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 --> 00:00:06.370 align:middle line:84% by MIT OpenCourseWare under a Creative Commons license. 00:00:06.370 --> 00:00:08.470 align:middle line:84% Additional information about our license 00:00:08.470 --> 00:00:10.870 align:middle line:84% and MIT OpenCourseWare in general 00:00:10.870 --> 00:00:13.090 align:middle line:90% is available at ocu.mit.edu. 00:00:13.090 --> 00:00:32.310 align:middle line:90% 00:00:32.310 --> 00:00:33.060 align:middle line:90% GEORGE CHURCH: OK. 00:00:33.060 --> 00:00:34.290 align:middle line:90% We're getting ready. 00:00:34.290 --> 00:00:36.710 align:middle line:90% And let's go. 00:00:36.710 --> 00:00:38.340 align:middle line:90% OK. 00:00:38.340 --> 00:00:45.090 align:middle line:84% So welcome to the first class of BIOE101 00:00:45.090 --> 00:00:51.170 align:middle line:84% or biophysics 101 or HST.508 or Genetics 224. 00:00:51.170 --> 00:00:55.440 align:middle line:84% You can see that this is a truly interdisciplinary class, 00:00:55.440 --> 00:00:58.320 align:middle line:84% and I hope that's going to be one of its strengths. 00:00:58.320 --> 00:00:59.790 align:middle line:90% It has been in the past. 00:00:59.790 --> 00:01:03.900 align:middle line:84% I hope that you get to know some of your colleagues in this room 00:01:03.900 --> 00:01:06.660 align:middle line:84% or maybe in the other half of this class, which 00:01:06.660 --> 00:01:10.260 align:middle line:84% is taught on the medical school campus, 00:01:10.260 --> 00:01:12.600 align:middle line:84% because they could be some of your greatest assets 00:01:12.600 --> 00:01:14.440 align:middle line:90% in years to come. 00:01:14.440 --> 00:01:19.290 align:middle line:84% This course-- basically, if you cannot, for some reason, 00:01:19.290 --> 00:01:21.780 align:middle line:84% on a particular day, make this meeting at 5:30, 00:01:21.780 --> 00:01:26.130 align:middle line:84% you have the option of going at noon over in the Canon room, 00:01:26.130 --> 00:01:31.260 align:middle line:84% or you have the option of tuning in to the video, 00:01:31.260 --> 00:01:34.470 align:middle line:84% which is intended mainly for the distance education students 00:01:34.470 --> 00:01:37.770 align:middle line:84% but also serves as a backup in case one of you gets sick or is 00:01:37.770 --> 00:01:39.310 align:middle line:84% out of town or something like that. 00:01:39.310 --> 00:01:42.600 align:middle line:84% So that's what's going on right there, 00:01:42.600 --> 00:01:47.340 align:middle line:84% is that we put this on our internet site. 00:01:47.340 --> 00:01:50.160 align:middle line:84% It's synchronized to the PowerPoint slides. 00:01:50.160 --> 00:01:54.240 align:middle line:84% You should have PowerPoint handouts right now. 00:01:54.240 --> 00:01:58.080 align:middle line:84% This course, as stated on the website 00:01:58.080 --> 00:02:03.330 align:middle line:84% and on this first slide, is based, essentially, 00:02:03.330 --> 00:02:06.690 align:middle line:84% entirely on six problem sets and a course project. 00:02:06.690 --> 00:02:13.050 align:middle line:84% In the past, the students and I have had a great time 00:02:13.050 --> 00:02:14.430 align:middle line:90% in the course projects. 00:02:14.430 --> 00:02:18.672 align:middle line:84% We see, really, the cutting edge of computational biology. 00:02:18.672 --> 00:02:20.130 align:middle line:84% And the six problem sets are really 00:02:20.130 --> 00:02:24.180 align:middle line:84% intended to get you up to speed so that you are as close as 00:02:24.180 --> 00:02:28.200 align:middle line:84% possible to a publication-grade computational biology 00:02:28.200 --> 00:02:31.020 align:middle line:84% project, which, generally speaking, 00:02:31.020 --> 00:02:35.790 align:middle line:84% is collaborative, involving two or more people. 00:02:35.790 --> 00:02:38.070 align:middle line:90% But you can do it solo as well. 00:02:38.070 --> 00:02:40.620 align:middle line:84% There'll be more details on the website 00:02:40.620 --> 00:02:44.400 align:middle line:90% and also in your sections. 00:02:44.400 --> 00:02:46.950 align:middle line:84% You're also free to work with other people 00:02:46.950 --> 00:02:49.950 align:middle line:84% to use any resources you want for the problem sets. 00:02:49.950 --> 00:02:52.800 align:middle line:90% I would just ask that you-- 00:02:52.800 --> 00:02:54.630 align:middle line:84% however you answer the question, whether it 00:02:54.630 --> 00:02:57.030 align:middle line:84% comes just purely from your head or from collaboration 00:02:57.030 --> 00:03:01.380 align:middle line:84% or from some website, that you just say for each problem, 00:03:01.380 --> 00:03:04.270 align:middle line:84% just briefly, where you got the answer. 00:03:04.270 --> 00:03:07.470 align:middle line:84% This is just a good academic discipline 00:03:07.470 --> 00:03:09.900 align:middle line:90% for acknowledging your sources. 00:03:09.900 --> 00:03:13.650 align:middle line:90% 00:03:13.650 --> 00:03:16.140 align:middle line:84% This course is intended to be an introductory course, 00:03:16.140 --> 00:03:21.180 align:middle line:84% introductory to almost every subject that it combines. 00:03:21.180 --> 00:03:23.670 align:middle line:90% It has minor prerequisites. 00:03:23.670 --> 00:03:26.010 align:middle line:84% And those who feel that they are a little 00:03:26.010 --> 00:03:28.650 align:middle line:84% weak in either molecular biology, statistics, 00:03:28.650 --> 00:03:31.800 align:middle line:84% or computing, there will be classes 00:03:31.800 --> 00:03:34.380 align:middle line:84% in addition to the regular sections which 00:03:34.380 --> 00:03:37.683 align:middle line:84% will address these intensively, the especially at 00:03:37.683 --> 00:03:38.850 align:middle line:90% the beginning of the course. 00:03:38.850 --> 00:03:40.320 align:middle line:84% Then the sections themselves will 00:03:40.320 --> 00:03:42.570 align:middle line:84% be crafted so that they are directed 00:03:42.570 --> 00:03:46.678 align:middle line:84% at people who feel strong in biology and slightly weaker 00:03:46.678 --> 00:03:47.220 align:middle line:90% in computing. 00:03:47.220 --> 00:03:50.700 align:middle line:84% So the focus will be on getting the computing up from-- 00:03:50.700 --> 00:03:51.900 align:middle line:90% more or less from scratch. 00:03:51.900 --> 00:03:55.110 align:middle line:84% And those that are stronger, say, in computing, 00:03:55.110 --> 00:03:59.580 align:middle line:84% then there'll be sections on advanced topics 00:03:59.580 --> 00:04:02.070 align:middle line:84% which will fit in with the course. 00:04:02.070 --> 00:04:05.250 align:middle line:84% By the middle of the course, almost all the sections 00:04:05.250 --> 00:04:06.870 align:middle line:90% will be very close to identical. 00:04:06.870 --> 00:04:12.030 align:middle line:84% But there still will be that slight difference. 00:04:12.030 --> 00:04:13.740 align:middle line:84% We will try to provide mechanisms 00:04:13.740 --> 00:04:16.860 align:middle line:84% by which you can interact with colleagues that 00:04:16.860 --> 00:04:20.730 align:middle line:84% have different strengths but similar interests 00:04:20.730 --> 00:04:26.580 align:middle line:84% so that you can get together on problem sets and the projects. 00:04:26.580 --> 00:04:28.080 align:middle line:84% You shouldn't feel it was essential. 00:04:28.080 --> 00:04:31.665 align:middle line:84% It's an opportunity, not an obligation. 00:04:31.665 --> 00:04:33.540 align:middle line:84% It's important you hand in your questionnaire 00:04:33.540 --> 00:04:38.670 align:middle line:84% immediately after class so that we can assign you the sections. 00:04:38.670 --> 00:04:42.780 align:middle line:84% The sections will-- the sections will be the way 00:04:42.780 --> 00:04:44.910 align:middle line:84% that you get information and the way 00:04:44.910 --> 00:04:47.550 align:middle line:84% that you interact with the teaching 00:04:47.550 --> 00:04:50.470 align:middle line:84% fellow who will be responsible for grading your problem sets. 00:04:50.470 --> 00:04:53.700 align:middle line:84% So that teaching fellow will interact with you 00:04:53.700 --> 00:04:56.170 align:middle line:84% on all six problem sets in your project, 00:04:56.170 --> 00:04:59.010 align:middle line:84% and so you should get to know that person very well. 00:04:59.010 --> 00:05:00.600 align:middle line:84% And I'm very indebted to this crew, 00:05:00.600 --> 00:05:03.750 align:middle line:84% by the way, which is up here at the top. 00:05:03.750 --> 00:05:07.890 align:middle line:84% Suzanne [? Camilli ?] is the head teaching fellow this year. 00:05:07.890 --> 00:05:11.400 align:middle line:84% Many of them-- she was a teaching fellow last year. 00:05:11.400 --> 00:05:12.750 align:middle line:90% And many of these-- 00:05:12.750 --> 00:05:14.250 align:middle line:84% almost all of these teaching fellows 00:05:14.250 --> 00:05:16.995 align:middle line:84% have taken this course last year. 00:05:16.995 --> 00:05:21.020 align:middle line:90% 00:05:21.020 --> 00:05:24.920 align:middle line:84% That's most of the bureaucratic aspects of it 00:05:24.920 --> 00:05:26.640 align:middle line:90% unless there are some questions. 