WEBVTT 00:00:00.000 --> 00:00:00.610 align:middle line:90% 00:00:00.610 --> 00:00:02.440 align:middle line:84% The following content is provided 00:00:02.440 --> 00:00:06.340 align:middle line:84% by MIT OpenCourseWare under a Creative Commons license. 00:00:06.340 --> 00:00:08.500 align:middle line:84% Additional information about our license 00:00:08.500 --> 00:00:10.870 align:middle line:84% and MIT OpenCourseWare in general 00:00:10.870 --> 00:00:12.730 align:middle line:90% is available at ocw.mit.edu. 00:00:12.730 --> 00:00:16.728 align:middle line:90% 00:00:16.728 --> 00:00:20.102 align:middle line:90% [SIDE CONVERSATION] 00:00:20.102 --> 00:00:41.560 align:middle line:90% 00:00:41.560 --> 00:00:44.560 align:middle line:84% GEORGE CHURCH: OK, welcome back to the second half 00:00:44.560 --> 00:00:47.530 align:middle line:90% of the second lecture. 00:00:47.530 --> 00:00:53.620 align:middle line:84% Here, we're going to take this beautiful example 00:00:53.620 --> 00:01:02.140 align:middle line:84% of an algorithm, which the cellular machinery applies 00:01:02.140 --> 00:01:04.190 align:middle line:90% to get from DNA to RNA protein. 00:01:04.190 --> 00:01:10.780 align:middle line:84% And so a simple example where that 00:01:10.780 --> 00:01:14.440 align:middle line:84% is incorporated into a program very similar to what you'll 00:01:14.440 --> 00:01:16.520 align:middle line:90% be doing on your problem set. 00:01:16.520 --> 00:01:19.750 align:middle line:84% So here, we've got it nicely color coded. 00:01:19.750 --> 00:01:24.010 align:middle line:84% This is unrelated, of course, to the color coding 00:01:24.010 --> 00:01:26.230 align:middle line:84% and the genetic code we've been using so far, 00:01:26.230 --> 00:01:29.560 align:middle line:84% but it includes the genetic code. 00:01:29.560 --> 00:01:33.280 align:middle line:84% You can see the comments here are preceded 00:01:33.280 --> 00:01:36.340 align:middle line:90% by this little number sign. 00:01:36.340 --> 00:01:40.720 align:middle line:84% We go from the genome, a DNA sequence, which is a string. 00:01:40.720 --> 00:01:43.930 align:middle line:84% Strings are one of the things that Perl does very well. 00:01:43.930 --> 00:01:46.630 align:middle line:84% It's not the only string-manipulation programming 00:01:46.630 --> 00:01:48.790 align:middle line:84% language, but it's a particularly easy one. 00:01:48.790 --> 00:01:51.340 align:middle line:84% And so here is how easy it is to enter 00:01:51.340 --> 00:01:53.960 align:middle line:84% a DNA sequence, one of the many ways of entering a DNA 00:01:53.960 --> 00:01:54.460 align:middle line:90% sequence. 00:01:54.460 --> 00:01:55.793 align:middle line:90% You can bring it in from a file. 00:01:55.793 --> 00:01:58.390 align:middle line:90% Here it is as part of the code. 00:01:58.390 --> 00:02:03.790 align:middle line:84% You transcribe it into RNA in silico here, 00:02:03.790 --> 00:02:07.000 align:middle line:84% by the simple command where you say 00:02:07.000 --> 00:02:10.539 align:middle line:84% RNA sequence is equal to DNA sequence, 00:02:10.539 --> 00:02:16.480 align:middle line:84% and then you substitute all the Ts for Us globally. 00:02:16.480 --> 00:02:21.550 align:middle line:84% That's what this 12th line is on slide 23. 00:02:21.550 --> 00:02:23.260 align:middle line:84% Now, off to the side here, is a reminder 00:02:23.260 --> 00:02:25.670 align:middle line:84% that it's really much more complicated than that. 00:02:25.670 --> 00:02:28.030 align:middle line:84% There's all these proteins involved and doing it 00:02:28.030 --> 00:02:31.100 align:middle line:84% accurately and with regulations, so forth. 00:02:31.100 --> 00:02:32.710 align:middle line:84% But for the sake of this Perl program, 00:02:32.710 --> 00:02:36.490 align:middle line:84% this is quite sufficient to get us to an RNA sequence, 00:02:36.490 --> 00:02:42.040 align:middle line:84% which we can then translate, and here, the translation process 00:02:42.040 --> 00:02:43.440 align:middle line:90% uses-- 00:02:43.440 --> 00:02:45.140 align:middle line:90% it's going to be a cycle. 00:02:45.140 --> 00:02:48.190 align:middle line:84% So in a sense, the cycle in the RNA m 00:02:48.190 --> 00:02:53.180 align:middle line:84% where you put in one nucleotide at a m is all compact here. 00:02:53.180 --> 00:02:55.360 align:middle line:84% It's all just substitute every T for U. 00:02:55.360 --> 00:02:59.500 align:middle line:84% Here, we're going to have a more explicit loop, this while loop. 00:02:59.500 --> 00:03:00.730 align:middle line:90% You can see it's indented. 00:03:00.730 --> 00:03:03.310 align:middle line:84% Everything within the loop, that's going to be iterated, 00:03:03.310 --> 00:03:04.293 align:middle line:90% is offset a bit. 00:03:04.293 --> 00:03:05.710 align:middle line:84% And what you're going to be doing, 00:03:05.710 --> 00:03:07.900 align:middle line:90% is looking in groups of 3. 00:03:07.900 --> 00:03:12.550 align:middle line:84% That's what this 3 is on line 17 is 00:03:12.550 --> 00:03:15.970 align:middle line:84% looking through the position you are in the RNA, 00:03:15.970 --> 00:03:18.130 align:middle line:84% and taking a chunk of 3 at a time. 00:03:18.130 --> 00:03:21.280 align:middle line:84% The chunk here is a substring, substring. 00:03:21.280 --> 00:03:26.860 align:middle line:84% And you pull out a codon as that substring. 00:03:26.860 --> 00:03:28.490 align:middle line:90% And then you do a translation. 00:03:28.490 --> 00:03:30.250 align:middle line:84% So now this is, in a sense, representing 00:03:30.250 --> 00:03:32.560 align:middle line:84% the modularity of biology and the modularity 00:03:32.560 --> 00:03:33.880 align:middle line:90% of good programming code. 00:03:33.880 --> 00:03:39.160 align:middle line:84% It's you put this whole business of translating 00:03:39.160 --> 00:03:40.790 align:middle line:84% in a separate part of the program, 00:03:40.790 --> 00:03:43.600 align:middle line:84% so you don't have to embed the code everywhere 00:03:43.600 --> 00:03:44.800 align:middle line:90% that it's going to be used. 00:03:44.800 --> 00:03:47.140 align:middle line:84% And the translation here, is that simple table. 00:03:47.140 --> 00:03:50.560 align:middle line:84% So this subroutine, S-U-B, line 22, again, 00:03:50.560 --> 00:03:53.410 align:middle line:84% indented all the code in the subroutine. 00:03:53.410 --> 00:03:55.660 align:middle line:84% It goes off the bottom of the slide 00:03:55.660 --> 00:04:00.370 align:middle line:84% and down through the floor all the different cases. 00:04:00.370 --> 00:04:03.970 align:middle line:84% Now, we could list 64 cases for the 64 trinucleotides, 00:04:03.970 --> 00:04:06.490 align:middle line:84% or we can use the more compact string manipulation 00:04:06.490 --> 00:04:08.230 align:middle line:84% that you can do in Perl to indicate-- 00:04:08.230 --> 00:04:11.950 align:middle line:84% a dot means any kind of character after G and C, 00:04:11.950 --> 00:04:17.440 align:middle line:84% would return alanine so GC, ACG, or U. 00:04:17.440 --> 00:04:19.740 align:middle line:84% And then for cysteine, which has two possibilities, 00:04:19.740 --> 00:04:22.060 align:middle line:84% so that's four possibilities represented by that dot. 00:04:22.060 --> 00:04:24.140 align:middle line:84% And cysteine has two possibilities. 00:04:24.140 --> 00:04:27.610 align:middle line:84% It's either UGC, or UGU, where the vertical line means or. 00:04:27.610 --> 00:04:28.550 align:middle line:90% And you get the idea. 00:04:28.550 --> 00:04:31.570 align:middle line:84% There's a whole, very compact, syntax 00:04:31.570 --> 00:04:35.470 align:middle line:90% here for doing the translation. 00:04:35.470 --> 00:04:43.210 align:middle line:84% And that's how we do one of the cleaner, more simple algorithms 00:04:43.210 --> 00:04:45.190 align:middle line:90% in computational biology. 00:04:45.190 --> 00:04:49.660 align:middle line:84% And now I'm going to make it complicated again. 00:04:49.660 --> 00:04:55.300 align:middle line:84% But first, to set the stage for how the genetic code is not 00:04:55.300 --> 00:04:58.420 align:middle line:84% universal, we have to explain by what we might 00:04:58.420 --> 00:05:00.100 align:middle line:90% mean by it being universal. 00:05:00.100 --> 00:05:06.130 align:middle line:84% This is the ultimate pedigree on slide 24 here. 00:05:06.130 --> 00:05:10.480 align:middle line:84% It's, in principle, some very simple organism, 00:05:10.480 --> 00:05:13.600 align:middle line:84% possibly an RNA-based organism that 00:05:13.600 --> 00:05:16.960 align:middle line:84% may have made a proto-ribosome that 00:05:16.960 --> 00:05:19.180 align:middle line:84% may have done protein synthesis or may have done 00:05:19.180 --> 00:05:23.620 align:middle line:84% some other chemical reaction at what is now [INAUDIBLE] 00:05:23.620 --> 00:05:24.670 align:middle line:90% transferase site. 00:05:24.670 --> 00:05:28.120 align:middle line:84% Something along this line, it's speculated by some biologists, 00:05:28.120 --> 00:05:32.980 align:middle line:84% was the common ancestor of all living species, all living 00:05:32.980 --> 00:05:34.000 align:middle line:90% cells. 00:05:34.000 --> 00:05:37.180 align:middle line:84% And certainly, by the time we started 00:05:37.180 --> 00:05:39.070 align:middle line:84% getting branching of the three major branches 00:05:39.070 --> 00:05:45.040 align:middle line:84% bacteria, the archaea bacteria, and the eukaryotes, by the time 00:05:45.040 --> 00:05:49.750 align:middle line:84% we get to that point, there was, probably 00:05:49.750 --> 00:05:53.710 align:middle line:84% a ribosome set of genes that encodes all the proteins 00:05:53.710 --> 00:05:56.260 align:middle line:84% and RNAs or the ribosome, and these were shared. 00:05:56.260 --> 00:05:59.838 align:middle line:84% And then at each cell division, you 00:05:59.838 --> 00:06:01.630 align:middle line:84% split off two cells were slightly different 00:06:01.630 --> 00:06:02.338 align:middle line:90% from one another. 00:06:02.338 --> 00:06:06.190 align:middle line:84% As they would mutate and differentiate and be selected, 00:06:06.190 --> 00:06:08.920 align:middle line:84% they would generate this huge diversity of organisms, 00:06:08.920 --> 00:06:10.520 align:middle line:90% which you see here. 00:06:10.520 --> 00:06:13.660 align:middle line:84% Now, this is a [? directed ?] basically graph in the sense 00:06:13.660 --> 00:06:17.980 align:middle line:84% that you can't have a descendant in this process 00:06:17.980 --> 00:06:22.780 align:middle line:84% being an ancestor of one of its parents. 00:06:22.780 --> 00:06:27.920 align:middle line:84% So time has, as an axis, going up in this case, 00:06:27.920 --> 00:06:31.210 align:middle line:84% unlike some of the more physics based diagrams we had before. 