:13:27.280 align:middle line:84% So now, let's get into proteins because we're coming close 01:13:27.280 --> 01:13:29.000 align:middle line:90% to the end of the hour here. 01:13:29.000 --> 01:13:31.120 align:middle line:84% So again, the Holy Grail is get to proteins. 01:13:31.120 --> 01:13:34.570 align:middle line:84% We all essentially use microarrays 01:13:34.570 --> 01:13:39.843 align:middle line:84% as a proxy for proteins, but we don't have an array. 01:13:39.843 --> 01:13:42.010 align:middle line:84% The same kind of array that we have for microarrays, 01:13:42.010 --> 01:13:43.840 align:middle line:84% we just don't have it for proteins. 01:13:43.840 --> 01:13:45.280 align:middle line:90% Why? 01:13:45.280 --> 01:13:49.120 align:middle line:84% Because proteins are much harder to measure. 01:13:49.120 --> 01:13:52.000 align:middle line:84% The RNA, basically, is As, Ts, Cs, Gs. 01:13:52.000 --> 01:13:53.350 align:middle line:90% And it's essentially a strand. 01:13:53.350 --> 01:13:55.630 align:middle line:84% And sometimes it buckles and has twists, 01:13:55.630 --> 01:13:58.780 align:middle line:84% but it's pretty easy to measure using these microarrays. 01:13:58.780 --> 01:14:02.230 align:middle line:84% Proteins can be positively charged, negatively charged. 01:14:02.230 --> 01:14:03.580 align:middle line:90% They can love water. 01:14:03.580 --> 01:14:04.840 align:middle line:90% They can hate water. 01:14:04.840 --> 01:14:05.840 align:middle line:90% They can be big. 01:14:05.840 --> 01:14:06.740 align:middle line:90% They can be small. 01:14:06.740 --> 01:14:08.410 align:middle line:84% If you look at it this way, it looks like one thing. 01:14:08.410 --> 01:14:10.900 align:middle line:84% If you look at it this way, it looks like another thing. 01:14:10.900 --> 01:14:13.630 align:middle line:84% So there's no easy consistent way 01:14:13.630 --> 01:14:17.200 align:middle line:84% to measure proteins like there is for RNA. 01:14:17.200 --> 01:14:20.290 align:middle line:84% Just to give you an idea, here is a commonly used technique 01:14:20.290 --> 01:14:25.840 align:middle line:84% called 2D Polyacrylamide Gel Electrophoresis, or 2D-PAGE. 01:14:25.840 --> 01:14:29.530 align:middle line:84% Take one sample, a gemish of proteins. 01:14:29.530 --> 01:14:31.330 align:middle line:84% All the proteins are radioactive, 01:14:31.330 --> 01:14:34.420 align:middle line:84% and I'm going to spread them out this way based on the size. 01:14:34.420 --> 01:14:37.390 align:middle line:84% Remember, if I put a charge, the smaller things 01:14:37.390 --> 01:14:39.050 align:middle line:90% move down faster. 01:14:39.050 --> 01:14:40.430 align:middle line:90% So here's size. 01:14:40.430 --> 01:14:41.770 align:middle line:90% And here's pH. 01:14:41.770 --> 01:14:43.450 align:middle line:84% Here's the pH that they like to hang out 01:14:43.450 --> 01:14:46.700 align:middle line:84% in, two different things, two different aspects of proteins. 01:14:46.700 --> 01:14:50.990 align:middle line:84% And in the one gemish, I have all of these spots. 01:14:50.990 --> 01:14:55.640 align:middle line:84% And if I do this again, they're going to be slightly different. 01:14:55.640 --> 01:14:58.480 align:middle line:84% If I do a third time, it'll be slightly different. 01:14:58.480 --> 01:14:59.470 align:middle line:90% This is a picture. 01:14:59.470 --> 01:15:02.410 align:middle line:84% This is the most irreducible thing you can imagine. 01:15:02.410 --> 01:15:04.570 align:middle line:90% The gel is set just right. 01:15:04.570 --> 01:15:06.530 align:middle line:84% It might just come out a different way. 01:15:06.530 --> 01:15:08.530 align:middle line:84% So we wish that we could look at this 01:15:08.530 --> 01:15:10.690 align:middle line:84% and say, oh, it's got this size and this pH. 01:15:10.690 --> 01:15:12.160 align:middle line:90% It's got to be this protein-- 01:15:12.160 --> 01:15:13.790 align:middle line:90% can't do that at all. 01:15:13.790 --> 01:15:18.160 align:middle line:84% So then what we'll do is we'll cut out one of these spots 01:15:18.160 --> 01:15:20.110 align:middle line:84% and use the next technique to figure out 01:15:20.110 --> 01:15:21.880 align:middle line:90% what protein might be in there. 01:15:21.880 --> 01:15:24.160 align:middle line:84% But literally, people look at these pictures 01:15:24.160 --> 01:15:26.590 align:middle line:84% and they say, well, here's a picture from cancer. 01:15:26.590 --> 01:15:27.910 align:middle line:90% Here's not cancer. 01:15:27.910 --> 01:15:30.160 align:middle line:84% And oh, I see this spot here, and I don't see it here. 01:15:30.160 --> 01:15:32.980 align:middle line:84% Maybe it's a protein of interest. 01:15:32.980 --> 01:15:35.470 align:middle line:84% And there's a whole informatics trying to figure out 01:15:35.470 --> 01:15:37.720 align:middle line:84% and automate and image process these things, 01:15:37.720 --> 01:15:39.288 align:middle line:90% but they're terrible. 01:15:39.288 --> 01:15:40.330 align:middle line:90% This is not reproducible. 