00:05:26.640 --> 00:05:28.610 align:middle line:84% Please feel free to interrupt at any time. 00:05:28.610 --> 00:05:32.450 align:middle line:84% This can be interactive if you want it to be. 00:05:32.450 --> 00:05:35.450 align:middle line:84% I can set aside enough time for that. 00:05:35.450 --> 00:05:39.510 align:middle line:84% You're certainly also welcome to come to me before class, 00:05:39.510 --> 00:05:41.090 align:middle line:84% after class, and during the break, 00:05:41.090 --> 00:05:43.280 align:middle line:84% and we can even set up additional times. 00:05:43.280 --> 00:05:45.800 align:middle line:90% 00:05:45.800 --> 00:05:47.560 align:middle line:90% Questions? 00:05:47.560 --> 00:05:48.170 align:middle line:90% Yes. 00:05:48.170 --> 00:05:50.080 align:middle line:84% AUDIENCE: The extra sections, those are more 00:05:50.080 --> 00:05:51.313 align:middle line:90% filling in the background. 00:05:51.313 --> 00:05:52.480 align:middle line:90% GEORGE CHURCH: That's right. 00:05:52.480 --> 00:05:54.040 align:middle line:90% There will be this-- 00:05:54.040 --> 00:05:57.418 align:middle line:84% AUDIENCE: There seems to be only one, which is Thursday evening. 00:05:57.418 --> 00:05:59.710 align:middle line:84% GEORGE CHURCH: So the question is about extra sessions. 00:05:59.710 --> 00:06:05.290 align:middle line:84% There will be a schedule of all the sections evolving 00:06:05.290 --> 00:06:08.480 align:middle line:84% on the website within the next day or two. 00:06:08.480 --> 00:06:10.870 align:middle line:84% Depending on what your questionnaires show up, 00:06:10.870 --> 00:06:13.510 align:middle line:84% there will probably be three or four extra sections 00:06:13.510 --> 00:06:17.020 align:middle line:84% within the first couple of weeks. 00:06:17.020 --> 00:06:20.290 align:middle line:84% AUDIENCE: And how do we [INAUDIBLE]?? 00:06:20.290 --> 00:06:22.630 align:middle line:84% GEORGE CHURCH: Through your section head. 00:06:22.630 --> 00:06:25.300 align:middle line:84% If you put your email address, it 00:06:25.300 --> 00:06:29.740 align:middle line:84% will allow the section head teaching fellow to contact you, 00:06:29.740 --> 00:06:36.260 align:middle line:84% or you can contact any of them if no one contacts you. 00:06:36.260 --> 00:06:36.970 align:middle line:90% OK. 00:06:36.970 --> 00:06:40.550 align:middle line:90% So the overview for the course-- 00:06:40.550 --> 00:06:47.410 align:middle line:84% this is constrained to fit into the calendar of both Harvard 00:06:47.410 --> 00:06:48.700 align:middle line:90% and MIT. 00:06:48.700 --> 00:06:50.890 align:middle line:84% It also tries to achieve the goal 00:06:50.890 --> 00:06:54.490 align:middle line:84% of keeping the Division of Continuing Education students 00:06:54.490 --> 00:06:59.180 align:middle line:84% so that they only have to come in for one day a week, 00:06:59.180 --> 00:07:02.170 align:middle line:84% which is a considerable plus section. 00:07:02.170 --> 00:07:06.340 align:middle line:84% Some of them choose to have section the same day. 00:07:06.340 --> 00:07:09.170 align:middle line:84% And some of them have considerable commutes. 00:07:09.170 --> 00:07:13.540 align:middle line:84% So that determines the timing of this. 00:07:13.540 --> 00:07:17.560 align:middle line:84% The topics-- we'll have two introductory lectures. 00:07:17.560 --> 00:07:20.980 align:middle line:84% We'll actually cover pretty interesting topics 00:07:20.980 --> 00:07:25.750 align:middle line:84% for introductory thematically, one on computing 00:07:25.750 --> 00:07:27.910 align:middle line:84% followed by one on biology, although both 00:07:27.910 --> 00:07:30.580 align:middle line:84% of them and the whole course is on computational biology 00:07:30.580 --> 00:07:32.410 align:middle line:90% and systems biology. 00:07:32.410 --> 00:07:35.410 align:middle line:84% Then as another way of focusing, we'll 00:07:35.410 --> 00:07:38.770 align:middle line:84% have a series of six lectures, two on DNA, two on RNA, 00:07:38.770 --> 00:07:42.760 align:middle line:84% and two on proteins or proteomics, which 00:07:42.760 --> 00:07:45.610 align:middle line:84% reflects the central dogma of molecular biology 00:07:45.610 --> 00:07:48.310 align:middle line:84% where DNA encodes RNA which encodes proteins. 00:07:48.310 --> 00:07:50.350 align:middle line:84% But much more so, it's telling us 00:07:50.350 --> 00:07:52.810 align:middle line:84% about the flow of information that we 00:07:52.810 --> 00:07:55.870 align:middle line:84% need to establish experimentally and in terms of modeling that 00:07:55.870 --> 00:07:58.300 align:middle line:84% allows us to do functional genomics, 00:07:58.300 --> 00:08:04.000 align:middle line:84% and hence, systems biology, and to mine 00:08:04.000 --> 00:08:08.650 align:middle line:84% the datasets which are on everyone's attention right now. 00:08:08.650 --> 00:08:14.650 align:middle line:84% Then finally, three topics which really 00:08:14.650 --> 00:08:17.530 align:middle line:84% focus on the overarching theme of the whole course, which 00:08:17.530 --> 00:08:21.460 align:middle line:84% is on networks and systems biology. 00:08:21.460 --> 00:08:26.500 align:middle line:84% And these really cover the gamut of network analysis 00:08:26.500 --> 00:08:33.820 align:middle line:84% on cellular all the way through ecological scale modeling 00:08:33.820 --> 00:08:37.510 align:middle line:84% in order to see how we integrate various data types. 00:08:37.510 --> 00:08:40.929 align:middle line:84% Then comes the very interesting section 00:08:40.929 --> 00:08:45.100 align:middle line:84% of the course, three of these two-hour sessions 00:08:45.100 --> 00:08:50.920 align:middle line:84% where you get to present to me the topics that you're 00:08:50.920 --> 00:08:53.360 align:middle line:84% excited about, how you have taken 00:08:53.360 --> 00:08:57.940 align:middle line:84% the course material problem sets and so forth 00:08:57.940 --> 00:09:01.930 align:middle line:84% and incorporated into your view of what's exciting 00:09:01.930 --> 00:09:05.950 align:middle line:84% and then present to all of us that information as team 00:09:05.950 --> 00:09:09.630 align:middle line:90% or individual presentations. 00:09:09.630 --> 00:09:10.130 align:middle line:90% OK. 00:09:10.130 --> 00:09:14.540 align:middle line:84% That's the overview for the whole course. 00:09:14.540 --> 00:09:21.200 align:middle line:84% Today's story-- I try to make this somewhat a narrative 00:09:21.200 --> 00:09:23.420 align:middle line:90% and with themes. 00:09:23.420 --> 00:09:27.260 align:middle line:84% So we're going to toggle back and forth 00:09:27.260 --> 00:09:31.610 align:middle line:84% throughout the lecture on living systems 00:09:31.610 --> 00:09:33.050 align:middle line:90% and computational systems-- 00:09:33.050 --> 00:09:35.060 align:middle line:84% similarities, differences, how we 00:09:35.060 --> 00:09:38.750 align:middle line:84% can use one to tell us something about the other. 00:09:38.750 --> 00:09:42.860 align:middle line:84% In particular, we'll focus in on the aspect of living systems 00:09:42.860 --> 00:09:48.950 align:middle line:84% which is fairly unique, which is self-assembly and replication. 00:09:48.950 --> 00:09:55.070 align:middle line:84% These are put in the context of another theme for today, which 00:09:55.070 --> 00:09:58.070 align:middle line:84% is discrete versus continuous data, discrete 00:09:58.070 --> 00:10:02.210 align:middle line:84% versus continuous modeling of data. 00:10:02.210 --> 00:10:04.070 align:middle line:84% Since this is an introductory lecture 00:10:04.070 --> 00:10:05.527 align:middle line:84% and we want to get you warmed up, 00:10:05.527 --> 00:10:07.610 align:middle line:84% I'm going to try to illustrate as much as possible 00:10:07.610 --> 00:10:11.430 align:middle line:84% with minimal examples, small examples, 00:10:11.430 --> 00:10:13.580 align:middle line:84% so you can see it all in one page 00:10:13.580 --> 00:10:18.710 align:middle line:84% or all in one line, even, examples of minimal life-- 00:10:18.710 --> 00:10:21.470 align:middle line:84% things that illustrate the minimal aspects of life 00:10:21.470 --> 00:10:24.920 align:middle line:84% replication, mainly, and minimal programs, 00:10:24.920 --> 00:10:27.680 align:middle line:84% which allow us to analyze some key aspects 00:10:27.680 --> 00:10:29.660 align:middle line:90% of modeling living systems. 00:10:29.660 --> 00:10:33.800 align:middle line:84% Some of the key aspects are catalysis and replication. 00:10:33.800 --> 00:10:37.370 align:middle line:84% And here we'll use differential equations. 00:10:37.370 --> 00:10:42.020 align:middle line:84% This will be quite, I think, an interesting approach to it, 00:10:42.020 --> 00:10:46.280 align:middle line:84% painless, and exciting the way it connects to biology. 