00:06:31.210 --> 00:06:36.040 align:middle line:84% And as you branch out to existing species, 00:06:36.040 --> 00:06:40.900 align:middle line:84% you see that things like plants, have actually inheritance, 00:06:40.900 --> 00:06:45.160 align:middle line:84% not just along this direct tree-like structure, 00:06:45.160 --> 00:06:48.130 align:middle line:84% but you've got more of a network-like structure, where 00:06:48.130 --> 00:06:49.660 align:middle line:84% some of the genomic material came 00:06:49.660 --> 00:06:53.110 align:middle line:84% in from one of the bacterial branches long ago. 00:06:53.110 --> 00:06:57.700 align:middle line:84% And this has recently been put on very firm footing 00:06:57.700 --> 00:07:01.990 align:middle line:84% on a genomic scale, in a recent article this month, 00:07:01.990 --> 00:07:05.110 align:middle line:84% where the bacterial genome of cyanobacteria, these 00:07:05.110 --> 00:07:08.050 align:middle line:84% are blue-green algae that fix carbon 00:07:08.050 --> 00:07:12.010 align:middle line:90% in all the oceans and plant. 00:07:12.010 --> 00:07:14.110 align:middle line:84% And the simplest plant that's been sequenced 00:07:14.110 --> 00:07:17.650 align:middle line:90% is Arabidopsis, a weed. 00:07:17.650 --> 00:07:20.245 align:middle line:84% And anyway, the DNA from the bacteria 00:07:20.245 --> 00:07:22.120 align:middle line:84% has not only gone into the chloroplast, which 00:07:22.120 --> 00:07:25.390 align:middle line:84% is an organelle, which has a very reduced genome, 00:07:25.390 --> 00:07:27.670 align:middle line:84% but it's been spread through thousands of genes 00:07:27.670 --> 00:07:30.460 align:middle line:84% in the nucleus, which is the major place where 00:07:30.460 --> 00:07:32.810 align:middle line:84% all the chromosomes are for the plants. 00:07:32.810 --> 00:07:38.860 align:middle line:84% And so this, possibly symbiotic relationship, 00:07:38.860 --> 00:07:42.050 align:middle line:84% has resulted in a complicated inheritance. 00:07:42.050 --> 00:07:43.780 align:middle line:90% And this is not unique. 00:07:43.780 --> 00:07:45.970 align:middle line:84% Another one coming in from the purple bacteria, 00:07:45.970 --> 00:07:48.190 align:middle line:84% has been incorporated into a separate membrane-bound 00:07:48.190 --> 00:07:52.840 align:middle line:84% organelle, which provides the ATP for both plants 00:07:52.840 --> 00:07:56.890 align:middle line:84% and non-photosynthetic multicellular 00:07:56.890 --> 00:07:58.780 align:middle line:90% organisms, animals. 00:07:58.780 --> 00:08:02.260 align:middle line:84% And these two arrows, that could be 00:08:02.260 --> 00:08:05.590 align:middle line:84% there could be thousands of arrows going 00:08:05.590 --> 00:08:08.170 align:middle line:84% over long periods of time at deep branches, 00:08:08.170 --> 00:08:10.690 align:middle line:84% and we know there are many interconnecting 00:08:10.690 --> 00:08:14.560 align:middle line:84% arrows in recent times in almost all these organisms, 00:08:14.560 --> 00:08:17.500 align:middle line:84% or certain representatives from all over this tree, 00:08:17.500 --> 00:08:20.980 align:middle line:84% can take up DNA in various ways, can even 00:08:20.980 --> 00:08:24.580 align:middle line:84% mate with organisms of various species and exchange DNA 00:08:24.580 --> 00:08:25.840 align:middle line:90% and incorporate it. 00:08:25.840 --> 00:08:28.090 align:middle line:84% And so it's not this simple tree, 00:08:28.090 --> 00:08:33.610 align:middle line:84% but it is, certainly, going forward in time. 00:08:33.610 --> 00:08:39.210 align:middle line:84% It is directed and acyclic in that regard. 00:08:39.210 --> 00:08:40.789 align:middle line:84% So how many living species are there? 00:08:40.789 --> 00:08:43.070 align:middle line:84% We're building to the point, connecting back 00:08:43.070 --> 00:08:46.670 align:middle line:84% to the central dogma, of how many different genetic codes 00:08:46.670 --> 00:08:47.280 align:middle line:90% are there. 00:08:47.280 --> 00:08:49.310 align:middle line:84% So we need to know how many species there are. 00:08:49.310 --> 00:08:56.600 align:middle line:84% If you take a gram of some thumbnail of soil, 00:08:56.600 --> 00:08:59.000 align:middle line:84% from any of a variety of different soils 00:08:59.000 --> 00:09:00.770 align:middle line:84% that have been tested, you can find 00:09:00.770 --> 00:09:02.283 align:middle line:90% about 5,000 bacterial species. 00:09:02.283 --> 00:09:03.950 align:middle line:84% Well, what does it mean to be a species? 00:09:03.950 --> 00:09:06.020 align:middle line:84% In animals, that typically means that they 00:09:06.020 --> 00:09:08.630 align:middle line:84% don't produce fertile offspring when they interbreed 00:09:08.630 --> 00:09:09.740 align:middle line:90% two different species. 00:09:09.740 --> 00:09:12.320 align:middle line:84% But there are books full of exceptions of this, 00:09:12.320 --> 00:09:13.080 align:middle line:90% even in animals. 00:09:13.080 --> 00:09:15.080 align:middle line:84% And bacteria, of course, where they exchange DNA 00:09:15.080 --> 00:09:17.840 align:middle line:84% all over the place, as I said, it breaks down even more. 00:09:17.840 --> 00:09:21.980 align:middle line:84% So working definition that many biologists adopt, 00:09:21.980 --> 00:09:27.950 align:middle line:84% is that if two microbial species share 20% of their DNA, 00:09:27.950 --> 00:09:30.350 align:middle line:84% if you take their DNA and align it by algorithms, 00:09:30.350 --> 00:09:33.080 align:middle line:84% such as the ones we we'll be using this course, 00:09:33.080 --> 00:09:36.750 align:middle line:84% and you find that 70% of the base pairs are conserved, 00:09:36.750 --> 00:09:38.030 align:middle line:90% then they're the same species. 00:09:38.030 --> 00:09:40.700 align:middle line:84% Otherwise, they're different species. 00:09:40.700 --> 00:09:44.160 align:middle line:84% And there are millions of non-microbial species, 00:09:44.160 --> 00:09:47.090 align:middle line:84% many of which harbor microbial species. 00:09:47.090 --> 00:09:53.420 align:middle line:84% This number is dropping slightly because of our inability 00:09:53.420 --> 00:09:58.610 align:middle line:84% to restrain our growth and other activities 00:09:58.610 --> 00:10:04.250 align:middle line:90% that cause species extinction. 00:10:04.250 --> 00:10:07.880 align:middle line:84% And the number of whole genomes is getting closer to 100, 00:10:07.880 --> 00:10:12.620 align:middle line:84% and the number in the pipeline is probably 600 or so, 00:10:12.620 --> 00:10:15.800 align:middle line:84% maybe in the thousands, with new technologies. 00:10:15.800 --> 00:10:20.540 align:middle line:84% And there's over 80,000 species defined 00:10:20.540 --> 00:10:24.620 align:middle line:84% by one or more nucleic acid sequences in the NCBI, National 00:10:24.620 --> 00:10:27.650 align:middle line:84% Center for Bioinformatics, which is 00:10:27.650 --> 00:10:31.173 align:middle line:84% one of the three major nucleic-acid databases 00:10:31.173 --> 00:10:31.715 align:middle line:90% in the world. 00:10:31.715 --> 00:10:36.170 align:middle line:90% 00:10:36.170 --> 00:10:37.900 align:middle line:84% Why do we study more than one species? 00:10:37.900 --> 00:10:43.660 align:middle line:84% The comparison between species allows subtle and not so subtle 00:10:43.660 --> 00:10:47.530 align:middle line:84% analyses of what are the important positions 00:10:47.530 --> 00:10:51.130 align:middle line:84% to stay constant because they provide some very fundamental 00:10:51.130 --> 00:10:52.690 align:middle line:90% biochemical activity. 00:10:52.690 --> 00:10:55.660 align:middle line:84% But what are the important ones to vary because they provide 00:10:55.660 --> 00:10:57.580 align:middle line:84% some important variants, for example, 00:10:57.580 --> 00:11:00.443 align:middle line:84% escaping immune surveillance and so on. 00:11:00.443 --> 00:11:01.860 align:middle line:84% So there's reasons to be constant, 00:11:01.860 --> 00:11:04.820 align:middle line:84% there's reason to be variable, and reasons to be neutral. 00:11:04.820 --> 00:11:06.910 align:middle line:84% So let's go back now and apply this 00:11:06.910 --> 00:11:08.500 align:middle line:84% to the genetic code, this particularly 00:11:08.500 --> 00:11:12.260 align:middle line:84% simple and elegant, nearly-universal code. 00:11:12.260 --> 00:11:14.920 align:middle line:84% This is how genetic codes are represented 00:11:14.920 --> 00:11:16.690 align:middle line:90% in CBI, one of the ways. 00:11:16.690 --> 00:11:19.420 align:middle line:84% And here, the three bases of the codon, 00:11:19.420 --> 00:11:22.060 align:middle line:84% bases 1, 2, and 3 remember, we said 00:11:22.060 --> 00:11:26.080 align:middle line:84% that you UUU was phenylalanine encoded by DNA TTT. 00:11:26.080 --> 00:11:29.410 align:middle line:84% So going down from the top on the leftmost column, 00:11:29.410 --> 00:11:33.740 align:middle line:84% it's TTT single letter code F for phenylalanine. 00:11:33.740 --> 00:11:40.190 align:middle line:84% So the amino acids go along the bottom row of this table. 00:11:40.190 --> 00:11:44.110 align:middle line:84% And you can see all the amino acids are represented. 00:11:44.110 --> 00:11:45.910 align:middle line:84% Stars represent stop codons which 00:11:45.910 --> 00:11:47.620 align:middle line:84% are not recognized by transfer RNAs 00:11:47.620 --> 00:11:50.020 align:middle line:84% but by proteins called release factors that 00:11:50.020 --> 00:11:51.970 align:middle line:84% simulate the function of the transfer RNA 00:11:51.970 --> 00:11:56.410 align:middle line:84% and cause the release of this cyclic incorporation 00:11:56.410 --> 00:11:58.450 align:middle line:84% of amino acids into polypeptides. 00:11:58.450 --> 00:12:01.290 align:middle line:90% 00:12:01.290 --> 00:12:05.790 align:middle line:84% Now, this is the standard code, so-called, 00:12:05.790 --> 00:12:08.790 align:middle line:84% where you have one methionine here in the middle 00:12:08.790 --> 00:12:15.420 align:middle line:84% here, encoded by ATG, and three stop codons in all the rest. 00:12:15.420 --> 00:12:17.230 align:middle line:84% Here's where it gets complicated. 00:12:17.230 --> 00:12:22.440 align:middle line:84% There are over 22 different genetic codes. 00:12:22.440 --> 00:12:24.300 align:middle line:84% Some of the changes from the standard code 00:12:24.300 --> 00:12:27.390 align:middle line:90% are indicated here in blue. 00:12:27.390 --> 00:12:29.910 align:middle line:84% You have here, for example, for the [INAUDIBLE] 00:12:29.910 --> 00:12:32.430 align:middle line:84% mitochondrial code, this is the code actually used 00:12:32.430 --> 00:12:35.580 align:middle line:84% in every cell in your body for the subset of the cell that 00:12:35.580 --> 00:12:38.760 align:middle line:84% is the powerhouse that makes ATP, the mitochondria that we 00:12:38.760 --> 00:12:40.830 align:middle line:84% talked about before, which was part 00:12:40.830 --> 00:12:43.920 align:middle line:84% of the horizontal transmission of information 00:12:43.920 --> 00:12:46.380 align:middle line:90% from purple bacteria long ago. 00:12:46.380 --> 00:12:50.430 align:middle line:84% But anyway, the normal stop codon is now tryptophan, 00:12:50.430 --> 00:12:54.570 align:middle line:84% abbreviated W. There's an extra methionine. 