01:15:40.330 --> 01:15:42.970 align:middle line:84% Look how these spots are streaky, right? 01:15:42.970 --> 01:15:44.920 align:middle line:90% I mean, it's just a mess. 01:15:44.920 --> 01:15:48.760 align:middle line:84% Each spot has dozens, hundreds of proteins in there. 01:15:48.760 --> 01:15:51.280 align:middle line:90% Here's one paper I picked. 01:15:51.280 --> 01:15:53.140 align:middle line:84% These guys looked at all of these spots 01:15:53.140 --> 01:15:55.332 align:middle line:84% and sequenced all of them just to identify 01:15:55.332 --> 01:15:56.540 align:middle line:90% all these different proteins. 01:15:56.540 --> 01:15:58.370 align:middle line:90% So they've numbered each one. 01:15:58.370 --> 01:16:00.250 align:middle line:90% So back in '99, this is the-- 01:16:00.250 --> 01:16:02.140 align:middle line:84% [INAUDIBLE] actually in this building. 01:16:02.140 --> 01:16:03.557 align:middle line:84% I'm laughing at this, but, I mean, 01:16:03.557 --> 01:16:05.890 align:middle line:84% it's a huge amount of work to do this. 01:16:05.890 --> 01:16:09.070 align:middle line:84% So from that one spot, we might want to figure out 01:16:09.070 --> 01:16:10.510 align:middle line:90% what proteins are there. 01:16:10.510 --> 01:16:12.460 align:middle line:84% And this is a technique that most people use. 01:16:12.460 --> 01:16:17.360 align:middle line:84% It's this technique called mass spectrometry. 01:16:17.360 --> 01:16:20.277 align:middle line:84% And so the way this works is you get this little spot. 01:16:20.277 --> 01:16:22.360 align:middle line:84% Still, there's a whole bunch of peptides in there. 01:16:22.360 --> 01:16:25.310 align:middle line:84% Even though we separated them a lot in these two axes, 01:16:25.310 --> 01:16:26.920 align:middle line:84% there's a bunch of proteins here. 01:16:26.920 --> 01:16:30.340 align:middle line:84% And if you actually shoot it into this detector-- 01:16:30.340 --> 01:16:33.250 align:middle line:84% so basically, what mass spectrometry does is is it 01:16:33.250 --> 01:16:35.470 align:middle line:84% takes a spot and hits it with a laser. 01:16:35.470 --> 01:16:38.260 align:middle line:84% And the things come flying off of there. 01:16:38.260 --> 01:16:40.870 align:middle line:84% And the way that they fly off of there 01:16:40.870 --> 01:16:45.010 align:middle line:84% depends a lot on the size of the protein 01:16:45.010 --> 01:16:47.690 align:middle line:90% and the charge of the protein. 01:16:47.690 --> 01:16:50.080 align:middle line:84% So basically, in the end through all of this, 01:16:50.080 --> 01:16:55.460 align:middle line:84% you basically have one axis here, mass divided by charge. 01:16:55.460 --> 01:16:58.240 align:middle line:84% And I have peaks, basically, that are coming off 01:16:58.240 --> 01:16:59.050 align:middle line:90% of the detector. 01:16:59.050 --> 01:17:00.370 align:middle line:84% Oh, there's a whole bunch of stuff at this. 01:17:00.370 --> 01:17:01.030 align:middle line:90% Now, there's nothing. 01:17:01.030 --> 01:17:01.947 align:middle line:90% There's a whole bunch. 01:17:01.947 --> 01:17:06.130 align:middle line:84% There's nothing, like that, mass versus charge. 01:17:06.130 --> 01:17:08.050 align:middle line:84% So the way this works is what you can do 01:17:08.050 --> 01:17:11.980 align:middle line:84% is, these proteins here, you can digest them. 01:17:11.980 --> 01:17:15.760 align:middle line:84% Remember how these are long proteins. 01:17:15.760 --> 01:17:17.980 align:middle line:84% That same enzyme that I have in my pancreas that's 01:17:17.980 --> 01:17:21.100 align:middle line:84% still working on my lunch, I can use it to my advantage. 01:17:21.100 --> 01:17:24.070 align:middle line:84% I could take a spot and digest these peptides 01:17:24.070 --> 01:17:27.340 align:middle line:90% into successfully-- 01:17:27.340 --> 01:17:30.020 align:middle line:84% to have sequentially shorter protein. 01:17:30.020 --> 01:17:33.250 align:middle line:84% So here's one protein that's 40 amino acids long. 01:17:33.250 --> 01:17:34.150 align:middle line:90% Here's one with 39. 01:17:34.150 --> 01:17:34.960 align:middle line:90% Here's 138. 01:17:34.960 --> 01:17:35.830 align:middle line:90% Here's 137. 01:17:35.830 --> 01:17:36.857 align:middle line:90% Here's 136. 01:17:36.857 --> 01:17:38.440 align:middle line:84% And basically, I do this in such a way 01:17:38.440 --> 01:17:41.050 align:middle line:84% so that all of these peptides get broken up 01:17:41.050 --> 01:17:45.190 align:middle line:84% into a series of smaller peptides. 01:17:45.190 --> 01:17:46.560 align:middle line:90% Why am I doing all this? 01:17:46.560 --> 01:17:47.560 align:middle line:90% Why am I doing all this? 01:17:47.560 --> 01:17:50.140 align:middle line:90% Because this is the trick. 