00:10:46.280 --> 00:10:52.100 align:middle line:84% And then after talking about that replication 00:10:52.100 --> 00:10:54.350 align:middle line:84% in the context of differential equations, 00:10:54.350 --> 00:10:56.540 align:middle line:84% we'll introduce [? directed ?] graphs, 00:10:56.540 --> 00:11:01.250 align:middle line:84% which indicate how growth occurs in a pedigree, not just growth 00:11:01.250 --> 00:11:02.810 align:middle line:90% in exponential mode. 00:11:02.810 --> 00:11:06.170 align:middle line:84% And then finally, we'll connect this 00:11:06.170 --> 00:11:11.270 align:middle line:84% to the issues that surround single molecules, which 00:11:11.270 --> 00:11:15.440 align:middle line:84% are actually very significant in biological systems 00:11:15.440 --> 00:11:19.070 align:middle line:84% and involve a different type of analysis 00:11:19.070 --> 00:11:21.500 align:middle line:84% than the continuous functions that we 00:11:21.500 --> 00:11:26.150 align:middle line:84% will use in ordinary differential equations. 00:11:26.150 --> 00:11:29.960 align:middle line:84% When we analyze errors, either in data 00:11:29.960 --> 00:11:35.000 align:middle line:84% or variance in biological populations, 00:11:35.000 --> 00:11:37.730 align:middle line:84% we use statistics, broadly speaking, 00:11:37.730 --> 00:11:40.280 align:middle line:84% that fall under bell curve statistics-- actually, 00:11:40.280 --> 00:11:41.810 align:middle line:90% more than that. 00:11:41.810 --> 00:11:45.350 align:middle line:90% And overall, the entire-- 00:11:45.350 --> 00:11:49.800 align:middle line:84% the theme that unites all of biology and most of the course, 00:11:49.800 --> 00:11:53.090 align:middle line:84% and this lecture in particular, is the idea 00:11:53.090 --> 00:11:58.460 align:middle line:84% that many of these functions in biological systems 00:11:58.460 --> 00:12:01.880 align:middle line:84% can be under selection, experimental selection 00:12:01.880 --> 00:12:07.310 align:middle line:84% in the laboratory in order to obtain an optimal growth 00:12:07.310 --> 00:12:09.490 align:middle line:84% rate or some other aspect of optionality. 00:12:09.490 --> 00:12:13.010 align:middle line:84% You generally will do well embedding the biological system 00:12:13.010 --> 00:12:18.470 align:middle line:84% either as optimal or can be made optimal for a particular task. 00:12:18.470 --> 00:12:20.540 align:middle line:90% OK. 00:12:20.540 --> 00:12:26.910 align:middle line:84% Now, I'm just poking fun at the number for this course. 00:12:26.910 --> 00:12:31.670 align:middle line:84% It is also an intrinsically important symbol 00:12:31.670 --> 00:12:34.670 align:middle line:84% for the 0's and 1's that occur in all of our computers 00:12:34.670 --> 00:12:43.100 align:middle line:84% and under underlie the way we either deal with discrete data 00:12:43.100 --> 00:12:46.820 align:middle line:84% or make continuous data appear to be discrete. 00:12:46.820 --> 00:12:53.330 align:middle line:84% We'll start with the biological entities most similar 00:12:53.330 --> 00:12:57.410 align:middle line:84% to the 0's and 1's that make your computers hum. 00:12:57.410 --> 00:13:01.400 align:middle line:84% And these are the nucleotides of DNA 00:13:01.400 --> 00:13:06.090 align:middle line:84% and RNA, A, C, G, and T, or A, C, G, and U in the case of RNA. 00:13:06.090 --> 00:13:10.670 align:middle line:84% And of course, all of this is symbols. 00:13:10.670 --> 00:13:13.010 align:middle line:90% The A stands for adenine. 00:13:13.010 --> 00:13:14.630 align:middle line:84% And adenine, of course, doesn't look 00:13:14.630 --> 00:13:17.210 align:middle line:84% like an A. It looks like some electron density. 00:13:17.210 --> 00:13:19.580 align:middle line:84% And we often represent it with chemical formula you'll 00:13:19.580 --> 00:13:21.750 align:middle line:90% see in subsequent slides. 00:13:21.750 --> 00:13:28.340 align:middle line:84% But here in slide 5, we're talking about, schematically, 00:13:28.340 --> 00:13:30.020 align:middle line:90% A, C, G, and T being-- 00:13:30.020 --> 00:13:35.810 align:middle line:84% A being represented by the two-digit binary number 00, 00:13:35.810 --> 00:13:38.930 align:middle line:90% and C 01 and 10 and 11. 00:13:38.930 --> 00:13:43.190 align:middle line:84% So you can see that two digits isn't 00:13:43.190 --> 00:13:47.540 align:middle line:84% quite enough to encode the four nucleotides which 00:13:47.540 --> 00:13:49.970 align:middle line:84% are strung together where you might have 00:13:49.970 --> 00:13:52.610 align:middle line:84% 3 billion such nucleotides making up 00:13:52.610 --> 00:13:56.135 align:middle line:90% one of your two human genomes. 00:13:56.135 --> 00:13:58.010 align:middle line:84% Now, things get a little bit more complicated 00:13:58.010 --> 00:14:04.610 align:middle line:84% once this digital 3 billion nucleotides in your genome 00:14:04.610 --> 00:14:09.980 align:middle line:84% starts being turned into molecules that actually 00:14:09.980 --> 00:14:12.980 align:middle line:84% do the work of the cell, the work of your body, which 00:14:12.980 --> 00:14:14.340 align:middle line:90% is RNA and proteins. 00:14:14.340 --> 00:14:19.190 align:middle line:84% This is an example of one of the primary transcripts 00:14:19.190 --> 00:14:22.590 align:middle line:84% common to all living organisms that we know of. 00:14:22.590 --> 00:14:24.440 align:middle line:90% This is a transfer RNA. 00:14:24.440 --> 00:14:27.140 align:middle line:84% I actually participated in solving 00:14:27.140 --> 00:14:31.400 align:middle line:84% this complicated structure in the '70s. 00:14:31.400 --> 00:14:35.750 align:middle line:84% And this is color coded, the DNA sequence that 00:14:35.750 --> 00:14:39.200 align:middle line:90% encodes this particular RNA. 00:14:39.200 --> 00:14:41.510 align:middle line:84% And when the RNA is made, it folds up 00:14:41.510 --> 00:14:43.310 align:middle line:84% into this three-dimensional structure 00:14:43.310 --> 00:14:44.480 align:middle line:90% which you see rotating here. 00:14:44.480 --> 00:14:46.640 align:middle line:84% This is actually a stereo image, which, 00:14:46.640 --> 00:14:48.560 align:middle line:84% if you cross your eyes in the right way, 00:14:48.560 --> 00:14:50.870 align:middle line:84% will appear to be even more three-dimensional than it 00:14:50.870 --> 00:14:52.500 align:middle line:90% is there. 00:14:52.500 --> 00:14:55.440 align:middle line:84% You will learn more about this later in the course 00:14:55.440 --> 00:14:56.570 align:middle line:90% if you don't know already. 00:14:56.570 --> 00:15:00.500 align:middle line:84% But the point is that this goes from a 5-prime end 00:15:00.500 --> 00:15:03.770 align:middle line:84% at blue all the way out to the 3-prime end in red. 00:15:03.770 --> 00:15:06.050 align:middle line:84% And it has to fold in this manner. 00:15:06.050 --> 00:15:07.710 align:middle line:84% It reproducibly folds in this manner, 00:15:07.710 --> 00:15:13.250 align:middle line:84% and it has to stay in order to perform its function, which is 00:15:13.250 --> 00:15:15.650 align:middle line:90% actually translating from DNA-- 00:15:15.650 --> 00:15:18.180 align:middle line:84% sorry-- from RNA sequences into protein. 00:15:18.180 --> 00:15:21.080 align:middle line:84% But what we're illustrating here is the relationship between 00:15:21.080 --> 00:15:28.940 align:middle line:84% this discrete binary digital code and DNA and the more 00:15:28.940 --> 00:15:35.340 align:middle line:84% continuous code that you have, which is the x, y, 00:15:35.340 --> 00:15:37.050 align:middle line:90% and z-coordinates of the atoms-- 00:15:37.050 --> 00:15:39.810 align:middle line:84% the continuous nature of the probability distribution 00:15:39.810 --> 00:15:43.590 align:middle line:84% of electrons around each of those atoms 00:15:43.590 --> 00:15:46.470 align:middle line:84% and the continuous nature of its position 00:15:46.470 --> 00:15:51.240 align:middle line:84% in space and its various binding constants, affinities 00:15:51.240 --> 00:15:54.230 align:middle line:90% for other molecules in the cell. 00:15:54.230 --> 00:15:58.290 align:middle line:84% Let me just take those two examples, the DNA sequence 00:15:58.290 --> 00:16:00.750 align:middle line:84% and the three-dimensional structure for the transfer RNA 00:16:00.750 --> 00:16:05.880 align:middle line:84% that it encodes, and expand upon them a little bit 00:16:05.880 --> 00:16:08.340 align:middle line:90% or give other examples. 00:16:08.340 --> 00:16:11.670 align:middle line:90% We have a sequence. 00:16:11.670 --> 00:16:17.610 align:middle line:84% In the previous slide, I showed a sequence of 76 nucleotides. 00:16:17.610 --> 00:16:22.110 align:middle line:84% On the left-hand side are examples of discrete concepts, 00:16:22.110 --> 00:16:24.210 align:middle line:84% concepts which are very naturally 00:16:24.210 --> 00:16:29.490 align:middle line:84% encoded in a discrete digital way, and then as close 00:16:29.490 --> 00:16:32.680 align:middle line:84% as we can get to the continuous version of that. 00:16:32.680 --> 00:16:34.290 align:middle line:84% So a continuous version of a sequence 00:16:34.290 --> 00:16:37.560 align:middle line:84% might be a probability of a sequence. 00:16:37.560 --> 00:16:39.210 align:middle line:84% In other words, at the first position, 00:16:39.210 --> 00:16:40.980 align:middle line:84% if you look at a population of sequences-- 00:16:40.980 --> 00:16:43.522 align:middle line:84% the number of different tRNAs, the number of different people 00:16:43.522 --> 00:16:44.490 align:middle line:90% in this room-- 00:16:44.490 --> 00:16:47.460 align:middle line:84% it may not be that there's always an A at position 1. 