00:12:54.570 --> 00:12:56.142 align:middle line:84% And there's two extra stop codons, 00:12:56.142 --> 00:12:57.600 align:middle line:84% which are replacing what would have 00:12:57.600 --> 00:12:59.257 align:middle line:84% been an arginine in the standard code. 00:12:59.257 --> 00:13:01.590 align:middle line:84% And you can see there's little blues all over the place. 00:13:01.590 --> 00:13:05.415 align:middle line:84% Changes in the number where you can start. 00:13:05.415 --> 00:13:08.760 align:middle line:84% The start sites that are indicated by 3, 00:13:08.760 --> 00:13:11.370 align:middle line:84% 1 to 3 starts in the standard code. 00:13:11.370 --> 00:13:18.630 align:middle line:84% And when you talk about the starts, you start getting to 00:13:18.630 --> 00:13:22.800 align:middle line:84% how much do you favor that particular start? 00:13:22.800 --> 00:13:25.443 align:middle line:84% What other signals are required to start 00:13:25.443 --> 00:13:26.610 align:middle line:90% at that particular position? 00:13:26.610 --> 00:13:27.990 align:middle line:84% It's not as simple as just having 00:13:27.990 --> 00:13:32.610 align:middle line:84% an ATG trinucleotide to get a start of protein synthesis. 00:13:32.610 --> 00:13:36.240 align:middle line:84% You need other nucleic acid components. 00:13:36.240 --> 00:13:41.460 align:middle line:84% Anyway, still, it's a slightly more complicated algorithm. 00:13:41.460 --> 00:13:44.070 align:middle line:84% You have to know exactly what organelle 00:13:44.070 --> 00:13:45.840 align:middle line:84% and what organism you're dealing with, 00:13:45.840 --> 00:13:49.650 align:middle line:84% but you can apply the same kind of computer codes 00:13:49.650 --> 00:13:52.680 align:middle line:84% that we had, a couple of slides back. 00:13:52.680 --> 00:13:54.780 align:middle line:84% But now we get into even more complicated. 00:13:54.780 --> 00:13:57.090 align:middle line:84% And part of the reason I'm showing you this early on, 00:13:57.090 --> 00:14:01.500 align:middle line:84% some of these things would not be in your textbooks 00:14:01.500 --> 00:14:02.940 align:middle line:90% for the first biology course. 00:14:02.940 --> 00:14:07.290 align:middle line:84% And they would not be in a first computational biology 00:14:07.290 --> 00:14:10.080 align:middle line:90% lecture or two. 00:14:10.080 --> 00:14:12.750 align:middle line:84% But I do this so that you'll have a healthy distrust 00:14:12.750 --> 00:14:15.210 align:middle line:84% of everything that you read and everything you hear, 00:14:15.210 --> 00:14:18.510 align:middle line:84% including everything you hear from me. 00:14:18.510 --> 00:14:22.770 align:middle line:84% And this should really make you distrust the genetic code. 00:14:22.770 --> 00:14:26.220 align:middle line:84% Because what these ribosomes do, in this particular sequence, 00:14:26.220 --> 00:14:29.880 align:middle line:84% which has been well-documented, eight years ago, 00:14:29.880 --> 00:14:32.490 align:middle line:84% is that they will hop over 50 nucleotides. 00:14:32.490 --> 00:14:36.360 align:middle line:84% They're not going 3 nucleotides at a time as they should, 00:14:36.360 --> 00:14:38.220 align:middle line:84% in fact, it's not even an integral multiple 00:14:38.220 --> 00:14:39.360 align:middle line:90% of 3 nucleotides. 00:14:39.360 --> 00:14:41.220 align:middle line:84% It is literally coming to a stop codon, 00:14:41.220 --> 00:14:45.570 align:middle line:84% and rather than stopping, if it has just the right sequence 00:14:45.570 --> 00:14:49.890 align:middle line:84% context, including this complicated RNA 00:14:49.890 --> 00:14:51.810 align:middle line:84% secondary structure called a pseudoknot. 00:14:51.810 --> 00:14:54.930 align:middle line:84% The RNA folds up the messenger, which really should just 00:14:54.930 --> 00:14:57.690 align:middle line:84% be a messenger, which the computer should just 00:14:57.690 --> 00:15:00.120 align:middle line:90% be recognizing-- 00:15:00.120 --> 00:15:03.077 align:middle line:84% or the biological, biochemical computer, as is the ribosome, 00:15:03.077 --> 00:15:05.160 align:middle line:84% should be recognizing three nucleotides at a time, 00:15:05.160 --> 00:15:08.200 align:middle line:84% instead it's recognizing a morphology. 00:15:08.200 --> 00:15:10.470 align:middle line:84% This thing folds up, and no longer 00:15:10.470 --> 00:15:11.730 align:middle line:90% is an informational molecule. 00:15:11.730 --> 00:15:14.790 align:middle line:84% It is a morphological recognition element. 00:15:14.790 --> 00:15:16.290 align:middle line:84% Anyway, when it finds that, it skips 00:15:16.290 --> 00:15:18.690 align:middle line:84% over 50 nucleotides, skips over the stop codon, 00:15:18.690 --> 00:15:22.890 align:middle line:84% and makes a, otherwise, perfectly normal protein. 00:15:22.890 --> 00:15:27.930 align:middle line:84% So don't even trust dogma, especially don't trust dogma, 00:15:27.930 --> 00:15:30.000 align:middle line:90% central dogma included. 00:15:30.000 --> 00:15:32.880 align:middle line:90% Plenty of counter-examples. 00:15:32.880 --> 00:15:39.600 align:middle line:84% Now, we're going to move from this very, very simple example 00:15:39.600 --> 00:15:44.430 align:middle line:84% of an algorithm, where we can model proteins directly 00:15:44.430 --> 00:15:49.140 align:middle line:84% from the nucleic acids that come out of DNA sequences wholesale. 00:15:49.140 --> 00:15:53.740 align:middle line:84% We now want to ask, how do we get the more quantitative data, 00:15:53.740 --> 00:15:55.740 align:middle line:84% which comes out of functional genomics rather 00:15:55.740 --> 00:15:58.380 align:middle line:90% than classical sequencing? 00:15:58.380 --> 00:16:00.960 align:middle line:84% How do we get that into quantitative models, 00:16:00.960 --> 00:16:03.600 align:middle line:84% and then get the quantitative models then repopulated 00:16:03.600 --> 00:16:09.180 align:middle line:84% with additional quantitative data to make a full model? 00:16:09.180 --> 00:16:13.050 align:middle line:84% Now, in order to say, came up earlier 00:16:13.050 --> 00:16:19.230 align:middle line:84% as a question, what is the function of a gene product? 00:16:19.230 --> 00:16:23.700 align:middle line:84% We dip into qualitative statements 00:16:23.700 --> 00:16:26.280 align:middle line:84% which are made in the literature, which 00:16:26.280 --> 00:16:28.620 align:middle line:84% have various ways of representing 00:16:28.620 --> 00:16:29.700 align:middle line:90% the evidence for this. 00:16:29.700 --> 00:16:32.970 align:middle line:84% Some of them very convoluted arguments, some of them 00:16:32.970 --> 00:16:35.760 align:middle line:90% very casual. 00:16:35.760 --> 00:16:38.190 align:middle line:84% But when an attempt is made to put these 00:16:38.190 --> 00:16:41.430 align:middle line:84% into a database or a data structure, 00:16:41.430 --> 00:16:44.220 align:middle line:84% as representative, gross oversimplification 00:16:44.220 --> 00:16:46.505 align:middle line:84% of the literature, this is what often comes out, 00:16:46.505 --> 00:16:47.880 align:middle line:84% something like this, where you'll 00:16:47.880 --> 00:16:49.260 align:middle line:90% have a hierarchical table. 00:16:49.260 --> 00:16:52.000 align:middle line:84% Here, I've blown up one of the levels of the hierarchy. 00:16:52.000 --> 00:16:54.840 align:middle line:84% You can think of as a list, where the list may not 00:16:54.840 --> 00:16:57.060 align:middle line:84% be in a particularly logical order, 00:16:57.060 --> 00:17:01.190 align:middle line:84% but the hierarchy is, so that under metabolism 00:17:01.190 --> 00:17:04.250 align:middle line:84% would be some covalent change in substrates, 00:17:04.250 --> 00:17:08.463 align:middle line:90% which enzymes would catalyze. 00:17:08.463 --> 00:17:10.130 align:middle line:84% And then you'd have information transfer 00:17:10.130 --> 00:17:13.075 align:middle line:84% we've been talking about, like the DNA to RNA 00:17:13.075 --> 00:17:14.540 align:middle line:90% protein, these biopolymers. 00:17:14.540 --> 00:17:18.200 align:middle line:84% Regulation of information transfer or of metabolism 00:17:18.200 --> 00:17:21.770 align:middle line:84% would have all these four subheadings type of regulation, 00:17:21.770 --> 00:17:23.460 align:middle line:90% trigger, and so on. 00:17:23.460 --> 00:17:26.089 align:middle line:84% And then transport in these various other processes. 00:17:26.089 --> 00:17:29.810 align:middle line:84% Each of these functions, such as illustrated by these references 00:17:29.810 --> 00:17:35.990 align:middle line:84% here, can be used in a way of connecting 00:17:35.990 --> 00:17:39.230 align:middle line:84% all the new information we get to some 00:17:39.230 --> 00:17:47.780 align:middle line:84% of systematic best guess, best estimate, of encapsulation 00:17:47.780 --> 00:17:49.100 align:middle line:90% of the literature. 00:17:49.100 --> 00:17:53.840 align:middle line:84% Another example of this, in addition to the MIPS for yeast, 00:17:53.840 --> 00:17:58.400 align:middle line:84% is gene ontology, which is derived from the word ontology, 00:17:58.400 --> 00:17:59.960 align:middle line:90% or nature of being. 00:17:59.960 --> 00:18:03.800 align:middle line:84% And the objective f GO, abbreviation of Gene Ontology, 00:18:03.800 --> 00:18:07.760 align:middle line:84% is to provide a controlled vocabulary. 00:18:07.760 --> 00:18:09.470 align:middle line:84% My vocabulary during this lecture 00:18:09.470 --> 00:18:14.000 align:middle line:84% has been uncontrolled, as you've probably guessed. 00:18:14.000 --> 00:18:16.280 align:middle line:84% But when you start talking about-- 00:18:16.280 --> 00:18:19.160 align:middle line:84% I have pointed out the problems that you 00:18:19.160 --> 00:18:22.220 align:middle line:84% get into when you casually refer to gene expression 00:18:22.220 --> 00:18:24.410 align:middle line:84% when you really mean RNA expression, 00:18:24.410 --> 00:18:27.470 align:middle line:84% and refer to genes as protein-coding entities, 00:18:27.470 --> 00:18:30.200 align:middle line:84% when you really mean protein or RNA-encoding entities. 00:18:30.200 --> 00:18:33.378 align:middle line:84% That process of being more precise about our use of terms, 00:18:33.378 --> 00:18:35.420 align:middle line:84% at least when we're communicating with computers, 00:18:35.420 --> 00:18:36.300 align:middle line:90% is very important. 