01:17:50.140 --> 01:17:53.320 align:middle line:84% If I want a protein that's 30 amino acids long 01:17:53.320 --> 01:17:56.830 align:middle line:84% and one the same protein, but it's only 29 amino acids long, 01:17:56.830 --> 01:18:02.530 align:middle line:84% the difference in size is the last amino acid. 01:18:02.530 --> 01:18:06.850 align:middle line:84% Then to go from 29 to 28, the difference in size there 01:18:06.850 --> 01:18:08.690 align:middle line:90% is that amino acid. 01:18:08.690 --> 01:18:11.140 align:middle line:84% So if I've basically successfully cut 01:18:11.140 --> 01:18:12.970 align:middle line:84% this one amino acid at a time and have 01:18:12.970 --> 01:18:14.860 align:middle line:84% a whole range of these, then I've 01:18:14.860 --> 01:18:18.100 align:middle line:84% got the difference in sizes to help me tell 01:18:18.100 --> 01:18:19.900 align:middle line:90% what that last amino acid was. 01:18:19.900 --> 01:18:21.010 align:middle line:90% And that's how we use it. 01:18:21.010 --> 01:18:22.720 align:middle line:84% We just look at the difference in peaks, 01:18:22.720 --> 01:18:25.950 align:middle line:84% the difference between the peaks is what we measure here. 01:18:25.950 --> 01:18:27.480 align:middle line:90% AUDIENCE: [INAUDIBLE]? 01:18:27.480 --> 01:18:29.310 align:middle line:90% ATUL J. BUTTE: No, not at all. 01:18:29.310 --> 01:18:30.505 align:middle line:90% It's a mess, right? 01:18:30.505 --> 01:18:31.380 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:18:31.380 --> 01:18:32.790 align:middle line:90% ATUL J. BUTTE: It's a ratio. 01:18:32.790 --> 01:18:34.620 align:middle line:84% So there's a bunch of peaks here, 01:18:34.620 --> 01:18:37.320 align:middle line:84% but that same peptide might be somewhere else, too, 01:18:37.320 --> 01:18:39.360 align:middle line:84% if it happened to get an extra charge. 01:18:39.360 --> 01:18:40.410 align:middle line:90% Absolutely. 01:18:40.410 --> 01:18:41.700 align:middle line:90% It's not perfect at all. 01:18:41.700 --> 01:18:44.670 align:middle line:84% It's not perfect at all, a whole new area here that's 01:18:44.670 --> 01:18:46.740 align:middle line:90% dying for new algorithms. 01:18:46.740 --> 01:18:47.980 align:middle line:90% You do it your best shot. 01:18:47.980 --> 01:18:49.170 align:middle line:90% You look at the peaks here. 01:18:49.170 --> 01:18:52.740 align:middle line:84% There's naive ways to do it, but they're not perfect at all. 01:18:52.740 --> 01:18:54.480 align:middle line:84% And of course, there's some proteins 01:18:54.480 --> 01:18:56.830 align:middle line:84% that love to have variable amounts of charge 01:18:56.830 --> 01:18:58.180 align:middle line:90% more than others. 01:18:58.180 --> 01:19:00.353 align:middle line:84% So some things are easier to detect than others. 01:19:00.353 --> 01:19:02.145 align:middle line:84% AUDIENCE: But I mean, the technique itself, 01:19:02.145 --> 01:19:08.820 align:middle line:84% there are known methods of how amino acids [INAUDIBLE] 01:19:08.820 --> 01:19:11.377 align:middle line:84% species themselves that they take [INAUDIBLE].. 01:19:11.377 --> 01:19:13.710 align:middle line:84% ATUL J. BUTTE: We know the size of each amino acid, each 01:19:13.710 --> 01:19:15.130 align:middle line:90% of the 20, 21 amino acids. 01:19:15.130 --> 01:19:17.910 align:middle line:84% So we can tell if there's a difference of peak here. 01:19:17.910 --> 01:19:19.850 align:middle line:84% But remember, there are multiple proteins. 01:19:19.850 --> 01:19:21.690 align:middle line:84% AUDIENCE: [INAUDIBLE] is not infallible, 01:19:21.690 --> 01:19:22.920 align:middle line:84% but there's characteristic [INAUDIBLE].. 01:19:22.920 --> 01:19:23.490 align:middle line:84% ATUL J. BUTTE: Absolutely, that's 01:19:23.490 --> 01:19:25.063 align:middle line:90% why we're able to use it at all. 01:19:25.063 --> 01:19:26.730 align:middle line:84% AUDIENCE: And there's libraries of that. 01:19:26.730 --> 01:19:27.370 align:middle line:90% There's a lot of-- 01:19:27.370 --> 01:19:28.410 align:middle line:90% ATUL J. BUTTE: Exactly. 01:19:28.410 --> 01:19:29.830 align:middle line:90% In fact, we can predict. 01:19:29.830 --> 01:19:31.710 align:middle line:84% We can look at every protein in the database 01:19:31.710 --> 01:19:33.870 align:middle line:84% and predict what it would look like when 01:19:33.870 --> 01:19:36.720 align:middle line:90% applied with trypsin to cut it. 01:19:36.720 --> 01:19:38.317 align:middle line:90% What would the peaks look like? 01:19:38.317 --> 01:19:39.900 align:middle line:84% And you can actually make predictions. 01:19:39.900 --> 01:19:41.665 align:middle line:84% And you basically take your pattern 01:19:41.665 --> 01:19:43.290 align:middle line:84% and compare it to the computer pattern. 01:19:43.290 --> 01:19:44.460 align:middle line:90% This might be predicted. 01:19:44.460 --> 01:19:45.420 align:middle line:90% This is the actual. 01:19:45.420 --> 01:19:47.220 align:middle line:84% And the computer, the algorithm here 01:19:47.220 --> 01:19:49.988 align:middle line:84% basically says, this is the protein. 01:19:49.988 --> 01:19:52.530 align:middle line:84% Now, there's another important point here I want to bring up. 01:19:52.530 --> 01:19:56.250 align:middle line:90% This is not quantitative. 