00:16:47.460 --> 00:16:50.147 align:middle line:84% It could be that it's got a different probability 00:16:50.147 --> 00:16:50.980 align:middle line:90% in different people. 00:16:50.980 --> 00:16:55.560 align:middle line:84% So you represent that as a probability A, C, G, and T. 00:16:55.560 --> 00:16:59.790 align:middle line:84% It's a vector with four numbers in it. 00:16:59.790 --> 00:17:02.760 align:middle line:90% That's more continuous. 00:17:02.760 --> 00:17:05.730 align:middle line:84% Similarly, we have digital analog devices 00:17:05.730 --> 00:17:08.880 align:middle line:84% in the instruments that collect the data that you'll 00:17:08.880 --> 00:17:10.980 align:middle line:90% be working with in this course. 00:17:10.980 --> 00:17:14.280 align:middle line:84% Integration can be represented, practically speaking, 00:17:14.280 --> 00:17:18.720 align:middle line:90% as a sum of small steps. 00:17:18.720 --> 00:17:24.300 align:middle line:84% When we're talking about, say, a neural network that's 00:17:24.300 --> 00:17:27.918 align:middle line:84% responsible for some of your thoughts 00:17:27.918 --> 00:17:31.530 align:middle line:84% or a regulatory network that causes homeostasis 00:17:31.530 --> 00:17:37.020 align:middle line:84% in your body, biologists casually often reflect on these 00:17:37.020 --> 00:17:38.470 align:middle line:90% as being on and off. 00:17:38.470 --> 00:17:42.120 align:middle line:90% It's a good approximation. 00:17:42.120 --> 00:17:44.550 align:middle line:84% It allows you to make very simple diagrams. 00:17:44.550 --> 00:17:47.100 align:middle line:84% But you must remember that many of these 00:17:47.100 --> 00:17:51.150 align:middle line:84% are actually composed of gradients or graded responses. 00:17:51.150 --> 00:17:53.850 align:middle line:84% Not necessarily all are [? on/off. ?] 00:17:53.850 --> 00:17:55.980 align:middle line:84% Some of these gradients or graded responses 00:17:55.980 --> 00:17:59.280 align:middle line:84% have this sigmoid shape that I've drawn as an icon 00:17:59.280 --> 00:18:01.590 align:middle line:84% down the middle of this, separating the discrete 00:18:01.590 --> 00:18:03.160 align:middle line:90% from the continuous. 00:18:03.160 --> 00:18:04.920 align:middle line:84% And so for all intents and purposes, 00:18:04.920 --> 00:18:07.980 align:middle line:84% things tend to hang out either off here at the bottom 00:18:07.980 --> 00:18:14.220 align:middle line:84% or on at the top of the concentration limits 00:18:14.220 --> 00:18:15.900 align:middle line:84% or the signal limits if the signal 00:18:15.900 --> 00:18:20.440 align:middle line:90% is electrical or so forth. 00:18:20.440 --> 00:18:24.450 align:middle line:84% Similarly, we'll have examples where some cells will be-- 00:18:24.450 --> 00:18:28.170 align:middle line:84% we'll have a field of cells which are either on or off, 00:18:28.170 --> 00:18:30.660 align:middle line:84% effectively very extreme ends of this. 00:18:30.660 --> 00:18:32.610 align:middle line:84% But because there's a mixture of them, 00:18:32.610 --> 00:18:36.420 align:middle line:84% if you were to mush them up and measure some property of them, 00:18:36.420 --> 00:18:38.470 align:middle line:84% it would appear to be somewhere intermediate. 00:18:38.470 --> 00:18:40.950 align:middle line:84% So this is an alarm that should be going off 00:18:40.950 --> 00:18:46.020 align:middle line:84% every time you see mixtures that they may be composed of-- 00:18:46.020 --> 00:18:47.730 align:middle line:84% you may need-- to model it, you may 00:18:47.730 --> 00:18:53.580 align:middle line:84% have to model it as a population, each individual 00:18:53.580 --> 00:18:59.610 align:middle line:84% behaving as more extreme than the average with very 00:18:59.610 --> 00:19:02.160 align:middle line:90% few in the middle. 00:19:02.160 --> 00:19:04.590 align:middle line:84% Similarly with mutations, you might 00:19:04.590 --> 00:19:07.740 align:middle line:84% say that certain mutations are essential for life. 00:19:07.740 --> 00:19:09.010 align:middle line:90% Others are neutral. 00:19:09.010 --> 00:19:10.590 align:middle line:90% They have no effect. 00:19:10.590 --> 00:19:15.811 align:middle line:84% Others, more gray zone, are conditional. 00:19:15.811 --> 00:19:18.520 align:middle line:90% 00:19:18.520 --> 00:19:20.110 align:middle line:90% OK. 00:19:20.110 --> 00:19:25.360 align:middle line:84% Just a point of orientation for how we can describe 00:19:25.360 --> 00:19:28.720 align:middle line:84% not only the bits, the discrete or digital components 00:19:28.720 --> 00:19:30.010 align:middle line:90% of this course. 00:19:30.010 --> 00:19:33.340 align:middle line:84% But also, many of these prefixes are 00:19:33.340 --> 00:19:37.160 align:middle line:84% useful for describing the continuous as well. 00:19:37.160 --> 00:19:38.207 align:middle line:90% But you can see this-- 00:19:38.207 --> 00:19:40.780 align:middle line:90% 00:19:40.780 --> 00:19:42.550 align:middle line:84% many of you will be familiar with these. 00:19:42.550 --> 00:19:45.490 align:middle line:84% All of you, I'm sure, have access to computers, 00:19:45.490 --> 00:19:48.160 align:middle line:84% and so you've used terms like kilobyte and megabyte 00:19:48.160 --> 00:19:51.820 align:middle line:90% and gigabyte. 00:19:51.820 --> 00:19:59.830 align:middle line:84% Technically, 2 to the 10th power is 1,024, not 1,000, 00:19:59.830 --> 00:20:01.810 align:middle line:84% and so it shouldn't be referred to as kilo. 00:20:01.810 --> 00:20:03.470 align:middle line:84% But for most intents and purposes, 00:20:03.470 --> 00:20:07.240 align:middle line:84% these are so close to 1,000 or a million or a billion 00:20:07.240 --> 00:20:09.040 align:middle line:84% that they're used interchangeably. 00:20:09.040 --> 00:20:11.650 align:middle line:84% But this is the official standard here. 00:20:11.650 --> 00:20:17.050 align:middle line:84% And certainly for the continuous numbers 00:20:17.050 --> 00:20:20.920 align:middle line:84% that we'll be talking about, the order 00:20:20.920 --> 00:20:23.620 align:middle line:84% we'll have is 10 to the plus 3 will [? be kilo. ?] 10 00:20:23.620 --> 00:20:27.550 align:middle line:84% to the minus 3 will be milliliters. 00:20:27.550 --> 00:20:31.690 align:middle line:84% Mega, micro, giga, nano, and so on, all the way up. 00:20:31.690 --> 00:20:34.850 align:middle line:84% [? For the ?] numbers, atom mole is getting-- 00:20:34.850 --> 00:20:37.090 align:middle line:84% zeptomole is getting close to the limit 00:20:37.090 --> 00:20:40.480 align:middle line:84% where you're getting close to single molecules. 00:20:40.480 --> 00:20:44.530 align:middle line:84% Petabyte is getting close to the limit of where your computers 00:20:44.530 --> 00:20:45.430 align:middle line:90% can go right now. 00:20:45.430 --> 00:20:49.990 align:middle line:90% 00:20:49.990 --> 00:20:53.680 align:middle line:84% Why is it important to have defined quantitative measures? 00:20:53.680 --> 00:20:57.280 align:middle line:84% Why can't we just casually say, oh, yeah, it's 00:20:57.280 --> 00:21:02.170 align:middle line:84% a long time rather than seconds, or it's really long distance 00:21:02.170 --> 00:21:05.620 align:middle line:90% to Harvard Square in meters? 00:21:05.620 --> 00:21:09.490 align:middle line:84% Why do we have to use meters, kilograms, moles, degrees, 00:21:09.490 --> 00:21:11.410 align:middle line:90% Kelvin, candelas, amperes? 00:21:11.410 --> 00:21:16.030 align:middle line:84% These are the seven basic international system units 00:21:16.030 --> 00:21:20.440 align:middle line:84% from which most of the other units that you use in science 00:21:20.440 --> 00:21:21.550 align:middle line:90% are derived. 00:21:21.550 --> 00:21:25.030 align:middle line:84% In a certain sense, there is some inter-convertibility even 00:21:25.030 --> 00:21:26.920 align:middle line:90% within these. 00:21:26.920 --> 00:21:29.410 align:middle line:84% We'll be talking about precision a little bit 00:21:29.410 --> 00:21:32.840 align:middle line:84% at the end of today's class and throughout the course. 00:21:32.840 --> 00:21:35.980 align:middle line:84% In fact, a theme today is that even things that are 00:21:35.980 --> 00:21:40.510 align:middle line:90% represented digitally are not-- 00:21:40.510 --> 00:21:42.400 align:middle line:84% in order to represent things digitally, 00:21:42.400 --> 00:21:44.650 align:middle line:84% there are approximations that are made. 00:21:44.650 --> 00:21:49.810 align:middle line:84% And you should always question what those assumptions are 00:21:49.810 --> 00:21:51.460 align:middle line:90% when you make an approximation. 00:21:51.460 --> 00:21:54.080 align:middle line:84% The precision for some of these units of measure-- 00:21:54.080 --> 00:21:57.070 align:middle line:84% for example, the time scale measured in seconds-- 00:21:57.070 --> 00:22:04.780 align:middle line:84% can be as precise as 14 significant decimal digits. 