00:18:36.300 --> 00:18:38.883 align:middle line:84% We communicate with each other, you guys will give me a little 00:18:38.883 --> 00:18:42.380 align:middle line:84% slack, some of you but computers won't. 00:18:42.380 --> 00:18:45.240 align:middle line:84% They will misinterpret every chance they get. 00:18:45.240 --> 00:18:48.900 align:middle line:84% And so that's what control of vocabulary is all about. 00:18:48.900 --> 00:18:50.550 align:middle line:84% And you'll have three different-- 00:18:50.550 --> 00:18:54.560 align:middle line:84% Dave, the inventors of gene ontology, 00:18:54.560 --> 00:18:58.040 align:middle line:84% have a hierarchy including molecular function, 00:18:58.040 --> 00:19:00.140 align:middle line:84% biological processes, and the cellular component, 00:19:00.140 --> 00:19:04.290 align:middle line:84% which we'll expand upon in the next slide. 00:19:04.290 --> 00:19:07.470 align:middle line:84% Cautionary note, whenever you do modeling or you 00:19:07.470 --> 00:19:09.630 align:middle line:84% will be assumptions in this case. 00:19:09.630 --> 00:19:12.210 align:middle line:84% Some of the assumptions exclude vast parts 00:19:12.210 --> 00:19:14.730 align:middle line:84% of biology, which are listed here 00:19:14.730 --> 00:19:18.210 align:middle line:90% as part of their documentation. 00:19:18.210 --> 00:19:23.460 align:middle line:84% They have things that are not modeled in the gene ontology 00:19:23.460 --> 00:19:26.520 align:middle line:84% are domain structure and three-dimensional structure, 00:19:26.520 --> 00:19:29.800 align:middle line:84% which, obviously, has played a big role in the two lectures 00:19:29.800 --> 00:19:31.380 align:middle line:90% so far. 00:19:31.380 --> 00:19:35.850 align:middle line:84% The evolution and gene expression, we've 00:19:35.850 --> 00:19:39.510 align:middle line:84% already talked about the phylogenetic tree of evolution, 00:19:39.510 --> 00:19:41.790 align:middle line:84% and the gene expression will be a big topic 00:19:41.790 --> 00:19:44.910 align:middle line:84% in the RNA and proteomics part of this course. 00:19:44.910 --> 00:19:47.190 align:middle line:84% The small molecules we've illustrated today. 00:19:47.190 --> 00:19:48.660 align:middle line:84% Almost everything in this course, 00:19:48.660 --> 00:19:50.970 align:middle line:84% seems to be excluded from the gene ontology. 00:19:50.970 --> 00:19:56.610 align:middle line:84% Nevertheless, here we go with just one slide 00:19:56.610 --> 00:19:58.560 align:middle line:90% talking about the functions. 00:19:58.560 --> 00:19:59.910 align:middle line:90% We have molecular function. 00:19:59.910 --> 00:20:03.160 align:middle line:84% What the gene product can do without specifying where 00:20:03.160 --> 00:20:04.440 align:middle line:90% or when. 00:20:04.440 --> 00:20:08.220 align:middle line:84% A broad example of this, would be enzyme, 00:20:08.220 --> 00:20:09.720 align:middle line:90% something that catalyzes. 00:20:09.720 --> 00:20:11.970 align:middle line:84% And then a very specific example of an enzyme 00:20:11.970 --> 00:20:14.310 align:middle line:84% would be, a adenylate cyclase, something 00:20:14.310 --> 00:20:22.210 align:middle line:84% that makes a cycle in the ribose of a adenylate. 00:20:22.210 --> 00:20:25.410 align:middle line:84% So both of these fall under the net molecular function 00:20:25.410 --> 00:20:29.280 align:middle line:84% when you're describing a function of a protein 00:20:29.280 --> 00:20:31.890 align:middle line:90% in describing a genome. 00:20:31.890 --> 00:20:35.010 align:middle line:84% A biological process has to have more than one step. 00:20:35.010 --> 00:20:37.950 align:middle line:84% If it's one step, that's not a process. 00:20:37.950 --> 00:20:41.130 align:middle line:84% It has to have a time component, typically, 00:20:41.130 --> 00:20:42.990 align:middle line:84% and there's a transformation that occurs. 00:20:42.990 --> 00:20:46.720 align:middle line:84% Examples of signal transduction is a broad biological process. 00:20:46.720 --> 00:20:48.360 align:middle line:84% An example of signal transduction 00:20:48.360 --> 00:20:51.600 align:middle line:90% is cyclic AMP biosynthesis. 00:20:51.600 --> 00:20:54.450 align:middle line:84% The cellular component, would be somehow reflecting 00:20:54.450 --> 00:20:57.720 align:middle line:84% this assembly to organelle that we were talking about earlier. 00:20:57.720 --> 00:21:01.110 align:middle line:84% And here, an example, you have a ribosomal protein 00:21:01.110 --> 00:21:03.770 align:middle line:90% being part of a ribosome. 00:21:03.770 --> 00:21:05.520 align:middle line:84% So it gets you some idea of the component, 00:21:05.520 --> 00:21:07.860 align:middle line:84% some molecular-function biological process 00:21:07.860 --> 00:21:09.090 align:middle line:90% and components. 00:21:09.090 --> 00:21:13.980 align:middle line:84% Now, as I said, this gene ontology is based on facts. 00:21:13.980 --> 00:21:18.780 align:middle line:84% The facts that are included, it's not-- 00:21:18.780 --> 00:21:21.750 align:middle line:84% ideally, there would be a direct logical connection 00:21:21.750 --> 00:21:24.300 align:middle line:84% between the facts that are summarized in the hierarchical 00:21:24.300 --> 00:21:26.580 align:middle line:84% gene ontology and the raw data that 00:21:26.580 --> 00:21:27.870 align:middle line:90% came out of some instrument. 00:21:27.870 --> 00:21:29.042 align:middle line:90% That is not the case. 00:21:29.042 --> 00:21:30.750 align:middle line:84% This is all from the literature, and it's 00:21:30.750 --> 00:21:33.420 align:middle line:90% done on a low budget, wow. 00:21:33.420 --> 00:21:37.110 align:middle line:84% And examples of how they summarize it, 00:21:37.110 --> 00:21:39.690 align:middle line:84% is it's inferred from a mutant phenotype 00:21:39.690 --> 00:21:41.230 align:middle line:90% or a genetic interaction. 00:21:41.230 --> 00:21:42.420 align:middle line:90% So these two are genetic. 00:21:42.420 --> 00:21:47.670 align:middle line:84% Or physical interaction, this passes for biophysics. 00:21:47.670 --> 00:21:49.620 align:middle line:84% Sequence similarity, now we're starting-- 00:21:49.620 --> 00:21:51.450 align:middle line:84% as we go down this list, we're starting 00:21:51.450 --> 00:21:55.350 align:middle line:84% to get into murkier and murkier evidence. 00:21:55.350 --> 00:21:58.740 align:middle line:84% Sequence similarity, as you'll see in a subsequence slide, 00:21:58.740 --> 00:22:02.700 align:middle line:90% has problems. 00:22:02.700 --> 00:22:08.910 align:middle line:84% A direct assay could be a physical interaction, 00:22:08.910 --> 00:22:11.670 align:middle line:84% or it could be some other biochemical assay. 00:22:11.670 --> 00:22:14.880 align:middle line:84% Expression pattern might be evidence 00:22:14.880 --> 00:22:18.570 align:middle line:84% of some of the associations that were 00:22:18.570 --> 00:22:20.790 align:middle line:90% mentioned in the gene ontology. 00:22:20.790 --> 00:22:22.920 align:middle line:84% Now we get to electronic annotation. 00:22:22.920 --> 00:22:25.290 align:middle line:84% In a certain sense, all of these things 00:22:25.290 --> 00:22:26.785 align:middle line:90% are electronic annotation. 00:22:26.785 --> 00:22:29.160 align:middle line:84% Sequence similarity might be a way that you automatically 00:22:29.160 --> 00:22:30.690 align:middle line:90% get electronic annotation. 00:22:30.690 --> 00:22:33.100 align:middle line:84% Then you get to a traceable author statement. 00:22:33.100 --> 00:22:35.880 align:middle line:84% This means that someone said something is true, 00:22:35.880 --> 00:22:39.360 align:middle line:84% without saying how he or she knows it's true, 00:22:39.360 --> 00:22:40.680 align:middle line:90% so we're getting really murky. 00:22:40.680 --> 00:22:44.820 align:middle line:84% And the murkiest of all is non-traceable authors' data. 00:22:44.820 --> 00:22:49.140 align:middle line:84% You don't even know who said something might be true, OK. 00:22:49.140 --> 00:22:51.180 align:middle line:84% Let's go back up to the top of that, 00:22:51.180 --> 00:22:52.650 align:middle line:84% in fact, go beyond the top of it, 00:22:52.650 --> 00:22:57.300 align:middle line:84% where we now will start tracking the data from the instruments 00:22:57.300 --> 00:22:59.250 align:middle line:90% to statements. 00:22:59.250 --> 00:23:02.100 align:middle line:84% And hopefully, in this course, you 00:23:02.100 --> 00:23:09.270 align:middle line:84% will see how we will, in the future and present, 00:23:09.270 --> 00:23:12.390 align:middle line:84% make models in a rigorous way where you can track it 00:23:12.390 --> 00:23:13.990 align:middle line:90% all the way back, to data. 00:23:13.990 --> 00:23:18.270 align:middle line:84% So one class, the most obvious class, of data collection, 00:23:18.270 --> 00:23:20.790 align:middle line:84% is what I would call direct observation, typically 00:23:20.790 --> 00:23:22.050 align:middle line:90% through a microscope. 00:23:22.050 --> 00:23:24.420 align:middle line:84% And here's a particularly powerful case. 00:23:24.420 --> 00:23:27.330 align:middle line:84% I promised you earlier that we would 00:23:27.330 --> 00:23:33.180 align:middle line:84% talk about how you have 959 cells in the non-gonadal cell 00:23:33.180 --> 00:23:34.350 align:middle line:90% lineage of the worm. 00:23:34.350 --> 00:23:36.120 align:middle line:84% It starts as a single cell up here 00:23:36.120 --> 00:23:38.730 align:middle line:84% at the top middle, a fertilized egg, 00:23:38.730 --> 00:23:40.890 align:middle line:84% as this egg-shaped thing at the top. 00:23:40.890 --> 00:23:45.085 align:middle line:84% And then it splits off, way off to the left 00:23:45.085 --> 00:23:45.960 align:middle line:90% and off to the right. 00:23:45.960 --> 00:23:48.150 align:middle line:84% And that makes two cells, two stem cells, 00:23:48.150 --> 00:23:51.060 align:middle line:84% that are capable of differentiating and dividing 00:23:51.060 --> 00:23:51.750 align:middle line:90% further. 00:23:51.750 --> 00:23:55.480 align:middle line:84% And they each make two more, and it keeps going. 00:23:55.480 --> 00:23:59.310 align:middle line:84% But you can see it starts getting symmetry breaking, 00:23:59.310 --> 00:24:00.580 align:middle line:90% almost immediately. 00:24:00.580 --> 00:24:05.320 align:middle line:84% In fact, the egg itself is an asymmetric entity. 00:24:05.320 --> 00:24:07.030 align:middle line:84% And you start getting lineages that 00:24:07.030 --> 00:24:10.300 align:middle line:84% will either die as they terminate, 00:24:10.300 --> 00:24:12.550 align:middle line:90% or they will just stop dividing. 