01:19:56.250 --> 01:19:59.250 align:middle line:84% This helps you identify a protein in a sample, 01:19:59.250 --> 01:20:00.600 align:middle line:90% but it's not quantitative. 01:20:00.600 --> 01:20:02.490 align:middle line:84% It doesn't tell you how much there is. 01:20:02.490 --> 01:20:04.830 align:middle line:84% There's all sorts of newer, fancier techniques 01:20:04.830 --> 01:20:06.900 align:middle line:84% to compare this to some other sample 01:20:06.900 --> 01:20:09.300 align:middle line:84% to try to get some kind of quantitative measurement. 01:20:09.300 --> 01:20:11.790 align:middle line:84% But the way this is commonly used, it's just identifying. 01:20:11.790 --> 01:20:15.060 align:middle line:84% It's not quantitative, [INAUDIBLE].. 01:20:15.060 --> 01:20:17.050 align:middle line:84% AUDIENCE: Just a point, in The Science Times, 01:20:17.050 --> 01:20:18.633 align:middle line:84% in The New York Times this week, there 01:20:18.633 --> 01:20:21.670 align:middle line:84% was an article I saw about a test, a diagnostic 01:20:21.670 --> 01:20:23.227 align:middle line:90% for cervical cancer [INAUDIBLE]. 01:20:23.227 --> 01:20:24.310 align:middle line:90% ATUL J. BUTTE: Absolutely. 01:20:24.310 --> 01:20:25.480 align:middle line:90% AUDIENCE: They don't know how. 01:20:25.480 --> 01:20:26.290 align:middle line:90% It's just a tumor. 01:20:26.290 --> 01:20:27.550 align:middle line:84% ATUL J. BUTTE: I think I have that picture. 01:20:27.550 --> 01:20:28.177 align:middle line:90% AUDIENCE: OK. 01:20:28.177 --> 01:20:29.010 align:middle line:90% ATUL J. BUTTE: Yeah. 01:20:29.010 --> 01:20:30.100 align:middle line:90% Here, here's the one. 01:20:30.100 --> 01:20:31.480 align:middle line:90% So here's ovarian cancer-- 01:20:31.480 --> 01:20:32.500 align:middle line:90% same thing, right? 01:20:32.500 --> 01:20:35.230 align:middle line:84% So these guys use a particular chip to do this, 01:20:35.230 --> 01:20:37.545 align:middle line:84% but the chip is the smallest piece of it. 01:20:37.545 --> 01:20:39.670 align:middle line:84% It's basically still a mass spec that's on the end. 01:20:39.670 --> 01:20:43.030 align:middle line:84% Here is unaffected, unaffected, cancer one, cancer two. 01:20:43.030 --> 01:20:44.617 align:middle line:84% Here are all the different bands. 01:20:44.617 --> 01:20:46.450 align:middle line:84% And they say, to hell with identifying them. 01:20:46.450 --> 01:20:48.080 align:middle line:90% Here's the pattern. 01:20:48.080 --> 01:20:48.580 align:middle line:90% Oh, wow. 01:20:48.580 --> 01:20:51.130 align:middle line:84% I see a band here, a band here, but no band here. 01:20:51.130 --> 01:20:53.330 align:middle line:90% This must be diagnostic. 01:20:53.330 --> 01:20:54.580 align:middle line:90% I'll easily answer back. 01:20:54.580 --> 01:20:58.390 align:middle line:84% I mean, we're in danger of overfitting the data here. 01:20:58.390 --> 01:21:01.720 align:middle line:84% You have hundreds of thousands of peaks here and only 01:21:01.720 --> 01:21:03.430 align:middle line:90% four samples. 01:21:03.430 --> 01:21:06.808 align:middle line:84% How hard is it to find them that will answer your question? 01:21:06.808 --> 01:21:09.100 align:middle line:84% I tend to believe these things when they have something 01:21:09.100 --> 01:21:10.747 align:middle line:90% to do with the biology. 01:21:10.747 --> 01:21:13.330 align:middle line:84% If we go ahead and identify and there's some causal mechanism, 01:21:13.330 --> 01:21:13.992 align:middle line:90% that's great. 01:21:13.992 --> 01:21:16.450 align:middle line:84% But others are going to be very happy with this, especially 01:21:16.450 --> 01:21:19.210 align:middle line:84% since we have no other test for ovarian cancer today. 01:21:19.210 --> 01:21:20.905 align:middle line:84% That's fine-- or cervical cancer. 01:21:20.905 --> 01:21:22.600 align:middle line:84% For cervical cancer, we do Pap smears. 01:21:22.600 --> 01:21:23.680 align:middle line:84% For ovarian cancer, we have nothing. 01:21:23.680 --> 01:21:24.910 align:middle line:84% If there is something in the blood, 01:21:24.910 --> 01:21:27.243 align:middle line:84% it's better than nothing is what they would answer back. 01:21:27.243 --> 01:21:29.790 align:middle line:90% 01:21:29.790 --> 01:21:31.840 align:middle line:84% So here's basically the peak differences here. 01:21:31.840 --> 01:21:33.240 align:middle line:90% So they say, here's this peak. 01:21:33.240 --> 01:21:33.948 align:middle line:90% Here's this peak. 01:21:33.948 --> 01:21:35.700 align:middle line:84% We think that's an arginine or-- that's 01:21:35.700 --> 01:21:37.620 align:middle line:90% basically how this works. 01:21:37.620 --> 01:21:39.000 align:middle line:84% You use the differences in peaks. 01:21:39.000 --> 01:21:42.460 align:middle line:84% These things are called peak lists. 01:21:42.460 --> 01:21:44.650 align:middle line:84% Quantitative, we're not going to go there, 01:21:44.650 --> 01:21:47.080 align:middle line:84% but, basically, you can take one sample and another sample 01:21:47.080 --> 01:21:50.920 align:middle line:84% and grow the one sample in heavy water. 