00:22:04.780 --> 00:22:08.320 align:middle line:84% Now, you may wonder from time to time in the course 00:22:08.320 --> 00:22:12.040 align:middle line:84% why biology doesn't have 14 significant figures. 00:22:12.040 --> 00:22:14.650 align:middle line:84% But I will leave that up as an exercise 00:22:14.650 --> 00:22:18.470 align:middle line:84% for one of your projects, maybe, to figure out how to get that. 00:22:18.470 --> 00:22:21.310 align:middle line:84% But for all practical purposes, most of biology 00:22:21.310 --> 00:22:26.440 align:middle line:84% is three or four significant figures. 00:22:26.440 --> 00:22:31.430 align:middle line:84% There are quantum limits to time and length, 00:22:31.430 --> 00:22:35.830 align:middle line:84% which are so short that we don't need to concern ourselves. 00:22:35.830 --> 00:22:40.660 align:middle line:84% But the quantum unit for the mole 00:22:40.660 --> 00:22:43.900 align:middle line:84% is actually of great importance to this course. 00:22:43.900 --> 00:22:48.800 align:middle line:84% A mole is 6 times 10 to the 23rd entities. 00:22:48.800 --> 00:22:51.580 align:middle line:84% These entities can be photons or molecules, 00:22:51.580 --> 00:22:59.020 align:middle line:84% or we even have over 6 times 10 to the 23rd 00:22:59.020 --> 00:23:03.490 align:middle line:90% of the bacteria in an ocean. 00:23:03.490 --> 00:23:09.790 align:middle line:84% So what's important here is that the quantum of this 00:23:09.790 --> 00:23:12.370 align:middle line:84% is the molecule, and many of the things we'll be dealing with 00:23:12.370 --> 00:23:13.420 align:middle line:90% are single molecules. 00:23:13.420 --> 00:23:17.110 align:middle line:84% And that'll be the last topic of today's lecture, 00:23:17.110 --> 00:23:20.680 align:middle line:84% will be how one deals with single molecules 00:23:20.680 --> 00:23:24.470 align:middle line:84% as opposed to large buckets of molecules. 00:23:24.470 --> 00:23:28.690 align:middle line:84% Now, we have all those great quantitative definitions 00:23:28.690 --> 00:23:33.590 align:middle line:84% for all the standard units used in physics and chemistry, 00:23:33.590 --> 00:23:37.180 align:middle line:84% and even in biology, but what about the definition of biology 00:23:37.180 --> 00:23:39.580 align:middle line:90% itself? 00:23:39.580 --> 00:23:42.070 align:middle line:84% Most biology books start with this. 00:23:42.070 --> 00:23:44.170 align:middle line:84% There are entire books dedicated to this question 00:23:44.170 --> 00:23:45.490 align:middle line:90% of what is life. 00:23:45.490 --> 00:23:52.150 align:middle line:84% I think that rather than, in this session, ask what is alive 00:23:52.150 --> 00:23:53.440 align:middle line:90% or not-- 00:23:53.440 --> 00:23:57.490 align:middle line:84% rather than make it a dichotomy, I'd prefer 00:23:57.490 --> 00:24:00.700 align:middle line:90% to say how alive is something. 00:24:00.700 --> 00:24:04.780 align:middle line:84% What is the probability that a given entity will replicate? 00:24:04.780 --> 00:24:09.940 align:middle line:84% We have, here on slide 10, the probability of replication, 00:24:09.940 --> 00:24:14.560 align:middle line:84% and not just how likely is it for the whole thing 00:24:14.560 --> 00:24:16.540 align:middle line:84% to replicate or some part of it to replicate, 00:24:16.540 --> 00:24:21.340 align:middle line:84% but is it doing so using simple parts, 00:24:21.340 --> 00:24:25.780 align:middle line:84% simple environmental components, and creating great complexity 00:24:25.780 --> 00:24:26.380 align:middle line:90% from that. 00:24:26.380 --> 00:24:29.510 align:middle line:90% How faithful is the replication? 00:24:29.510 --> 00:24:31.780 align:middle line:84% Now, this probability of replication 00:24:31.780 --> 00:24:35.830 align:middle line:84% from simplicity to complexity can be defined 00:24:35.830 --> 00:24:37.510 align:middle line:90% for a specific environment. 00:24:37.510 --> 00:24:40.630 align:middle line:84% Sometimes, the environment required for replication 00:24:40.630 --> 00:24:42.670 align:middle line:90% is, of necessity, very specific. 00:24:42.670 --> 00:24:46.210 align:middle line:84% Certain organisms do not do well except 00:24:46.210 --> 00:24:47.950 align:middle line:90% in their native environments. 00:24:47.950 --> 00:24:50.950 align:middle line:84% Zoos discover this, and we keep killing off 00:24:50.950 --> 00:24:53.230 align:middle line:84% species for a variety of reasons, one 00:24:53.230 --> 00:24:56.530 align:middle line:90% of them being this specificity. 00:24:56.530 --> 00:24:58.190 align:middle line:90% So when we talk about-- 00:24:58.190 --> 00:25:01.480 align:middle line:90% so another aspect of life-- 00:25:01.480 --> 00:25:04.780 align:middle line:84% which each of these, in principle, can be quantitated-- 00:25:04.780 --> 00:25:06.040 align:middle line:90% is how robust is it. 00:25:06.040 --> 00:25:09.428 align:middle line:84% How many environments can it handle? 00:25:09.428 --> 00:25:11.470 align:middle line:84% Inevitably, there are environments that cannot be 00:25:11.470 --> 00:25:11.990 align:middle line:90% handled. 00:25:11.990 --> 00:25:14.540 align:middle line:84% But the more robust and adaptable 00:25:14.540 --> 00:25:18.070 align:middle line:84% it is, in a certain sense, the more alive it is 00:25:18.070 --> 00:25:21.100 align:middle line:84% or the more alive it will be, its descendants 00:25:21.100 --> 00:25:25.000 align:middle line:84% will be, millions of years from now, probably. 00:25:25.000 --> 00:25:27.250 align:middle line:84% Now, some very challenging examples I list here. 00:25:27.250 --> 00:25:29.000 align:middle line:84% I'm not going to walk through all of them. 00:25:29.000 --> 00:25:33.250 align:middle line:84% But things that have challenged people's definitions of life 00:25:33.250 --> 00:25:34.900 align:middle line:90% before are mules-- 00:25:34.900 --> 00:25:38.560 align:middle line:84% that is to say, sterile hybrids which 00:25:38.560 --> 00:25:41.920 align:middle line:84% will not leave behind progeny but seem quite as 00:25:41.920 --> 00:25:48.430 align:middle line:84% alive as their parents were but will not replicate, generally. 00:25:48.430 --> 00:25:49.870 align:middle line:84% So the probability of replication 00:25:49.870 --> 00:25:53.230 align:middle line:84% is low for the entire organism, but maybe for individual cells. 00:25:53.230 --> 00:25:56.260 align:middle line:84% Fires replicate quite well, but most people 00:25:56.260 --> 00:25:59.020 align:middle line:84% would like to exclude them from this definition. 00:25:59.020 --> 00:26:03.070 align:middle line:84% Or perhaps if we're in the mode of not excluding things 00:26:03.070 --> 00:26:06.640 align:middle line:84% from life, the probability of replication complexity 00:26:06.640 --> 00:26:10.300 align:middle line:84% versus simplicity might impinge upon fires. 00:26:10.300 --> 00:26:14.410 align:middle line:84% Crystals-- if you nucleate a supersaturated solution 00:26:14.410 --> 00:26:17.140 align:middle line:84% with a crystal, it will make, in a certain sense, 00:26:17.140 --> 00:26:18.130 align:middle line:90% copies of itself. 00:26:18.130 --> 00:26:19.870 align:middle line:90% How [? painful ?] is that? 00:26:19.870 --> 00:26:23.290 align:middle line:90% How simple is it? 00:26:23.290 --> 00:26:25.570 align:middle line:84% Flowers, viruses, predators-- these require 00:26:25.570 --> 00:26:27.070 align:middle line:90% very complex environments. 00:26:27.070 --> 00:26:30.340 align:middle line:84% They require environments often more complex than themselves 00:26:30.340 --> 00:26:32.080 align:middle line:90% in order to replicate. 00:26:32.080 --> 00:26:33.460 align:middle line:90% Does that make them less alive? 00:26:33.460 --> 00:26:36.160 align:middle line:90% 00:26:36.160 --> 00:26:38.470 align:middle line:84% I'll show an example of molecular ligation 00:26:38.470 --> 00:26:41.080 align:middle line:84% as the simplest case just to get us warmed up 00:26:41.080 --> 00:26:43.390 align:middle line:90% for this introductory class. 00:26:43.390 --> 00:26:49.040 align:middle line:90% 00:26:49.040 --> 00:26:52.400 align:middle line:84% Since we're in the topic here, briefly, of general biology, 00:26:52.400 --> 00:26:54.650 align:middle line:84% not just the historical terrestrial biology 00:26:54.650 --> 00:26:58.160 align:middle line:84% from which all of us feel particular affinity-- 00:26:58.160 --> 00:27:04.640 align:middle line:84% if we had visitors from some other planet 00:27:04.640 --> 00:27:07.820 align:middle line:84% or if we started making our own self-assembling machines-- 00:27:07.820 --> 00:27:10.880 align:middle line:84% and to some extent, we already are 00:27:10.880 --> 00:27:14.630 align:middle line:84% making self-assembling machines in factories. 00:27:14.630 --> 00:27:16.593 align:middle line:84% They require a very complex environment 00:27:16.593 --> 00:27:17.510 align:middle line:90% which includes humans. 00:27:17.510 --> 00:27:21.230 align:middle line:84% But how do we define these things? 00:27:21.230 --> 00:27:23.570 align:middle line:90% Now, in order to define-- 00:27:23.570 --> 00:27:28.880 align:middle line:84% to get at that complexity versus simplicity issue 00:27:28.880 --> 00:27:30.650 align:middle line:84% and faithfulness of replication, we 00:27:30.650 --> 00:27:34.190 align:middle line:84% need to define both replication and complexity. 