00:24:12.550 --> 00:24:18.190 align:middle line:84% And eventually, you end up with, after about 1,000 cell 00:24:18.190 --> 00:24:23.470 align:middle line:84% divisions or so, you end up with these 959 non-gonadal cells. 00:24:23.470 --> 00:24:26.050 align:middle line:90% 00:24:26.050 --> 00:24:28.780 align:middle line:84% And this lineage has been completely mapped out 00:24:28.780 --> 00:24:31.510 align:middle line:84% by direct microscopic observation, where 00:24:31.510 --> 00:24:33.820 align:middle line:84% a series of photographs you can show 00:24:33.820 --> 00:24:36.970 align:middle line:84% that this single cell turns into these two cells, 00:24:36.970 --> 00:24:38.830 align:middle line:84% so you have a time axis, and you have 00:24:38.830 --> 00:24:44.650 align:middle line:84% a lineage axis, which is one of these directed acyclic graphs. 00:24:44.650 --> 00:24:47.320 align:middle line:84% In addition, and even more amazing, 00:24:47.320 --> 00:24:50.680 align:middle line:84% to me, anyway, is that you have a complete neural connection 00:24:50.680 --> 00:24:52.160 align:middle line:90% for this multicellular organism. 00:24:52.160 --> 00:24:55.900 align:middle line:84% It has a fairly simple brain if you've ever had a conversation 00:24:55.900 --> 00:24:57.010 align:middle line:90% with one of these things. 00:24:57.010 --> 00:25:09.550 align:middle line:84% But each neuron can have dozens to hundreds of connections. 00:25:09.550 --> 00:25:11.830 align:middle line:84% And these have been mapped by a serial section 00:25:11.830 --> 00:25:14.830 align:middle line:84% through the entire worm, very thin sections 00:25:14.830 --> 00:25:16.280 align:middle line:90% and electron microscopy. 00:25:16.280 --> 00:25:18.350 align:middle line:84% And then checking out the whole wiring diagram. 00:25:18.350 --> 00:25:19.900 align:middle line:90% This is really a tour de force. 00:25:19.900 --> 00:25:21.580 align:middle line:84% And part of the reason it's possible, 00:25:21.580 --> 00:25:24.190 align:middle line:84% this would even be hard to do in a variety of organisms, 00:25:24.190 --> 00:25:26.590 align:middle line:84% but this is another case where biology cooperates, 00:25:26.590 --> 00:25:30.190 align:middle line:84% just like with the genetic code, in this particular organism, 00:25:30.190 --> 00:25:34.120 align:middle line:84% that lineage happens the same way every time. 00:25:34.120 --> 00:25:37.180 align:middle line:84% In even slightly more complicated organisms, 00:25:37.180 --> 00:25:40.870 align:middle line:84% like the Drosophila fruit fly or humans. 00:25:40.870 --> 00:25:45.460 align:middle line:84% The lineages are not so strict, and a cell 00:25:45.460 --> 00:25:48.100 align:middle line:84% can take on a number of different directions 00:25:48.100 --> 00:25:52.300 align:middle line:84% depending on the exact physical environment it finds itself in. 00:25:52.300 --> 00:25:54.850 align:middle line:84% But nevertheless, for this one, the neural connections 00:25:54.850 --> 00:25:57.550 align:middle line:84% are reproducible, and the cell lineage is reproducible. 00:25:57.550 --> 00:26:00.100 align:middle line:90% And so you can map this all out. 00:26:00.100 --> 00:26:02.710 align:middle line:84% For other organisms, it doesn't mean you shouldn't try it, 00:26:02.710 --> 00:26:05.002 align:middle line:84% it just means you'll have to represent it in a slightly 00:26:05.002 --> 00:26:07.240 align:middle line:90% less-fixed pattern. 00:26:07.240 --> 00:26:10.420 align:middle line:84% You'll have to represent it as probabilistic set of divisions 00:26:10.420 --> 00:26:12.610 align:middle line:84% and probabilistic set of neural connections. 00:26:12.610 --> 00:26:17.500 align:middle line:84% And maybe even, conditional, on various conditions. 00:26:17.500 --> 00:26:19.840 align:middle line:84% OK, that's direct observation as a class 00:26:19.840 --> 00:26:23.470 align:middle line:90% of source of data for modeling. 00:26:23.470 --> 00:26:25.810 align:middle line:84% Here's three other sources of data. 00:26:25.810 --> 00:26:28.930 align:middle line:84% In each case, I've shown pretty raw representations 00:26:28.930 --> 00:26:30.160 align:middle line:90% of the data. 00:26:30.160 --> 00:26:33.070 align:middle line:84% You can think of these all as representing an intensity 00:26:33.070 --> 00:26:38.200 align:middle line:84% readout, with some sort of separation, 00:26:38.200 --> 00:26:44.570 align:middle line:84% as the horizontal axis or is some cases, both axes. 00:26:44.570 --> 00:26:46.810 align:middle line:84% So the intensity here, is indicated 00:26:46.810 --> 00:26:50.140 align:middle line:84% by a line plot of four different color 00:26:50.140 --> 00:26:54.550 align:middle line:84% fluorophores in an electrophoretic separation, 00:26:54.550 --> 00:26:56.800 align:middle line:84% which is the basis for the genomic sequencing 00:26:56.800 --> 00:27:00.880 align:middle line:90% that we're so proud of here. 00:27:00.880 --> 00:27:06.070 align:middle line:84% Then so here, the detection of this fluorescence 00:27:06.070 --> 00:27:09.940 align:middle line:84% of the terminated chains of DNA, we'll 00:27:09.940 --> 00:27:11.890 align:middle line:90% get to that later in the course. 00:27:11.890 --> 00:27:13.510 align:middle line:84% Here you have mass spectrometry where 00:27:13.510 --> 00:27:16.570 align:middle line:84% you're measuring differences in masses, 00:27:16.570 --> 00:27:18.910 align:middle line:84% even more accurately than in sequencing. 00:27:18.910 --> 00:27:21.370 align:middle line:84% You're separating nucleic acids by their differences 00:27:21.370 --> 00:27:24.070 align:middle line:84% in mass, so about 1 part in 1,000. 00:27:24.070 --> 00:27:26.710 align:middle line:84% The mass spectrometry is more like 1 part in 10,000, 00:27:26.710 --> 00:27:28.780 align:middle line:90% or even better. 00:27:28.780 --> 00:27:31.480 align:middle line:84% Because here, you're separating in a gas phase, 00:27:31.480 --> 00:27:34.780 align:middle line:84% based on the electrical and magnetic properties. 00:27:34.780 --> 00:27:37.450 align:middle line:84% Here, you're separating by charge 00:27:37.450 --> 00:27:42.610 align:middle line:84% in a liquid phase, liquid and gel phase. 00:27:42.610 --> 00:27:46.570 align:middle line:84% Each of these, you can specify the throughput per day 00:27:46.570 --> 00:27:49.360 align:middle line:84% or the throughput per unit dollar. 00:27:49.360 --> 00:27:52.390 align:middle line:84% This becomes important in planning these structures. 00:27:52.390 --> 00:27:56.590 align:middle line:84% The third category here, is arrays. 00:27:56.590 --> 00:28:00.670 align:middle line:84% These can be arrays of nucleic acids for quantitating RNAs, 00:28:00.670 --> 00:28:04.450 align:middle line:84% or arrays of antibodies, proteins, 00:28:04.450 --> 00:28:08.050 align:middle line:84% small-molecule chemicals, which we can quantitate 00:28:08.050 --> 00:28:10.030 align:middle line:84% the binding of one kind of molecule 00:28:10.030 --> 00:28:13.090 align:middle line:90% to an array of other molecules. 00:28:13.090 --> 00:28:16.810 align:middle line:84% In both the top and the bottom, you can have multiple colors. 00:28:16.810 --> 00:28:19.720 align:middle line:84% And these can be used quantitatively 00:28:19.720 --> 00:28:23.512 align:middle line:84% as internal standards so that you can monitor this process. 00:28:23.512 --> 00:28:25.970 align:middle line:84% See, we're going to go into this in great detail, later on. 00:28:25.970 --> 00:28:27.340 align:middle line:84% But I wanted to give you a feeling for where 00:28:27.340 --> 00:28:28.632 align:middle line:90% the source of these things are. 00:28:28.632 --> 00:28:31.120 align:middle line:84% This array analysis, in a sense, is another example 00:28:31.120 --> 00:28:31.780 align:middle line:90% of microscopy. 00:28:31.780 --> 00:28:33.238 align:middle line:84% Just like in the previous slide, we 00:28:33.238 --> 00:28:34.840 align:middle line:84% used direct observation microscopy 00:28:34.840 --> 00:28:37.010 align:middle line:90% to monitor cell lineages. 00:28:37.010 --> 00:28:39.535 align:middle line:84% So too we can do it/ We can make a-- 00:28:39.535 --> 00:28:43.067 align:middle line:90% 00:28:43.067 --> 00:28:43.900 align:middle line:90% it's just wonderful. 00:28:43.900 --> 00:28:45.340 align:middle line:90% The battery's charged, OK. 00:28:45.340 --> 00:28:53.070 align:middle line:90% 00:28:53.070 --> 00:28:57.930 align:middle line:84% We can take the microscopy of artificial patterns, 00:28:57.930 --> 00:28:59.730 align:middle line:90% such as arrays. 00:28:59.730 --> 00:29:02.760 align:middle line:84% Just as we have separation here by mass, 00:29:02.760 --> 00:29:05.340 align:middle line:84% we can also have separation on a variety of other properties, 00:29:05.340 --> 00:29:08.040 align:middle line:84% sometimes called multidimensional separation. 00:29:08.040 --> 00:29:12.840 align:middle line:84% This gets back to the first slide of the lecture, which 00:29:12.840 --> 00:29:15.870 align:middle line:90% was the purification aspect. 00:29:15.870 --> 00:29:19.440 align:middle line:84% Now, how do we jump from that kind of raw data 00:29:19.440 --> 00:29:24.570 align:middle line:84% to this common way that biologists communicate 00:29:24.570 --> 00:29:27.240 align:middle line:84% in journals, where they have circles and arrows, where 00:29:27.240 --> 00:29:29.280 align:middle line:84% the circles might be some kind of protein 00:29:29.280 --> 00:29:33.300 align:middle line:84% molecule such as a stat, and arrow 00:29:33.300 --> 00:29:36.210 align:middle line:84% indicates some of interaction, or regulation, 00:29:36.210 --> 00:29:39.810 align:middle line:84% or quantitative influence that one protein 00:29:39.810 --> 00:29:43.770 align:middle line:90% has on another, or a protein. 00:29:43.770 --> 00:29:48.150 align:middle line:84% So in alternative diagrams, nodes could be small molecules, 00:29:48.150 --> 00:29:51.570 align:middle line:84% and the edges, the links between the nodes, 00:29:51.570 --> 00:29:54.900 align:middle line:84% could be an enzymatic reaction catalyzed by a protein. 00:29:54.900 --> 00:29:57.570 align:middle line:84% There are about 500 biological databases 00:29:57.570 --> 00:30:01.230 align:middle line:84% that we'll talk about in the database talk. 00:30:01.230 --> 00:30:04.230 align:middle line:84% How the data and models were entered into these databases, 00:30:04.230 --> 00:30:07.770 align:middle line:90% is a huge issue. 00:30:07.770 --> 00:30:10.050 align:middle line:84% Many of them have been done very casually. 00:30:10.050 --> 00:30:14.050 align:middle line:84% For DNA sequencing and crystallography, 00:30:14.050 --> 00:30:16.510 align:middle line:84% I think the process by which you go from the raw data 00:30:16.510 --> 00:30:19.000 align:middle line:84% to the models, is very well understood, very well 00:30:19.000 --> 00:30:21.520 align:middle line:84% communicated, for this sort of thing. 