01:21:50.920 --> 01:21:53.770 align:middle line:90% Instead of H2O, we'll use H3O. 01:21:53.770 --> 01:21:56.530 align:middle line:84% And that extra H, it's not extra-- well, actually-- 01:21:56.530 --> 01:21:57.580 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:21:57.580 --> 01:21:58.966 align:middle line:90% ATUL J. BUTTE: It's-- 01:21:58.966 --> 01:22:00.058 align:middle line:90% AUDIENCE: Deuterium. 01:22:00.058 --> 01:22:01.600 align:middle line:84% ATUL J. BUTTE: It's deuterium, right. 01:22:01.600 --> 01:22:03.662 align:middle line:84% So it's not just deuterium, but you can also 01:22:03.662 --> 01:22:04.870 align:middle line:90% use other different isotopes. 01:22:04.870 --> 01:22:06.453 align:middle line:84% But basically, you have the same atom. 01:22:06.453 --> 01:22:09.520 align:middle line:84% Like, carbon can have 12 protons, 01:22:09.520 --> 01:22:11.410 align:middle line:84% or you can have 12 plus a neutron. 01:22:11.410 --> 01:22:13.030 align:middle line:84% You can have all sorts of other things 01:22:13.030 --> 01:22:15.998 align:middle line:84% in the nucleus of the atom that could change 01:22:15.998 --> 01:22:18.040 align:middle line:84% how it looks on mass spec because you're changing 01:22:18.040 --> 01:22:18.998 align:middle line:90% the mass just slightly. 01:22:18.998 --> 01:22:21.332 align:middle line:84% And then you can tell which sample it came from and then 01:22:21.332 --> 01:22:22.490 align:middle line:90% try to quantitate that way. 01:22:22.490 --> 01:22:25.340 align:middle line:84% But again, the charge could change. 01:22:25.340 --> 01:22:27.530 align:middle line:84% So one protein's not in one spot. 01:22:27.530 --> 01:22:28.930 align:middle line:90% It's in a whole bunch of slots. 01:22:28.930 --> 01:22:31.660 align:middle line:84% You need to sum them all up, and they're overlapping. 01:22:31.660 --> 01:22:35.110 align:middle line:84% AUDIENCE: [INAUDIBLE] trypsin [INAUDIBLE] peptide one 01:22:35.110 --> 01:22:37.852 align:middle line:90% [INAUDIBLE] then-- 01:22:37.852 --> 01:22:39.310 align:middle line:84% ATUL J. BUTTE: OK, let me be clear. 01:22:39.310 --> 01:22:41.530 align:middle line:84% The trypsin cuts into manageable fragments. 01:22:41.530 --> 01:22:43.807 align:middle line:84% And there's another fragmenter in front 01:22:43.807 --> 01:22:45.640 align:middle line:84% of the machine that's actually cutting it up 01:22:45.640 --> 01:22:47.290 align:middle line:90% into these one amino acids. 01:22:47.290 --> 01:22:47.800 align:middle line:90% Go ahead. 01:22:47.800 --> 01:22:50.095 align:middle line:84% AUDIENCE: You have [INAUDIBLE] different ways [INAUDIBLE].. 01:22:50.095 --> 01:22:50.887 align:middle line:90% ATUL J. BUTTE: Yes. 01:22:50.887 --> 01:22:55.510 align:middle line:84% AUDIENCE: [INAUDIBLE] and you have maybe [INAUDIBLE] number 01:22:55.510 --> 01:22:56.290 align:middle line:90% of [? charges-- ?] 01:22:56.290 --> 01:22:56.830 align:middle line:90% ATUL J. BUTTE: Absolutely. 01:22:56.830 --> 01:22:57.490 align:middle line:90% That's right. 01:22:57.490 --> 01:23:02.600 align:middle line:84% AUDIENCE: [INAUDIBLE] pretty much [INAUDIBLE] 01:23:02.600 --> 01:23:04.555 align:middle line:90% on what exactly [INAUDIBLE]. 01:23:04.555 --> 01:23:05.930 align:middle line:84% ATUL J. BUTTE: Well, the thing is 01:23:05.930 --> 01:23:08.270 align:middle line:84% the only hard part is that you don't 01:23:08.270 --> 01:23:10.430 align:middle line:90% know where the protein is. 01:23:10.430 --> 01:23:12.170 align:middle line:84% You see a whole bunch of spots there. 01:23:12.170 --> 01:23:13.640 align:middle line:84% And you see that it's here, but it 01:23:13.640 --> 01:23:16.980 align:middle line:84% could be on either side of the spectrum you're looking at. 01:23:16.980 --> 01:23:19.410 align:middle line:84% And if you let the machine run long enough, 01:23:19.410 --> 01:23:21.710 align:middle line:84% you get a huge amount of data off 01:23:21.710 --> 01:23:27.420 align:middle line:84% of that machine, an unmanageable amount of peaks. 01:23:27.420 --> 01:23:29.690 align:middle line:84% And we can look at and isolate one peak. 01:23:29.690 --> 01:23:33.560 align:middle line:84% Here, let me-- I think from this peak, we go to this peak. 01:23:33.560 --> 01:23:35.390 align:middle line:84% From that peak, we go to this peak. 01:23:35.390 --> 01:23:38.450 align:middle line:84% We've zoomed in so far to get the differences, 01:23:38.450 --> 01:23:40.670 align:middle line:84% but it could be elsewhere depending on the charge. 01:23:40.670 --> 01:23:43.250 align:middle line:90% 01:23:43.250 --> 01:23:44.570 align:middle line:90% It's doable. 01:23:44.570 --> 01:23:47.570 align:middle line:84% People do get great results off of this, but by no means 01:23:47.570 --> 01:23:50.270 align:middle line:84% is the informatics done in this field yet. 01:23:50.270 --> 01:23:52.010 align:middle line:84% And as more people use this technique, 01:23:52.010 --> 01:23:54.550 align:middle line:84% it's screaming for new algorithms. 