00:27:34.190 --> 00:27:42.020 align:middle line:84% Replication is not perfectly faithful in many of the things 00:27:42.020 --> 00:27:45.930 align:middle line:90% that you would consider alive. 00:27:45.930 --> 00:27:48.860 align:middle line:84% So simple bacterium, even though it 00:27:48.860 --> 00:27:52.337 align:middle line:84% will make a copy that does many of the same things-- 00:27:52.337 --> 00:27:54.920 align:middle line:84% we know it when we see it-- it looks like it's the same thing. 00:27:54.920 --> 00:27:56.780 align:middle line:84% If you actually counted the molecules, 00:27:56.780 --> 00:27:58.550 align:middle line:90% no two bacteria would be alike. 00:27:58.550 --> 00:28:00.440 align:middle line:84% No two humans, certainly, are alike. 00:28:00.440 --> 00:28:04.730 align:middle line:84% But even supposedly genetically identical bacteria 00:28:04.730 --> 00:28:08.240 align:middle line:84% will have different numbers of proteins and small molecules. 00:28:08.240 --> 00:28:11.490 align:middle line:84% Complexity has at least four definitions 00:28:11.490 --> 00:28:14.250 align:middle line:90% we will use in this course. 00:28:14.250 --> 00:28:16.400 align:middle line:84% There is computational complexity, 00:28:16.400 --> 00:28:18.590 align:middle line:84% which is of practical significance 00:28:18.590 --> 00:28:22.100 align:middle line:84% in computational biology, as the computer scientists in the room 00:28:22.100 --> 00:28:26.000 align:middle line:84% will know, in that this tells us the speed and memory 00:28:26.000 --> 00:28:29.390 align:middle line:84% tradeoffs we have in scaling up any problem. 00:28:29.390 --> 00:28:34.670 align:middle line:84% Does a problem scale up as a simple linear function 00:28:34.670 --> 00:28:37.790 align:middle line:84% of the number of inputs that you give to the computer? 00:28:37.790 --> 00:28:40.400 align:middle line:84% Does it scale up as a simple polynomial, 00:28:40.400 --> 00:28:41.810 align:middle line:90% or is it worse than that? 00:28:41.810 --> 00:28:43.520 align:middle line:90% Is it exponential in behavior? 00:28:43.520 --> 00:28:46.850 align:middle line:84% Is it something that you can prove 00:28:46.850 --> 00:28:49.410 align:middle line:84% you got the right answer in polynomial time, and so on? 00:28:49.410 --> 00:28:51.560 align:middle line:84% We'll get to this later on, but I just 00:28:51.560 --> 00:28:53.360 align:middle line:84% want to introduce it as something where 00:28:53.360 --> 00:28:55.550 align:middle line:90% the word "complexity" is used. 00:28:55.550 --> 00:29:00.000 align:middle line:84% Number two sounds similar but is actually quite a bit different. 00:29:00.000 --> 00:29:04.400 align:middle line:84% It's algorithmic complexity or algorithmic randomness. 00:29:04.400 --> 00:29:06.243 align:middle line:84% And this is basically any string-- 00:29:06.243 --> 00:29:08.660 align:middle line:84% not just a computer program, but a computer program can be 00:29:08.660 --> 00:29:10.220 align:middle line:90% represented as a string-- 00:29:10.220 --> 00:29:16.220 align:middle line:84% can be reduced down where you get rid of obvious redundancy, 00:29:16.220 --> 00:29:18.530 align:middle line:84% and what's left is the randomness. 00:29:18.530 --> 00:29:21.560 align:middle line:84% And the number of bits it takes to encode that algorithm 00:29:21.560 --> 00:29:23.840 align:middle line:84% is a reflection of the complexity. 00:29:23.840 --> 00:29:26.580 align:middle line:90% 00:29:26.580 --> 00:29:28.080 align:middle line:84% That doesn't necessarily give you 00:29:28.080 --> 00:29:30.690 align:middle line:84% any predictions about how long it will run 00:29:30.690 --> 00:29:35.220 align:middle line:84% or how much memory it will require during computation. 00:29:35.220 --> 00:29:40.680 align:middle line:84% Entropy and information, number 3, is related to item number 2 00:29:40.680 --> 00:29:44.400 align:middle line:84% in that the more complex the string-- 00:29:44.400 --> 00:29:46.870 align:middle line:84% a string is just a series of symbols like this, 00:29:46.870 --> 00:29:48.750 align:middle line:84% and that was what we were talking about with 00:29:48.750 --> 00:29:50.850 align:middle line:84% the randomness-- the more complex the string, 00:29:50.850 --> 00:29:52.920 align:middle line:84% or an image can be turned into a string-- 00:29:52.920 --> 00:29:55.740 align:middle line:90% these are three images here-- 00:29:55.740 --> 00:29:58.080 align:middle line:84% then the more bits you need to encode it 00:29:58.080 --> 00:29:59.970 align:middle line:90% in order to, say, make a file. 00:29:59.970 --> 00:30:03.240 align:middle line:84% You might compress the data, but ultimately, 00:30:03.240 --> 00:30:05.040 align:middle line:84% after it's fully compressed, that's 00:30:05.040 --> 00:30:07.680 align:middle line:84% the amount of information you need. 00:30:07.680 --> 00:30:11.520 align:middle line:90% Entropy is a chemical term. 00:30:11.520 --> 00:30:15.510 align:middle line:84% This information definition was championed by Shannon-- 00:30:15.510 --> 00:30:17.190 align:middle line:84% we'll come back to it-- and entropy 00:30:17.190 --> 00:30:18.540 align:middle line:90% by Boltzmann and others. 00:30:18.540 --> 00:30:22.530 align:middle line:84% And the fusion of these two into a unified theme 00:30:22.530 --> 00:30:25.765 align:middle line:84% is extremely important in both chemistry and information 00:30:25.765 --> 00:30:26.265 align:middle line:90% theory. 00:30:26.265 --> 00:30:29.180 align:middle line:90% 00:30:29.180 --> 00:30:32.390 align:middle line:84% Nevertheless, none of these above 00:30:32.390 --> 00:30:34.760 align:middle line:84% reflect our intuitive feeling for when 00:30:34.760 --> 00:30:37.430 align:middle line:90% we look at these three panels. 00:30:37.430 --> 00:30:43.190 align:middle line:84% We have a highly ordered, which would be a very low entropy, 00:30:43.190 --> 00:30:46.620 align:middle line:90% array on the left-hand side. 00:30:46.620 --> 00:30:49.790 align:middle line:84% We have, on the far right-hand side, something highly 00:30:49.790 --> 00:30:52.100 align:middle line:90% disordered, essentially random. 00:30:52.100 --> 00:30:55.640 align:middle line:84% This may have been generated by a coin toss 00:30:55.640 --> 00:30:57.650 align:middle line:84% where, as you fill up the array, you just say 00:30:57.650 --> 00:31:01.460 align:middle line:90% black or white, toss a coin. 00:31:01.460 --> 00:31:02.970 align:middle line:90% The one in the middle-- 00:31:02.970 --> 00:31:06.890 align:middle line:84% and so the low entropy on the left could be easily 00:31:06.890 --> 00:31:10.940 align:middle line:84% represented as 010101 in a simple array-- 00:31:10.940 --> 00:31:14.630 align:middle line:84% very little information, very little entropy, 00:31:14.630 --> 00:31:17.960 align:middle line:84% algorithmically, not very random. 00:31:17.960 --> 00:31:21.500 align:middle line:84% At the other end, though, it has high entropy, high information 00:31:21.500 --> 00:31:22.050 align:middle line:90% content. 00:31:22.050 --> 00:31:23.690 align:middle line:84% It takes a lot of bits to represent it 00:31:23.690 --> 00:31:26.900 align:middle line:84% even though you know you got it just by a coin toss, 00:31:26.900 --> 00:31:29.690 align:middle line:84% and it has high algorithmic randomness. 00:31:29.690 --> 00:31:34.950 align:middle line:90% 00:31:34.950 --> 00:31:37.310 align:middle line:84% But if you allow a coin toss to be 00:31:37.310 --> 00:31:40.250 align:middle line:84% part of your description of the physical complexity 00:31:40.250 --> 00:31:43.340 align:middle line:84% of this pattern on the far right, 00:31:43.340 --> 00:31:48.260 align:middle line:84% say a gas or something like this generated by coin toss, 00:31:48.260 --> 00:31:53.630 align:middle line:84% then you can represent it as a particular kind 00:31:53.630 --> 00:31:57.860 align:middle line:84% of random description, which this is almost 00:31:57.860 --> 00:31:59.780 align:middle line:84% as-- this is about as easy to describe 00:31:59.780 --> 00:32:01.290 align:middle line:90% as a highly ordered system. 00:32:01.290 --> 00:32:03.860 align:middle line:84% So even though it's high entropy, it has low complexity. 00:32:03.860 --> 00:32:05.840 align:middle line:84% And complexity lies somewhere in between where 00:32:05.840 --> 00:32:08.990 align:middle line:84% you have lots of different kinds of symmetries, lots 00:32:08.990 --> 00:32:12.020 align:middle line:84% of different scales of structures, 00:32:12.020 --> 00:32:15.620 align:middle line:84% and it's very hard to represent it either as a random coin toss 00:32:15.620 --> 00:32:19.130 align:middle line:90% or as a highly ordered system. 00:32:19.130 --> 00:32:22.070 align:middle line:90% How do we quantitate this? 00:32:22.070 --> 00:32:25.460 align:middle line:84% We're really trying to move from the vaguer definitions 00:32:25.460 --> 00:32:29.660 align:middle line:84% to something that really encompasses what we intuitively 00:32:29.660 --> 00:32:31.580 align:middle line:90% feel about complexity. 