00:30:21.520 --> 00:30:24.820 align:middle line:84% It will take this whole course for us to even to scratch 00:30:24.820 --> 00:30:25.970 align:middle line:90% the surface. 00:30:25.970 --> 00:30:27.550 align:middle line:84% Here's another example-- that one was 00:30:27.550 --> 00:30:29.480 align:middle line:90% protein-protein interactions. 00:30:29.480 --> 00:30:32.530 align:middle line:84% This is an example where the nodes now are not 00:30:32.530 --> 00:30:34.240 align:middle line:90% proteins but small molecules. 00:30:34.240 --> 00:30:36.730 align:middle line:84% And they're connected by an enzymatic pathway. 00:30:36.730 --> 00:30:39.370 align:middle line:84% This is another example of an application 00:30:39.370 --> 00:30:41.380 align:middle line:84% of ordinary differentiable equations, 00:30:41.380 --> 00:30:43.630 align:middle line:90% just like the one last class. 00:30:43.630 --> 00:30:45.610 align:middle line:90% We had exponential growth. 00:30:45.610 --> 00:30:49.410 align:middle line:84% Here, you have simple fluxes, where 00:30:49.410 --> 00:30:54.010 align:middle line:84% a catalytic reaction occurs, not autocatalytic, but catalytic. 00:30:54.010 --> 00:30:56.470 align:middle line:84% There's no there's no exponential growth occurring 00:30:56.470 --> 00:30:58.600 align:middle line:90% in this cell. 00:30:58.600 --> 00:31:03.550 align:middle line:84% It doesn't have any biopolymers in it, biopolymer synthesis. 00:31:03.550 --> 00:31:07.000 align:middle line:84% But these catalytic reactions that form this network, 00:31:07.000 --> 00:31:11.260 align:middle line:84% and you can model the influx of fresh molecules. 00:31:11.260 --> 00:31:15.980 align:middle line:84% And its utilization within the cell and the efflux. 00:31:15.980 --> 00:31:17.510 align:middle line:90% We'll come back to that. 00:31:17.510 --> 00:31:20.810 align:middle line:84% Inside there, are a set of kinetic equations. 00:31:20.810 --> 00:31:23.150 align:middle line:84% We need to figure out how to get from the raw data types 00:31:23.150 --> 00:31:25.890 align:middle line:84% that I've shown you to this kind of equation. 00:31:25.890 --> 00:31:28.190 align:middle line:84% This will be one of the goals of the course. 00:31:28.190 --> 00:31:31.670 align:middle line:84% Here, you have a velocity on the far-left hand 00:31:31.670 --> 00:31:34.460 align:middle line:84% side of the top equation, which is 00:31:34.460 --> 00:31:39.230 align:middle line:84% related to a maximum velocity on the numerator. 00:31:39.230 --> 00:31:44.300 align:middle line:84% And then a series of linear sums and quotients. 00:31:44.300 --> 00:31:48.500 align:middle line:84% Now, some of the terms will be nonlinear. 00:31:48.500 --> 00:31:50.960 align:middle line:84% Here's an exponent of 4 that enters in, 00:31:50.960 --> 00:31:53.307 align:middle line:84% because you have one of these [? propertivities ?] that 00:31:53.307 --> 00:31:54.890 align:middle line:84% gives you that kind of sigmoidal curve 00:31:54.890 --> 00:31:57.260 align:middle line:84% that we showed for transistors and will enter 00:31:57.260 --> 00:31:59.810 align:middle line:84% to a number of biological consequences, 00:31:59.810 --> 00:32:02.180 align:middle line:84% where the steepness of that sigmoidal curve 00:32:02.180 --> 00:32:04.610 align:middle line:84% is determined by this exponent, sometimes called a Hill 00:32:04.610 --> 00:32:06.500 align:middle line:90% coefficient. 00:32:06.500 --> 00:32:10.580 align:middle line:84% But other than that, you'll get these simple, linear sums 00:32:10.580 --> 00:32:11.630 align:middle line:90% and quotients. 00:32:11.630 --> 00:32:14.120 align:middle line:90% And we'll come back to that. 00:32:14.120 --> 00:32:17.990 align:middle line:84% What actually constitutes these networks? 00:32:17.990 --> 00:32:21.860 align:middle line:84% I want you to feel less limited than you 00:32:21.860 --> 00:32:24.140 align:middle line:90% might get in a simple textbook. 00:32:24.140 --> 00:32:27.860 align:middle line:84% A simple, textbook definition of a catalytic enzyme-catalyzed 00:32:27.860 --> 00:32:31.820 align:middle line:84% process you might have, A is a substrate 00:32:31.820 --> 00:32:33.590 align:middle line:90% that turns into B as a product. 00:32:33.590 --> 00:32:36.920 align:middle line:84% This is a process that A could go to B spontaneously, 00:32:36.920 --> 00:32:39.197 align:middle line:84% but in the presence of enzyme, it goes faster. 00:32:39.197 --> 00:32:41.280 align:middle line:84% Or it could be, that for all intents and purposes, 00:32:41.280 --> 00:32:44.420 align:middle line:84% A never turns into B. It's so slow that you 00:32:44.420 --> 00:32:46.437 align:middle line:84% need this enzyme here to even detect it. 00:32:46.437 --> 00:32:48.020 align:middle line:84% The enzyme will form a complex with A. 00:32:48.020 --> 00:32:51.380 align:middle line:84% This could be a non-covalent complex or a covalent one. 00:32:51.380 --> 00:32:54.200 align:middle line:84% It then produces a covalent change 00:32:54.200 --> 00:32:59.120 align:middle line:84% in A. Ant it becomes an enzyme-bound B. B is released. 00:32:59.120 --> 00:33:01.940 align:middle line:90% Enzyme E is regenerated. 00:33:01.940 --> 00:33:05.330 align:middle line:84% And so in a certain sense, in this process of turning A to B, 00:33:05.330 --> 00:33:07.610 align:middle line:90% E is not consumed. 00:33:07.610 --> 00:33:11.390 align:middle line:84% But let's think about an increasingly important class 00:33:11.390 --> 00:33:14.300 align:middle line:84% of biochemistry, such as signal transduction, where 00:33:14.300 --> 00:33:16.250 align:middle line:90% the enzyme now has a new role. 00:33:16.250 --> 00:33:18.620 align:middle line:84% It changes places with the substrate. 00:33:18.620 --> 00:33:20.300 align:middle line:90% It becomes a substrate. 00:33:20.300 --> 00:33:25.280 align:middle line:84% The E now is a substrate in which a small molecule, ATP, 00:33:25.280 --> 00:33:31.370 align:middle line:84% which might have been the A up here, combines with the E. 00:33:31.370 --> 00:33:35.930 align:middle line:84% And the E could either catalyze its own phosphorylation 00:33:35.930 --> 00:33:38.910 align:middle line:84% or in context with another enzyme, but in any case, 00:33:38.910 --> 00:33:41.330 align:middle line:84% it becomes covalently modified to produce 00:33:41.330 --> 00:33:44.360 align:middle line:84% a phosphorylated enzyme, a phosphorylated protein. 00:33:44.360 --> 00:33:48.630 align:middle line:84% And then the ATP is regenerated by a simple enzymatic process. 00:33:48.630 --> 00:33:50.630 align:middle line:84% And so in a certain sense, formally, it's 00:33:50.630 --> 00:33:53.150 align:middle line:84% very similar to this process, except you now flip 00:33:53.150 --> 00:33:54.380 align:middle line:90% the enzyme and the substrate. 00:33:54.380 --> 00:33:58.070 align:middle line:84% ATP is not consumed, the small molecule is not consumed. 00:33:58.070 --> 00:33:59.450 align:middle line:90% The enzyme is consumed. 00:33:59.450 --> 00:34:03.140 align:middle line:84% So think of these things, these networks, as symmetrically 00:34:03.140 --> 00:34:04.280 align:middle line:90% as you can. 00:34:04.280 --> 00:34:09.248 align:middle line:84% Try not to get too embedded in the names, this is an enzyme, 00:34:09.248 --> 00:34:11.540 align:middle line:84% this is a substrate, and think more about the concepts. 00:34:11.540 --> 00:34:14.510 align:middle line:84% The concept here, is some things are consumed 00:34:14.510 --> 00:34:20.469 align:middle line:84% and some things are catalytic and regenerated. 00:34:20.469 --> 00:34:22.690 align:middle line:84% So again, we are going to integrate 00:34:22.690 --> 00:34:24.969 align:middle line:84% this metabolic processes we were talking about, 00:34:24.969 --> 00:34:28.570 align:middle line:84% in the last couple of slides, with the information 00:34:28.570 --> 00:34:33.370 align:middle line:84% flow, which was the topic of central dogma, 00:34:33.370 --> 00:34:37.150 align:middle line:84% in order to get functional genomics, 00:34:37.150 --> 00:34:41.679 align:middle line:84% which measures those information molecules mostly and produces 00:34:41.679 --> 00:34:43.730 align:middle line:90% quantitative modeling. 00:34:43.730 --> 00:34:45.909 align:middle line:84% You need to have the qualitative models to know 00:34:45.909 --> 00:34:47.409 align:middle line:90% what's connected to what. 00:34:47.409 --> 00:34:50.530 align:middle line:84% You need to have the raw data as illustrated in slide 41. 00:34:50.530 --> 00:34:54.050 align:middle line:84% Again, to remind you, the source of quantitative data 00:34:54.050 --> 00:34:57.940 align:middle line:84% here, you can measure RNA, or proteins, or peptides 00:34:57.940 --> 00:34:59.350 align:middle line:90% in the mass spectrometry. 00:34:59.350 --> 00:35:02.620 align:middle line:84% RNA in the arrays connected to the DNA 00:35:02.620 --> 00:35:05.410 align:middle line:90% provided by the DNA sequencing. 00:35:05.410 --> 00:35:11.560 align:middle line:84% I warned you that one of the gene ontology data type 00:35:11.560 --> 00:35:15.790 align:middle line:84% sources of data was sequenced-- electronic sequence annotation 00:35:15.790 --> 00:35:17.590 align:middle line:90% by sequence similarity. 00:35:17.590 --> 00:35:20.830 align:middle line:84% I want to elaborate on this warning 00:35:20.830 --> 00:35:27.460 align:middle line:84% with this slide, where we say, we have various justifications 00:35:27.460 --> 00:35:30.550 align:middle line:84% for looking for distant homologs, examples 00:35:30.550 --> 00:35:34.780 align:middle line:84% of gene products, which are related by, 00:35:34.780 --> 00:35:37.960 align:middle line:84% on that ultimate pedigree tree of life, 00:35:37.960 --> 00:35:39.710 align:middle line:90% by very long distances. 00:35:39.710 --> 00:35:41.770 align:middle line:84% It's been a long time since those things were 00:35:41.770 --> 00:35:43.667 align:middle line:90% present as a common ancestor. 00:35:43.667 --> 00:35:45.250 align:middle line:84% And we want to find those because they 00:35:45.250 --> 00:35:48.130 align:middle line:84% help us limit the number of hypotheses 00:35:48.130 --> 00:35:51.080 align:middle line:84% that we need to test whenever we find a new molecule. 00:35:51.080 --> 00:35:53.230 align:middle line:84% If we can connect it to another molecule, 00:35:53.230 --> 00:35:56.155 align:middle line:84% however distant, then we feel that we 00:35:56.155 --> 00:35:58.030 align:middle line:84% don't have to test every possible hypothesis. 