01:23:54.550 --> 01:23:56.000 align:middle line:90% AUDIENCE: A quick statement-- 01:23:56.000 --> 01:23:59.270 align:middle line:84% is it obvious to everybody that the concentration of protein 01:23:59.270 --> 01:24:03.110 align:middle line:84% is not the same as concentration of RNA? 01:24:03.110 --> 01:24:05.900 align:middle line:90% Everybody should know that. 01:24:05.900 --> 01:24:06.560 align:middle line:90% In fact-- 01:24:06.560 --> 01:24:07.435 align:middle line:90% ATUL J. BUTTE: Right. 01:24:07.435 --> 01:24:10.130 align:middle line:84% Like I was saying, we're using it as a proxy. 01:24:10.130 --> 01:24:11.690 align:middle line:90% But let's be clear. 01:24:11.690 --> 01:24:14.960 align:middle line:84% If the cell has a whole bunch of this protein there 01:24:14.960 --> 01:24:18.950 align:middle line:84% and it's happy with that amount and it's not going anywhere, 01:24:18.950 --> 01:24:21.273 align:middle line:90% it doesn't need to make more. 01:24:21.273 --> 01:24:22.940 align:middle line:84% That's the simplest example of something 01:24:22.940 --> 01:24:25.835 align:middle line:84% where there might be zero RNA, but a whole lot of protein. 01:24:25.835 --> 01:24:28.400 align:middle line:90% 01:24:28.400 --> 01:24:30.447 align:middle line:84% If that protein's degrading really fast, 01:24:30.447 --> 01:24:32.030 align:middle line:84% then the cell might need to make more. 01:24:32.030 --> 01:24:33.897 align:middle line:84% But there's a massive disconnect here. 01:24:33.897 --> 01:24:36.230 align:middle line:84% Some things are going to be correlating, a lot of things 01:24:36.230 --> 01:24:37.880 align:middle line:90% are not. 01:24:37.880 --> 01:24:39.833 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:24:39.833 --> 01:24:41.750 align:middle line:84% ATUL J. BUTTE: All these things about proteins 01:24:41.750 --> 01:24:42.840 align:middle line:90% which we don't know-- 01:24:42.840 --> 01:24:45.206 align:middle line:90% depends on the protein. 01:24:45.206 --> 01:24:46.850 align:middle line:84% AUDIENCE: I just have a question. 01:24:46.850 --> 01:24:48.410 align:middle line:90% I'm just not sure. 01:24:48.410 --> 01:24:50.890 align:middle line:84% When you do small molecules [INAUDIBLE] compare 01:24:50.890 --> 01:24:54.110 align:middle line:84% the [INAUDIBLE] you eliminate the charges, 01:24:54.110 --> 01:24:56.953 align:middle line:84% is there any equivalent way that we could prepare-- 01:24:56.953 --> 01:24:58.620 align:middle line:84% ATUL J. BUTTE: To eliminate the charges? 01:24:58.620 --> 01:24:59.830 align:middle line:90% AUDIENCE: Yeah. 01:24:59.830 --> 01:25:01.830 align:middle line:84% ATUL J. BUTTE: You asking a detail I don't know. 01:25:01.830 --> 01:25:04.020 align:middle line:84% AUDIENCE: No, because that's the reason why 01:25:04.020 --> 01:25:08.467 align:middle line:84% [INAUDIBLE] small molecule [INAUDIBLE] proxy [INAUDIBLE].. 01:25:08.467 --> 01:25:09.550 align:middle line:90% ATUL J. BUTTE: Absolutely. 01:25:09.550 --> 01:25:11.842 align:middle line:84% But the thing is is that it's picking up charge, right? 01:25:11.842 --> 01:25:12.925 align:middle line:90% AUDIENCE: Yeah, of course. 01:25:12.925 --> 01:25:14.140 align:middle line:90% It is a charge, ultimately. 01:25:14.140 --> 01:25:17.463 align:middle line:84% But the reason why [INAUDIBLE] use 01:25:17.463 --> 01:25:18.630 align:middle line:90% that as a proxy [INAUDIBLE]. 01:25:18.630 --> 01:25:19.713 align:middle line:90% ATUL J. BUTTE: Absolutely. 01:25:19.713 --> 01:25:21.930 align:middle line:84% AUDIENCE: [INAUDIBLE] is not [INAUDIBLE].. 01:25:21.930 --> 01:25:22.170 align:middle line:90% ATUL J. BUTTE: Yeah. 01:25:22.170 --> 01:25:24.087 align:middle line:84% The biggest problem, though, is that these are 01:25:24.087 --> 01:25:25.380 align:middle line:90% longer than small molecules. 01:25:25.380 --> 01:25:27.930 align:middle line:84% I mean, these are much larger masses. 01:25:27.930 --> 01:25:28.890 align:middle line:90% That's the problem. 01:25:28.890 --> 01:25:32.658 align:middle line:84% And the proteins themselves, if you just 01:25:32.658 --> 01:25:34.950 align:middle line:84% chopped off the thing that's holding the charge or that 01:25:34.950 --> 01:25:35.640 align:middle line:90% loves-- 01:25:35.640 --> 01:25:37.663 align:middle line:84% if you just chopped off the polar part of it, 01:25:37.663 --> 01:25:39.330 align:middle line:84% it's going to have different properties. 01:25:39.330 --> 01:25:42.090 align:middle line:84% That next peak could be wildly different. 01:25:42.090 --> 01:25:43.230 align:middle line:90% It totally depends on that. 01:25:43.230 --> 01:25:46.650 align:middle line:84% AUDIENCE: But for my microarray lecture is, why is it, 01:25:46.650 --> 01:25:50.370 align:middle line:84% if they, in fact, as [INAUDIBLE] explain it, protein function, 01:25:50.370 --> 01:25:53.490 align:middle line:84% protein synthesis is not in a 1 to 1 01:25:53.490 --> 01:25:56.640 align:middle line:84% relationship with RNA synthesis [INAUDIBLE] 01:25:56.640 --> 01:26:01.170 align:middle line:84% so successful in defining disease classes 01:26:01.170 --> 01:26:05.840 align:middle line:84% and the [INAUDIBLE] just based on RNA expression? 