00:32:31.580 --> 00:32:34.220 align:middle line:84% Here's an example I happen to like. 00:32:34.220 --> 00:32:36.170 align:middle line:84% I would say that this is not something 00:32:36.170 --> 00:32:39.080 align:middle line:84% that's broadly adapted, but I want to expose you to it. 00:32:39.080 --> 00:32:44.990 align:middle line:84% Here we have entropy or the randomness, the information-- 00:32:44.990 --> 00:32:49.640 align:middle line:84% Shannon information on the horizontal axis of slide 12 00:32:49.640 --> 00:32:51.718 align:middle line:84% where the low entropy that's highly ordered 00:32:51.718 --> 00:32:52.760 align:middle line:90% is on the left-hand side. 00:32:52.760 --> 00:32:53.750 align:middle line:90% It's 0. 00:32:53.750 --> 00:32:58.940 align:middle line:84% And the highly disordered, random high entropy 00:32:58.940 --> 00:33:04.640 align:middle line:84% of 1 on this type of scale is on the right-hand side. 00:33:04.640 --> 00:33:08.780 align:middle line:84% And complexity, as we said, we expect 00:33:08.780 --> 00:33:15.710 align:middle line:84% does not correlate perfectly with information or entropy 00:33:15.710 --> 00:33:19.400 align:middle line:84% in that this highly disordered structure 00:33:19.400 --> 00:33:22.190 align:middle line:84% on the right-hand side actually has very low complexity. 00:33:22.190 --> 00:33:24.770 align:middle line:84% And complexity is not even a single-valued function 00:33:24.770 --> 00:33:29.990 align:middle line:84% of entropy or information because you 00:33:29.990 --> 00:33:31.850 align:middle line:84% can have multiple different structures that 00:33:31.850 --> 00:33:34.250 align:middle line:84% have the same entropy but have different complexities. 00:33:34.250 --> 00:33:38.180 align:middle line:84% That's what we see here when you take a slice or a line 00:33:38.180 --> 00:33:42.090 align:middle line:84% up from the horizontal axis up through the complexity. 00:33:42.090 --> 00:33:46.780 align:middle line:84% You can have multiple different values in here. 00:33:46.780 --> 00:33:47.280 align:middle line:90% OK. 00:33:47.280 --> 00:33:50.560 align:middle line:90% 00:33:50.560 --> 00:33:52.240 align:middle line:90% That was an example of a model. 00:33:52.240 --> 00:33:53.550 align:middle line:90% Yeah. 00:33:53.550 --> 00:33:56.300 align:middle line:84% AUDIENCE: [INAUDIBLE] despite how complexity is measured 00:33:56.300 --> 00:34:00.442 align:middle line:90% or [INAUDIBLE] definition-- 00:34:00.442 --> 00:34:02.900 align:middle line:84% GEORGE CHURCH: You'll really have to refer to this article. 00:34:02.900 --> 00:34:04.150 align:middle line:90% It's a fairly complicated one. 00:34:04.150 --> 00:34:16.639 align:middle line:84% But the idea here was you take a complicated set of data 00:34:16.639 --> 00:34:19.340 align:middle line:84% generated by a logistic map that we'll 00:34:19.340 --> 00:34:22.429 align:middle line:84% come to in just a few slides, and you 00:34:22.429 --> 00:34:25.520 align:middle line:84% ask-- as you increase the complexity 00:34:25.520 --> 00:34:31.489 align:middle line:84% of the map in a way that appeals to the intuitive 00:34:31.489 --> 00:34:36.320 align:middle line:84% definition of complexity you had in the previous one, 00:34:36.320 --> 00:34:39.620 align:middle line:84% you calculate the number of symmetries, 00:34:39.620 --> 00:34:43.250 align:middle line:84% the physical symmetry that it would take to represent this, 00:34:43.250 --> 00:34:45.980 align:middle line:84% allowing coin tosses as part of the algorithm 00:34:45.980 --> 00:34:48.380 align:middle line:90% for simplifying it. 00:34:48.380 --> 00:34:52.730 align:middle line:84% It's just barred from all the previous definitions. 00:34:52.730 --> 00:34:56.100 align:middle line:84% Crutchfield and coworkers have championed this. 00:34:56.100 --> 00:34:59.060 align:middle line:84% And I would say it's not widely accepted, 00:34:59.060 --> 00:35:00.890 align:middle line:84% but it's certainly not rejected either. 00:35:00.890 --> 00:35:04.400 align:middle line:90% And I find it appealing. 00:35:04.400 --> 00:35:06.260 align:middle line:90% So why do we model? 00:35:06.260 --> 00:35:09.980 align:middle line:84% That was an example of modeling complexity. 00:35:09.980 --> 00:35:14.150 align:middle line:84% I pointed out earlier the models that we 00:35:14.150 --> 00:35:17.990 align:middle line:84% had for a three-dimensional structure. 00:35:17.990 --> 00:35:21.410 align:middle line:84% This course is mainly about measurements. 00:35:21.410 --> 00:35:22.640 align:middle line:90% So why model? 00:35:22.640 --> 00:35:26.270 align:middle line:84% I would argue that when we measure, 00:35:26.270 --> 00:35:27.620 align:middle line:90% we actually must model. 00:35:27.620 --> 00:35:30.200 align:middle line:84% And we need to do that not merely 00:35:30.200 --> 00:35:33.020 align:middle line:84% to understand the biological, chemical data 00:35:33.020 --> 00:35:35.720 align:middle line:90% that we are collecting. 00:35:35.720 --> 00:35:39.020 align:middle line:84% And if I understand, we can have various tests 00:35:39.020 --> 00:35:40.640 align:middle line:90% for our understanding. 00:35:40.640 --> 00:35:43.760 align:middle line:84% Tests might include, but not be limited to, 00:35:43.760 --> 00:35:48.030 align:middle line:84% designing useful modifications of a system. 00:35:48.030 --> 00:35:49.520 align:middle line:84% The course is mainly about systems. 00:35:49.520 --> 00:35:57.500 align:middle line:84% And by using our models to design a new variation 00:35:57.500 --> 00:36:00.830 align:middle line:84% on the system and then obtaining additional data 00:36:00.830 --> 00:36:05.180 align:middle line:84% and modeling that new data, we get a better successful 00:36:05.180 --> 00:36:09.710 align:middle line:84% approximation to what the underlying data means. 00:36:09.710 --> 00:36:13.610 align:middle line:84% And we prove it, often with useful modifications 00:36:13.610 --> 00:36:17.540 align:middle line:84% that can have impact on society or impact 00:36:17.540 --> 00:36:20.360 align:middle line:90% on other research agenda. 00:36:20.360 --> 00:36:23.060 align:middle line:84% Another reason why we model is in order to share data. 00:36:23.060 --> 00:36:26.030 align:middle line:84% Typically, when you look in a database-- 00:36:26.030 --> 00:36:28.730 align:middle line:84% those of you who have looked at biological or chemical 00:36:28.730 --> 00:36:32.600 align:middle line:84% databases, that is not data in the database, 00:36:32.600 --> 00:36:35.420 align:middle line:90% generally speaking. 00:36:35.420 --> 00:36:38.810 align:middle line:90% There are models in there. 00:36:38.810 --> 00:36:40.880 align:middle line:84% And the things that you download are models. 00:36:40.880 --> 00:36:42.630 align:middle line:84% So for example, the sequence that I showed 00:36:42.630 --> 00:36:44.300 align:middle line:84% at the beginning-- that's 76 nucleotides 00:36:44.300 --> 00:36:45.710 align:middle line:90% of A's, C's, G's and P's-- 00:36:45.710 --> 00:36:48.860 align:middle line:84% that's a model that represents our interpretation 00:36:48.860 --> 00:36:50.690 align:middle line:84% of some kind of fluorescent patterns 00:36:50.690 --> 00:36:55.580 align:middle line:84% that we see when we run our instruments on amplified 00:36:55.580 --> 00:36:58.370 align:middle line:84% DNA taken from a variety of organisms. 00:36:58.370 --> 00:37:01.940 align:middle line:90% That's the model. 00:37:01.940 --> 00:37:04.880 align:middle line:84% Similarly, the three-dimensional rotating transfer 00:37:04.880 --> 00:37:08.210 align:middle line:84% RNA that we saw was probably more obviously a model. 00:37:08.210 --> 00:37:15.560 align:middle line:84% There, we integrate the chemistry 00:37:15.560 --> 00:37:20.720 align:middle line:84% of molecular mechanics and the physics of the diffraction 00:37:20.720 --> 00:37:24.020 align:middle line:90% data where the atoms-- 00:37:24.020 --> 00:37:26.390 align:middle line:84% the electron density in a crystal lattice 00:37:26.390 --> 00:37:28.100 align:middle line:90% diffract and make a pattern. 00:37:28.100 --> 00:37:32.450 align:middle line:84% We integrate those models, and that's 00:37:32.450 --> 00:37:34.400 align:middle line:90% what shared in databases. 00:37:34.400 --> 00:37:36.660 align:middle line:84% Not only does it allow us to share data, 00:37:36.660 --> 00:37:39.350 align:middle line:84% but the way we share it is by searching. 00:37:39.350 --> 00:37:42.020 align:middle line:84% Those of you who have searched either biological databases 00:37:42.020 --> 00:37:44.150 align:middle line:84% or the internet with Google or something like that 00:37:44.150 --> 00:37:48.620 align:middle line:84% know how powerful a search can be. 00:37:48.620 --> 00:37:51.530 align:middle line:84% Typically, it is easier to search with a model 00:37:51.530 --> 00:37:54.500 align:middle line:84% and more powerful and more accurate to search with a model 00:37:54.500 --> 00:37:56.960 align:middle line:90% than to search through raw data. 00:37:56.960 --> 00:38:00.890 align:middle line:84% When we merge datasets, we will align them. 00:38:00.890 --> 00:38:04.190 align:middle line:90% We will find redundancies. 