00:35:58.030 --> 00:36:00.070 align:middle line:84% We just have to test that little narrow one. 00:36:00.070 --> 00:36:02.980 align:middle line:84% But what happens when we do that? 00:36:02.980 --> 00:36:04.990 align:middle line:84% Let's say, instead of some distant homology, 00:36:04.990 --> 00:36:07.930 align:middle line:84% where we have, say 20% amino acid identity. 00:36:07.930 --> 00:36:09.610 align:middle line:84% You line up the sequences by methods 00:36:09.610 --> 00:36:11.000 align:middle line:90% that we'll talk about later. 00:36:11.000 --> 00:36:13.570 align:middle line:84% And you have 20% of the positions 00:36:13.570 --> 00:36:19.100 align:middle line:84% that are the same, or even less, can sometimes be meaningful. 00:36:19.100 --> 00:36:20.710 align:middle line:90% But how good is that? 00:36:20.710 --> 00:36:22.630 align:middle line:84% There's going to be some kind of curve that 00:36:22.630 --> 00:36:26.860 align:middle line:84% relates how close two proteins are with a probability 00:36:26.860 --> 00:36:30.580 align:middle line:84% that they will have the same biochemical, or cell 00:36:30.580 --> 00:36:33.550 align:middle line:90% biological, or genetic function. 00:36:33.550 --> 00:36:35.620 align:middle line:84% And here's some worst case scenarios. 00:36:35.620 --> 00:36:37.810 align:middle line:84% And I don't mean to represent these as typical, 00:36:37.810 --> 00:36:40.360 align:middle line:84% but they get you doubting again so 00:36:40.360 --> 00:36:43.990 align:middle line:90% that you don't trust anything. 00:36:43.990 --> 00:36:45.460 align:middle line:90% 100% sequence identity. 00:36:45.460 --> 00:36:48.850 align:middle line:84% This should be a best case scenario, but it's not. 00:36:48.850 --> 00:36:52.600 align:middle line:84% The amylase enzyme, which catalyzes a carbon metabolism 00:36:52.600 --> 00:36:55.450 align:middle line:84% in most cells when it's expressed to high levels 00:36:55.450 --> 00:37:01.810 align:middle line:84% in a vertebrate like our friend this marine tortoise, 00:37:01.810 --> 00:37:04.592 align:middle line:84% turtle, it turns into the major eye lens protein. 00:37:04.592 --> 00:37:06.550 align:middle line:84% And actually, this is true of most vertebrates. 00:37:06.550 --> 00:37:09.520 align:middle line:84% They have some kind of enzyme, like a glycolic enzyme, 00:37:09.520 --> 00:37:12.700 align:middle line:84% which is overproduced and aggregates 00:37:12.700 --> 00:37:17.170 align:middle line:84% and makes a clear lens more morphologically 00:37:17.170 --> 00:37:19.690 align:middle line:84% interesting feature, which just focuses light. 00:37:19.690 --> 00:37:22.900 align:middle line:84% Completely new function by all those definitions of function. 00:37:22.900 --> 00:37:24.820 align:middle line:84% No longer does the enzymatic activity, 00:37:24.820 --> 00:37:27.130 align:middle line:84% does an optical activity instead. 00:37:27.130 --> 00:37:29.320 align:middle line:84% Another example, we have 100% sequence identity. 00:37:29.320 --> 00:37:32.500 align:middle line:84% Not something really distant homolog like 20% or 10%, 00:37:32.500 --> 00:37:34.480 align:middle line:90% but 100% sequence identity. 00:37:34.480 --> 00:37:37.120 align:middle line:84% [? Thyroxine, ?] which is involved in redox reactions 00:37:37.120 --> 00:37:39.640 align:middle line:84% involving [INAUDIBLE] and other things. 00:37:39.640 --> 00:37:41.740 align:middle line:84% In the right context with other proteins, 00:37:41.740 --> 00:37:44.110 align:middle line:84% it can now be part of a DNA polymerase, 00:37:44.110 --> 00:37:48.010 align:middle line:84% when it globs on the DNA, it goes really without stopping 00:37:48.010 --> 00:37:51.340 align:middle line:84% with [? thyroxine, ?] but it falls off if [? thyroxine ?] 00:37:51.340 --> 00:37:52.060 align:middle line:90% isn't around. 00:37:52.060 --> 00:37:54.400 align:middle line:84% That's not a redox function completely different 00:37:54.400 --> 00:37:56.927 align:middle line:90% biochemical function. 00:37:56.927 --> 00:37:58.510 align:middle line:84% But like I say, there will be a curve. 00:37:58.510 --> 00:38:03.580 align:middle line:84% Sometimes, there will be very great hypothesis 00:38:03.580 --> 00:38:07.150 align:middle line:84% limitation that can come from very distant relatives. 00:38:07.150 --> 00:38:09.840 align:middle line:90% 00:38:09.840 --> 00:38:15.750 align:middle line:84% These are more examples of the quantitative data 00:38:15.750 --> 00:38:19.380 align:middle line:84% that we will use to get hints at relationships among genes 00:38:19.380 --> 00:38:21.330 align:middle line:90% that go up and down together. 00:38:21.330 --> 00:38:24.680 align:middle line:90% 00:38:24.680 --> 00:38:28.090 align:middle line:84% They form the basis of asking, what 00:38:28.090 --> 00:38:30.820 align:middle line:84% is function not based just on sequence homology, 00:38:30.820 --> 00:38:33.940 align:middle line:84% but based on a variety of quantitative data, 00:38:33.940 --> 00:38:38.170 align:middle line:84% such as the RNA data and the microarrays. 00:38:38.170 --> 00:38:42.160 align:middle line:84% This is three more ways of looking 00:38:42.160 --> 00:38:44.260 align:middle line:90% at how we define functions. 00:38:44.260 --> 00:38:46.810 align:middle line:84% Function definition number one, is the effects 00:38:46.810 --> 00:38:47.890 align:middle line:90% of mutation on fitness. 00:38:47.890 --> 00:38:49.540 align:middle line:84% This is, in a certain sense, what 00:38:49.540 --> 00:38:54.640 align:middle line:84% the organism cares about the function of a gene product. 00:38:54.640 --> 00:38:58.330 align:middle line:84% It's how many grandchildren am I going to have? 00:38:58.330 --> 00:38:59.960 align:middle line:90% That's what it cares about. 00:38:59.960 --> 00:39:02.662 align:middle line:84% And that's what shaped the function over time. 00:39:02.662 --> 00:39:04.120 align:middle line:84% And so if we're going to understand 00:39:04.120 --> 00:39:06.520 align:middle line:84% any of our other definition functions, 00:39:06.520 --> 00:39:09.880 align:middle line:84% we have to at least give some attention 00:39:09.880 --> 00:39:12.790 align:middle line:84% to what shaped it over billions of years 00:39:12.790 --> 00:39:14.410 align:middle line:84% and over many different environments. 00:39:14.410 --> 00:39:18.790 align:middle line:84% We need to have some feeling for the ecology of these organisms. 00:39:18.790 --> 00:39:22.390 align:middle line:84% The second definition is the more commonly used one, 00:39:22.390 --> 00:39:25.750 align:middle line:84% which is what is actually its function 00:39:25.750 --> 00:39:27.970 align:middle line:90% in a machine-like sense. 00:39:27.970 --> 00:39:32.135 align:middle line:84% In the cogs, in the wheels, how does it function structurally? 00:39:32.135 --> 00:39:33.760 align:middle line:84% What's the three-dimensional structure? 00:39:33.760 --> 00:39:35.350 align:middle line:90% What's the mechanism? 00:39:35.350 --> 00:39:38.470 align:middle line:84% The third function, is more forward-looking, 00:39:38.470 --> 00:39:41.770 align:middle line:84% not what good has it been to organisms in the past, 00:39:41.770 --> 00:39:43.810 align:middle line:84% but what good can it be to us in the future 00:39:43.810 --> 00:39:45.940 align:middle line:84% or to other organisms in the future? 00:39:45.940 --> 00:39:49.870 align:middle line:84% This may not involve reproducing the organism, 00:39:49.870 --> 00:39:50.710 align:middle line:90% making copies of it. 00:39:50.710 --> 00:39:53.320 align:middle line:84% It could be that there's some other engineering goal 00:39:53.320 --> 00:39:56.050 align:middle line:90% or objective function. 00:39:56.050 --> 00:39:58.120 align:middle line:84% When we say that we've proven something. 00:39:58.120 --> 00:40:01.930 align:middle line:84% We've proven a biological hypothesis. 00:40:01.930 --> 00:40:04.510 align:middle line:84% What we mean is, given the assumption, 00:40:04.510 --> 00:40:06.580 align:middle line:84% it's a statistical statement that the odds 00:40:06.580 --> 00:40:09.070 align:middle line:84% of the hypothesis being wrong are less than 5% 00:40:09.070 --> 00:40:13.240 align:middle line:84% of the time keeping in mind hidden hypotheses 00:40:13.240 --> 00:40:14.260 align:middle line:90% and multiple hypotheses. 00:40:14.260 --> 00:40:16.810 align:middle line:84% In genomics, it's all too easy to collect a lot of data, 00:40:16.810 --> 00:40:19.090 align:middle line:84% and therefore, when you mine the data, 00:40:19.090 --> 00:40:20.920 align:middle line:84% you can make a lot of hypotheses. 00:40:20.920 --> 00:40:23.950 align:middle line:84% And you test them, and you find you will find thousands 00:40:23.950 --> 00:40:25.540 align:middle line:84% of things, which by themselves would 00:40:25.540 --> 00:40:30.670 align:middle line:84% be significant at the 5% level, the standard statistical test, 00:40:30.670 --> 00:40:34.810 align:middle line:84% but you've got to correct for the number of hypotheses 00:40:34.810 --> 00:40:37.210 align:middle line:84% you implicitly or explicitly test. 00:40:37.210 --> 00:40:40.492 align:middle line:84% We'll mention this time and again in specific cases 00:40:40.492 --> 00:40:41.200 align:middle line:90% as we go forward. 00:40:41.200 --> 00:40:43.770 align:middle line:90% 00:40:43.770 --> 00:40:48.600 align:middle line:84% The systems biology manifesto that I mentioned earlier, 00:40:48.600 --> 00:40:51.150 align:middle line:84% had this little loop where you would generate perturbations 00:40:51.150 --> 00:40:52.660 align:middle line:90% and test things and so forth. 00:40:52.660 --> 00:40:54.960 align:middle line:84% But an alternative way rather than doing 00:40:54.960 --> 00:40:57.240 align:middle line:84% additional experiments, is if you really 00:40:57.240 --> 00:40:59.287 align:middle line:84% have bought-in fully into systems biology 00:40:59.287 --> 00:41:00.870 align:middle line:84% and you really have all the components 00:41:00.870 --> 00:41:03.210 align:middle line:84% and a systematic perturbations, then 00:41:03.210 --> 00:41:05.580 align:middle line:84% you might be able to test the hypothesis generated 00:41:05.580 --> 00:41:08.520 align:middle line:84% by data mining one data set by going into another data set. 00:41:08.520 --> 00:41:10.780 align:middle line:84% You need to ensure that they are independent. 00:41:10.780 --> 00:41:15.643 align:middle line:84% And you need to ensure that the hypothesis itself, came 00:41:15.643 --> 00:41:18.060 align:middle line:84% from the first data set and not the second when you go out 00:41:18.060 --> 00:41:18.640 align:middle line:90% and test it. 00:41:18.640 --> 00:41:21.990 align:middle line:84% But that would be a pure data mining loop, systems biology 00:41:21.990 --> 00:41:22.890 align:middle line:90% loop. 00:41:22.890 --> 00:41:29.970 align:middle line:84% Now, just like when we say we have a proof, 00:41:29.970 --> 00:41:31.920 align:middle line:84% you should be distrusting of anybody that 00:41:31.920 --> 00:41:33.180 align:middle line:90% says I have an absolute proof. 