01:26:05.840 --> 01:26:08.630 align:middle line:90% 01:26:08.630 --> 01:26:10.730 align:middle line:84% Why is it-- why should we [INAUDIBLE]?? 01:26:10.730 --> 01:26:14.510 align:middle line:90% 01:26:14.510 --> 01:26:16.817 align:middle line:90% Or could it? 01:26:16.817 --> 01:26:19.150 align:middle line:84% ATUL J. BUTTE: I think this is my-- this or the next one 01:26:19.150 --> 01:26:21.315 align:middle line:90% is the last slide here. 01:26:21.315 --> 01:26:23.440 align:middle line:84% Now, despite it being hard to measure these things, 01:26:23.440 --> 01:26:24.898 align:middle line:84% there are definitely companies that 01:26:24.898 --> 01:26:26.680 align:middle line:84% are trying to make protein chips here. 01:26:26.680 --> 01:26:28.810 align:middle line:84% One way to detect proteins is to have 01:26:28.810 --> 01:26:30.173 align:middle line:90% an antibody against a protein. 01:26:30.173 --> 01:26:31.840 align:middle line:84% That's how people have been doing things 01:26:31.840 --> 01:26:33.347 align:middle line:90% like Western blots for decades. 01:26:33.347 --> 01:26:34.930 align:middle line:84% And you can radiolabel them and stuff. 01:26:34.930 --> 01:26:37.390 align:middle line:84% So you can make a plate here where each well 01:26:37.390 --> 01:26:41.680 align:middle line:84% has an antibody against one particular protein. 01:26:41.680 --> 01:26:44.300 align:middle line:84% Or you could make-- for example, people make a cytokine array. 01:26:44.300 --> 01:26:46.000 align:middle line:84% So this array just basically-- each spot 01:26:46.000 --> 01:26:47.292 align:middle line:90% looks for a different cytokine. 01:26:47.292 --> 01:26:49.000 align:middle line:84% You can buy these things, but you 01:26:49.000 --> 01:26:51.760 align:middle line:84% can't do this comprehensively like can with microarrays. 01:26:51.760 --> 01:26:52.900 align:middle line:90% That's the point. 01:26:52.900 --> 01:26:54.970 align:middle line:84% If you look across all possible proteins 01:26:54.970 --> 01:26:57.805 align:middle line:84% plus alternative splice products in proteins, 01:26:57.805 --> 01:27:00.190 align:middle line:90% you just can't do it today. 01:27:00.190 --> 01:27:02.490 align:middle line:84% You can do functional assays as well. 01:27:02.490 --> 01:27:05.620 align:middle line:90% I mean, it's antibody array. 01:27:05.620 --> 01:27:06.120 align:middle line:90% That's it. 01:27:06.120 --> 01:27:08.100 align:middle line:84% So we talked about sequencing, polymorphisms. 01:27:08.100 --> 01:27:09.570 align:middle line:90% We skipped SAGE. 01:27:09.570 --> 01:27:11.940 align:middle line:84% Most people are using microarrays now. 01:27:11.940 --> 01:27:13.320 align:middle line:90% We didn't even cover wafers. 01:27:13.320 --> 01:27:15.180 align:middle line:90% Let's cover wafers in a second-- 01:27:15.180 --> 01:27:17.160 align:middle line:84% 2D-PAGE, mass spec protein arrays. 01:27:17.160 --> 01:27:18.780 align:middle line:84% Let me at least cover the wafer thing 01:27:18.780 --> 01:27:25.110 align:middle line:90% because it's worth knowing that. 01:27:25.110 --> 01:27:27.000 align:middle line:90% It's all slides. 01:27:27.000 --> 01:27:29.375 align:middle line:90% So the wafers blow my mind. 01:27:29.375 --> 01:27:31.000 align:middle line:84% And that's why I like to end with this. 01:27:31.000 --> 01:27:33.120 align:middle line:84% So those companies that make these microarrays, 01:27:33.120 --> 01:27:34.530 align:middle line:84% they don't just make them one at a time. 01:27:34.530 --> 01:27:36.697 align:middle line:84% They make them the same way computer chips are made. 01:27:36.697 --> 01:27:38.850 align:middle line:84% They're made 40 at a time in a wafer, 01:27:38.850 --> 01:27:40.578 align:middle line:90% and then they're just cut out. 01:27:40.578 --> 01:27:42.120 align:middle line:84% So there's a company called Perlegen, 01:27:42.120 --> 01:27:43.320 align:middle line:84% which is owned by that other company 01:27:43.320 --> 01:27:45.112 align:middle line:84% Affymetrix, that basically said, why are we 01:27:45.112 --> 01:27:46.320 align:middle line:90% making 40 of the same array? 01:27:46.320 --> 01:27:49.560 align:middle line:84% Let's just make a wafer one big array. 01:27:49.560 --> 01:27:51.870 align:middle line:84% And with that, you have so many spots. 01:27:51.870 --> 01:27:54.300 align:middle line:84% You have 60 million spots in one wafer. 01:27:54.300 --> 01:27:55.560 align:middle line:90% It's about 5 inches. 01:27:55.560 --> 01:27:58.077 align:middle line:84% I'm drawing it like this, but it's really this small. 01:27:58.077 --> 01:27:59.160 align:middle line:90% You can see the size here. 01:27:59.160 --> 01:27:59.827 align:middle line:90% This is a wafer. 01:27:59.827 --> 01:28:01.830 align:middle line:84% A 5 inch square has 60 million probes. 