00:41:33.180 --> 00:41:36.220 align:middle line:84% What they really mean, is a statistical statement. 00:41:36.220 --> 00:41:38.352 align:middle line:84% So too, when someone says, when they 00:41:38.352 --> 00:41:39.810 align:middle line:84% refer to the quality of their data, 00:41:39.810 --> 00:41:45.210 align:middle line:84% is this is the answer at the raw data level, what they really 00:41:45.210 --> 00:41:47.520 align:middle line:84% mean, is that they have some error 00:41:47.520 --> 00:41:49.950 align:middle line:90% level that they can quantitate. 00:41:49.950 --> 00:41:52.770 align:middle line:84% And you should be especially distrustful 00:41:52.770 --> 00:41:57.750 align:middle line:84% if someone doesn't attempt to give you any feeling for that. 00:41:57.750 --> 00:42:00.600 align:middle line:84% Not to say that everybody that gives error bars or error 00:42:00.600 --> 00:42:04.300 align:middle line:84% estimates is to be trusted, but you get the idea. 00:42:04.300 --> 00:42:07.410 align:middle line:84% So for DNA sequencing, there's a standard of practice. 00:42:07.410 --> 00:42:12.390 align:middle line:84% It was not always such, but a meeting in Bermuda, it's 00:42:12.390 --> 00:42:14.130 align:middle line:84% called the Bermuda standard, this 00:42:14.130 --> 00:42:23.100 align:middle line:84% is the best place to establish standards, is 99.99% accurate. 00:42:23.100 --> 00:42:25.620 align:middle line:84% You can see they have very high standards in Bermuda. 00:42:25.620 --> 00:42:28.350 align:middle line:84% But that's across the Genome Project. 00:42:28.350 --> 00:42:31.540 align:middle line:90% 00:42:31.540 --> 00:42:35.250 align:middle line:84% These are aspects that, I think, we got from genomics, 00:42:35.250 --> 00:42:39.310 align:middle line:84% in addition to the raw data, we've got kind of an attitude. 00:42:39.310 --> 00:42:42.900 align:middle line:84% The attitude is, we can start looking at whole systems again, 00:42:42.900 --> 00:42:47.040 align:middle line:84% less on the individual gene-hypothesis driven 00:42:47.040 --> 00:42:51.120 align:middle line:84% standard NIH grant proposals that predated the Genome 00:42:51.120 --> 00:42:51.690 align:middle line:90% Project. 00:42:51.690 --> 00:42:54.300 align:middle line:84% Now you can do less hypothesis driven, 00:42:54.300 --> 00:42:58.090 align:middle line:84% you can do data mining, and so on. 00:42:58.090 --> 00:43:02.130 align:middle line:84% We've also inherited the concept of automation the modeling 00:43:02.130 --> 00:43:03.450 align:middle line:90% and completion. 00:43:03.450 --> 00:43:06.447 align:middle line:84% Completion is something which still is not 00:43:06.447 --> 00:43:08.280 align:middle line:84% reduced to practice for functional genomics, 00:43:08.280 --> 00:43:11.050 align:middle line:84% but it has been reduced to practice for sequencing. 00:43:11.050 --> 00:43:13.890 align:middle line:84% And there is hope that we can approach it 00:43:13.890 --> 00:43:15.510 align:middle line:90% for functional genomics. 00:43:15.510 --> 00:43:17.910 align:middle line:84% Be careful using the word impossible. 00:43:17.910 --> 00:43:19.710 align:middle line:84% There certainly are things that appear 00:43:19.710 --> 00:43:21.930 align:middle line:84% not to be cost effective at any given moment, 00:43:21.930 --> 00:43:24.780 align:middle line:84% but technology is moving quickly enough. 00:43:24.780 --> 00:43:27.510 align:middle line:84% Remember those greater than exponential curves 00:43:27.510 --> 00:43:29.490 align:middle line:90% in the last lecture. 00:43:29.490 --> 00:43:33.600 align:middle line:84% There are technologies arriving that make things suddenly 00:43:33.600 --> 00:43:36.340 align:middle line:90% become cost effective. 00:43:36.340 --> 00:43:40.320 align:middle line:84% And that's a particularly important warning when you're 00:43:40.320 --> 00:43:42.390 align:middle line:84% designing a computational method that 00:43:42.390 --> 00:43:45.190 align:middle line:84% will compete with an experimental method, 00:43:45.190 --> 00:43:47.920 align:middle line:84% the experimental method suddenly becomes cost effective, 00:43:47.920 --> 00:43:53.310 align:middle line:84% then you need to revise your computational goals. 00:43:53.310 --> 00:43:56.880 align:middle line:84% We have types of mutations that we've talked about. 00:43:56.880 --> 00:43:59.970 align:middle line:84% We have a null mutation, for example, phenylketonuria, 00:43:59.970 --> 00:44:05.080 align:middle line:84% which is tested in newborns, in almost all newborns, 00:44:05.080 --> 00:44:07.080 align:middle line:84% that are born in the United States and certainly 00:44:07.080 --> 00:44:09.750 align:middle line:90% Massachusetts. 00:44:09.750 --> 00:44:13.050 align:middle line:84% This is a very serious source of mental retardation 00:44:13.050 --> 00:44:15.120 align:middle line:90% completely wiping out that gene. 00:44:15.120 --> 00:44:18.630 align:middle line:84% Small dosage effects, like a 1.5-fold effect that we talked 00:44:18.630 --> 00:44:22.330 align:middle line:84% about in trisomies, like Down syndrome, are important. 00:44:22.330 --> 00:44:23.610 align:middle line:90% You have conditional mutants. 00:44:23.610 --> 00:44:27.150 align:middle line:84% Classically temperature sensitivity of a mutation, 00:44:27.150 --> 00:44:28.770 align:middle line:90% meaning the protein unfolds. 00:44:28.770 --> 00:44:34.560 align:middle line:84% Or more recent enthusiasm for chemicals, 00:44:34.560 --> 00:44:38.850 align:middle line:84% a mutation which depends upon a chemical for producing 00:44:38.850 --> 00:44:40.135 align:middle line:90% its phenotype. 00:44:40.135 --> 00:44:41.760 align:middle line:84% You can not only have these things that 00:44:41.760 --> 00:44:48.870 align:middle line:84% affect dosage or condition or complete knockout, 00:44:48.870 --> 00:44:51.060 align:middle line:84% you can have a new function to obtain 00:44:51.060 --> 00:44:52.620 align:middle line:84% for changing the ligand specificity 00:44:52.620 --> 00:44:54.900 align:middle line:84% or changing the aggregation of a protein. 00:44:54.900 --> 00:44:58.800 align:middle line:84% Here in the background, are how a change in the hemoglobin, 00:44:58.800 --> 00:45:00.270 align:middle line:84% which normally transports oxygen, 00:45:00.270 --> 00:45:04.320 align:middle line:84% can change the morphology of a cell and hence the function 00:45:04.320 --> 00:45:06.720 align:middle line:90% in transporting oxygen. 00:45:06.720 --> 00:45:09.870 align:middle line:84% I just want to end on two slides on how 00:45:09.870 --> 00:45:17.670 align:middle line:84% you can represent the competition among cells 00:45:17.670 --> 00:45:20.430 align:middle line:84% or among organisms, which represents 00:45:20.430 --> 00:45:24.060 align:middle line:84% the Darwinian function, function number one, a few slides back. 00:45:24.060 --> 00:45:27.360 align:middle line:84% Here you have mutants in a population. 00:45:27.360 --> 00:45:30.090 align:middle line:84% Selection acts on populations, and mutations 00:45:30.090 --> 00:45:33.930 align:middle line:84% are tagged, by definition, by their nucleic acid. 00:45:33.930 --> 00:45:42.300 align:middle line:84% If you can use the tags, you make a pool of such mutants, 00:45:42.300 --> 00:45:44.263 align:middle line:84% these are naturally occurring population. 00:45:44.263 --> 00:45:45.680 align:middle line:84% And when these pools are subjected 00:45:45.680 --> 00:45:49.340 align:middle line:84% to selections natural, or complex, or simple, 00:45:49.340 --> 00:45:54.560 align:middle line:84% or in the laboratory you can now read out these tags in many 00:45:54.560 --> 00:45:57.590 align:middle line:84% of the quantitative ways we talked about, for instance, 00:45:57.590 --> 00:46:00.450 align:middle line:84% mass spectrometry, arrays, so on. 00:46:00.450 --> 00:46:06.433 align:middle line:84% And as you go through more rounds of selection, 00:46:06.433 --> 00:46:07.850 align:middle line:84% you'll eventually pick the winner, 00:46:07.850 --> 00:46:10.220 align:middle line:84% which is the most highly selected 00:46:10.220 --> 00:46:14.960 align:middle line:90% of the mutations, the winner. 00:46:14.960 --> 00:46:16.383 align:middle line:84% Or you might have a mixture if you 00:46:16.383 --> 00:46:18.050 align:middle line:84% through a very limited number of rounds. 00:46:18.050 --> 00:46:19.970 align:middle line:84% This will follow the exponential curve 00:46:19.970 --> 00:46:22.430 align:middle line:84% that we had here, whether it's exponential decay 00:46:22.430 --> 00:46:23.690 align:middle line:90% or exponential growth. 00:46:23.690 --> 00:46:26.420 align:middle line:84% You can have a very subtle difference in growth, 00:46:26.420 --> 00:46:31.100 align:middle line:84% due to the function of that mutated gene product, 00:46:31.100 --> 00:46:36.230 align:middle line:84% but that small, say 1% turns into complete 00:46:36.230 --> 00:46:39.887 align:middle line:84% all or none replacement if you have enough generations. 00:46:39.887 --> 00:46:41.720 align:middle line:84% This is the awesome power of the exponential 00:46:41.720 --> 00:46:43.025 align:middle line:90% that we talked about last time. 00:46:43.025 --> 00:46:45.710 align:middle line:90% 00:46:45.710 --> 00:46:50.030 align:middle line:84% And in real world, and also in the laboratory, 00:46:50.030 --> 00:46:53.060 align:middle line:84% you can think of this as going over a variety of environments, 00:46:53.060 --> 00:46:55.130 align:middle line:90% E, over different times. 00:46:55.130 --> 00:47:00.050 align:middle line:84% So the time you spend in each of these different environments, 00:47:00.050 --> 00:47:03.710 align:middle line:90% has some unit that happens. 00:47:03.710 --> 00:47:06.230 align:middle line:84% In natural environment, you'll spend say, more time 00:47:06.230 --> 00:47:07.950 align:middle line:90% in one condition than another. 00:47:07.950 --> 00:47:11.090 align:middle line:84% And the selection coefficients are a simple sum, 00:47:11.090 --> 00:47:14.990 align:middle line:84% and this exponential gives you the ratio of the organisms. 00:47:14.990 --> 00:47:17.310 align:middle line:90% Here's some references on this. 00:47:17.310 --> 00:47:21.140 align:middle line:84% And I urge you to take a look at these, where actual experiments 00:47:21.140 --> 00:47:22.910 align:middle line:90% have been done getting these. 00:47:22.910 --> 00:47:26.760 align:middle line:84% And we'll come back to this later in the course. 00:47:26.760 --> 00:47:31.070 align:middle line:84% So this is the end of this lecture number two. 00:47:31.070 --> 00:47:33.085 align:middle line:90% Thank you very much. 00:47:33.085 --> 00:48:14.000 align:middle line:90%