01:28:01.830 --> 01:28:03.938 align:middle line:84% So even two years ago, they showed 01:28:03.938 --> 01:28:05.730 align:middle line:84% how you can use these wafers to re-sequence 01:28:05.730 --> 01:28:07.290 align:middle line:90% an entire chromosome. 01:28:07.290 --> 01:28:08.680 align:middle line:90% And that was one paper. 01:28:08.680 --> 01:28:10.710 align:middle line:84% You think about how many years it 01:28:10.710 --> 01:28:13.090 align:middle line:84% took to sequence of first chromosome, the Human Genome 01:28:13.090 --> 01:28:13.590 align:middle line:90% Project. 01:28:13.590 --> 01:28:17.130 align:middle line:84% They can just do that now over a weekend because they basically 01:28:17.130 --> 01:28:22.000 align:middle line:84% have each spot is a moving window of 25 base pairs. 01:28:22.000 --> 01:28:23.140 align:middle line:90% So what do I mean by that? 01:28:23.140 --> 01:28:24.557 align:middle line:84% They took the smallest chromosome. 01:28:24.557 --> 01:28:25.920 align:middle line:90% I think it was, like, 22, 23. 01:28:25.920 --> 01:28:27.450 align:middle line:84% They took the smallest chromosome, 01:28:27.450 --> 01:28:30.180 align:middle line:84% and they took the first 25 nucleotides. 01:28:30.180 --> 01:28:31.347 align:middle line:90% And they put it in one spot. 01:28:31.347 --> 01:28:33.763 align:middle line:84% And then the middle one they said, well, it could be an A, 01:28:33.763 --> 01:28:35.340 align:middle line:84% T, C, or G. So that's four spots. 01:28:35.340 --> 01:28:37.660 align:middle line:84% Now, let's move to the next one, to the next one. 01:28:37.660 --> 01:28:40.740 align:middle line:84% So they have a moving window of 25 nucleotides going 01:28:40.740 --> 01:28:43.710 align:middle line:84% from one end to the other end of the chromosome 01:28:43.710 --> 01:28:45.893 align:middle line:90% across a series of wafers. 01:28:45.893 --> 01:28:47.310 align:middle line:84% And once they have that done, they 01:28:47.310 --> 01:28:49.435 align:middle line:84% can just take anyone's blood and just basically get 01:28:49.435 --> 01:28:50.640 align:middle line:90% the entire pattern now. 01:28:50.640 --> 01:28:51.690 align:middle line:84% And they don't have to say, well, 01:28:51.690 --> 01:28:53.440 align:middle line:84% I think there's a SNP here or a SNP there. 01:28:53.440 --> 01:28:55.558 align:middle line:84% They just know all the SNPs now because it's just 01:28:55.558 --> 01:28:57.600 align:middle line:84% a matter of how many humans you put on the array. 01:28:57.600 --> 01:28:59.490 align:middle line:90% And it's just blood. 01:28:59.490 --> 01:29:05.550 align:middle line:84% The big issue-- each scan, the TIFF image, takes 10 terabytes. 01:29:05.550 --> 01:29:07.650 align:middle line:84% We were just learning about gigabytes. 01:29:07.650 --> 01:29:10.740 align:middle line:84% Terabyte is 1,000 of those, and this is 10 of them 01:29:10.740 --> 01:29:13.770 align:middle line:90% for just one of these wafers. 01:29:13.770 --> 01:29:16.140 align:middle line:84% So it's well-known that life science data 01:29:16.140 --> 01:29:19.260 align:middle line:84% is growing much faster than this famous Moore's law. 01:29:19.260 --> 01:29:21.300 align:middle line:84% Moore's law is from Gordon Moore from Intel 01:29:21.300 --> 01:29:24.390 align:middle line:84% that said that microprocessor power doubles every 18 months. 01:29:24.390 --> 01:29:27.370 align:middle line:84% Life science data is growing way faster than that. 01:29:27.370 --> 01:29:29.100 align:middle line:84% This is just one example of this. 01:29:29.100 --> 01:29:30.750 align:middle line:84% They've publicly stated that they 01:29:30.750 --> 01:29:33.240 align:middle line:84% can sequence all SNPs in a human in 10 days, 01:29:33.240 --> 01:29:34.600 align:middle line:90% not just the ones we know about. 01:29:34.600 --> 01:29:36.070 align:middle line:90% They just know all of them. 01:29:36.070 --> 01:29:37.800 align:middle line:84% And this is just one of many companies. 01:29:37.800 --> 01:29:39.720 align:middle line:84% There's many in the 95 Beltway here 01:29:39.720 --> 01:29:42.180 align:middle line:84% that are doing the same thing as competition. 01:29:42.180 --> 01:29:43.770 align:middle line:84% So remember how I was joking around 01:29:43.770 --> 01:29:45.960 align:middle line:84% that you could fit your entire genome on a CD-ROM? 01:29:45.960 --> 01:29:47.760 align:middle line:84% It is absolutely conceivable that we 01:29:47.760 --> 01:29:51.990 align:middle line:84% can have this within the next one or two years if we want it. 01:29:51.990 --> 01:29:53.833 align:middle line:84% The technology is there to do this. 01:29:53.833 --> 01:29:56.250 align:middle line:84% They certainly have the hard drive space there at Perlegen 01:29:56.250 --> 01:29:57.562 align:middle line:90% to do it. 01:29:57.562 --> 01:29:58.770 align:middle line:90% All right, I think that's it. 01:29:58.770 --> 01:30:01.490 align:middle line:84% And we should end since we're a little bit over.