00:38:04.190 --> 00:38:10.220 align:middle line:84% We will integrate different concepts. 00:38:10.220 --> 00:38:11.570 align:middle line:90% That requires modeling. 00:38:11.570 --> 00:38:13.058 align:middle line:90% And finally, checking data-- 00:38:13.058 --> 00:38:14.600 align:middle line:84% one of the themes of this course will 00:38:14.600 --> 00:38:18.500 align:middle line:90% be to embrace your outliers. 00:38:18.500 --> 00:38:21.920 align:middle line:84% When you model your raw data and you 00:38:21.920 --> 00:38:26.900 align:middle line:84% find data that are very far away from expectation, 00:38:26.900 --> 00:38:28.470 align:middle line:90% this is a good thing. 00:38:28.470 --> 00:38:33.320 align:middle line:84% It allows you to find errors in your model or errors 00:38:33.320 --> 00:38:36.150 align:middle line:90% in your data or discoveries. 00:38:36.150 --> 00:38:38.120 align:middle line:90% So checking is another. 00:38:38.120 --> 00:38:39.930 align:middle line:84% Finally, integrating-- as I said, 00:38:39.930 --> 00:38:41.870 align:middle line:84% this course is about measures and models. 00:38:41.870 --> 00:38:44.870 align:middle line:84% And it's not just a survey of all the things 00:38:44.870 --> 00:38:47.630 align:middle line:84% that you can do with computation in biology and computational 00:38:47.630 --> 00:38:51.050 align:middle line:84% biology, but most importantly, it's about integration. 00:38:51.050 --> 00:38:57.080 align:middle line:90% 00:38:57.080 --> 00:39:00.260 align:middle line:84% We need more than one data type to make progress 00:39:00.260 --> 00:39:03.290 align:middle line:84% in biology and medicine, agriculture. 00:39:03.290 --> 00:39:05.840 align:middle line:84% And integration is one of the biggest challenges 00:39:05.840 --> 00:39:06.800 align:middle line:90% that we have right now. 00:39:06.800 --> 00:39:10.038 align:middle line:84% It's relatively easy to collect a single homogeneous dataset, 00:39:10.038 --> 00:39:12.080 align:middle line:84% but then to connect that to the rest of the world 00:39:12.080 --> 00:39:16.700 align:middle line:84% is the big challenge that all of you will have. 00:39:16.700 --> 00:39:19.490 align:middle line:84% A theme that will go throughout the course 00:39:19.490 --> 00:39:23.420 align:middle line:84% will be this business about errors, two types of errors, 00:39:23.420 --> 00:39:26.870 align:middle line:90% random and systematic. 00:39:26.870 --> 00:39:28.910 align:middle line:84% And every time that somebody hands you some data 00:39:28.910 --> 00:39:31.452 align:middle line:84% or you collect your own data or you're dealing with something 00:39:31.452 --> 00:39:33.920 align:middle line:84% from the literature, you should immediately 00:39:33.920 --> 00:39:36.242 align:middle line:84% assume there are both random and systematic errors. 00:39:36.242 --> 00:39:38.450 align:middle line:84% You should know which is which and how much there is. 00:39:38.450 --> 00:39:41.420 align:middle line:84% You should not accept anything as being 00:39:41.420 --> 00:39:45.150 align:middle line:90% true without qualification. 00:39:45.150 --> 00:39:48.240 align:middle line:84% Random errors mean that if you repeat the experiment again 00:39:48.240 --> 00:39:52.950 align:middle line:84% and again, you will get slightly different variations 00:39:52.950 --> 00:39:54.810 align:middle line:90% on the error types. 00:39:54.810 --> 00:39:57.210 align:middle line:84% And to a certain extent, they will average out 00:39:57.210 --> 00:39:59.550 align:middle line:90% over enough determination. 00:39:59.550 --> 00:40:01.420 align:middle line:84% Systematic errors, on the other hand-- 00:40:01.420 --> 00:40:03.630 align:middle line:84% you have a high probability of getting almost exactly 00:40:03.630 --> 00:40:06.150 align:middle line:84% the same error again and again, or a very small subset 00:40:06.150 --> 00:40:09.420 align:middle line:84% of a certain class of errors, which means that just doing it 00:40:09.420 --> 00:40:11.430 align:middle line:84% over and over will not improve your statistics 00:40:11.430 --> 00:40:13.980 align:middle line:90% or improve your accuracy. 00:40:13.980 --> 00:40:18.840 align:middle line:84% You need to change paradigms altogether, collecting data 00:40:18.840 --> 00:40:21.750 align:middle line:90% by more than one method. 00:40:21.750 --> 00:40:27.990 align:middle line:84% So those are the types of reasons that we model. 00:40:27.990 --> 00:40:31.650 align:middle line:84% And now we should go through the kinds of models that we'll do. 00:40:31.650 --> 00:40:35.460 align:middle line:84% This is a course that's basically on sequence, 00:40:35.460 --> 00:40:37.560 align:middle line:84% how sequence leads to three-dimensional or even 00:40:37.560 --> 00:40:39.930 align:middle line:84% four-dimensional structures with time, 00:40:39.930 --> 00:40:43.740 align:middle line:84% how three-dimensional structures lead to function, how function 00:40:43.740 --> 00:40:47.740 align:middle line:84% is embedded in complicated systems, 00:40:47.740 --> 00:40:51.540 align:middle line:84% so which models we will be searching, merging, 00:40:51.540 --> 00:40:56.020 align:middle line:84% and checking, as we said in the previous slide. 00:40:56.020 --> 00:40:57.630 align:middle line:90% Which models will we sequence? 00:40:57.630 --> 00:41:01.950 align:middle line:84% We will be making the assumption in dynamic programming, which 00:41:01.950 --> 00:41:05.790 align:middle line:84% is just a fancy way of saying searching and aligning-- 00:41:05.790 --> 00:41:07.680 align:middle line:84% dynamic programming will make the assumption 00:41:07.680 --> 00:41:10.470 align:middle line:84% that sequences are related to one another. 00:41:10.470 --> 00:41:13.740 align:middle line:84% They make the further assumption they are related by ancestry. 00:41:13.740 --> 00:41:16.620 align:middle line:84% That is to say, you mutated them in the laboratory, 00:41:16.620 --> 00:41:19.320 align:middle line:84% or perhaps you were mutated before you 00:41:19.320 --> 00:41:21.210 align:middle line:90% came to the laboratory. 00:41:21.210 --> 00:41:25.342 align:middle line:84% That's dynamic programming that can align sequences 00:41:25.342 --> 00:41:27.300 align:middle line:84% or three-dimensional structures which are quite 00:41:27.300 --> 00:41:28.750 align:middle line:90% different from one another. 00:41:28.750 --> 00:41:30.900 align:middle line:84% A few slides back when we talk about replication, 00:41:30.900 --> 00:41:32.970 align:middle line:90% if two things are replicated-- 00:41:32.970 --> 00:41:35.220 align:middle line:90% showing common ancestry, say-- 00:41:35.220 --> 00:41:36.900 align:middle line:84% the faithfulness of that replication 00:41:36.900 --> 00:41:38.430 align:middle line:84% is key to-- the dynamic programming 00:41:38.430 --> 00:41:41.340 align:middle line:84% is key to your accepting that something is actually 00:41:41.340 --> 00:41:41.880 align:middle line:90% replicated. 00:41:41.880 --> 00:41:45.450 align:middle line:84% If it changes into something completely new 00:41:45.450 --> 00:41:48.810 align:middle line:84% in the process of replication, then it's 00:41:48.810 --> 00:41:52.770 align:middle line:84% unlikely it will be able to maintain that. 00:41:52.770 --> 00:41:54.630 align:middle line:84% Three-dimensional structure-- we will 00:41:54.630 --> 00:41:57.240 align:middle line:84% be talking about motifs, catalysis, 00:41:57.240 --> 00:41:58.475 align:middle line:90% complementary surfaces. 00:41:58.475 --> 00:41:59.850 align:middle line:84% I'll give you a beautiful example 00:41:59.850 --> 00:42:02.280 align:middle line:84% of complimentary surface in a few slides. 00:42:02.280 --> 00:42:05.580 align:middle line:84% And all of this is dealing with the continuous functions 00:42:05.580 --> 00:42:11.520 align:middle line:84% of energy and kinetics where different complementary 00:42:11.520 --> 00:42:15.630 align:middle line:84% surfaces will bind to one another with very specific rate 00:42:15.630 --> 00:42:20.160 align:middle line:84% constants covering many orders of magnitude. 00:42:20.160 --> 00:42:25.530 align:middle line:84% These energetic and kinetic phenomenon 00:42:25.530 --> 00:42:28.020 align:middle line:84% is what underlies the functional genomics. 00:42:28.020 --> 00:42:30.060 align:middle line:84% When we functional genomic data, one 00:42:30.060 --> 00:42:31.890 align:middle line:84% of the ways we model it is we ask 00:42:31.890 --> 00:42:37.440 align:middle line:84% whether phenomena that we're studying 00:42:37.440 --> 00:42:40.770 align:middle line:84% inside our bodies or inside of microorganisms, so forth-- 00:42:40.770 --> 00:42:42.565 align:middle line:84% the protein levels, the RNA levels, 00:42:42.565 --> 00:42:44.190 align:middle line:84% and so forth, where they go up and down 00:42:44.190 --> 00:42:46.200 align:middle line:84% together means that they're a cluster. 00:42:46.200 --> 00:42:47.910 align:middle line:90% They have common properties. 00:42:47.910 --> 00:42:51.600 align:middle line:84% Does this common properties reflect yet another aspect 00:42:51.600 --> 00:42:54.540 align:middle line:84% of them, like their common functions 00:42:54.540 --> 00:42:58.440 align:middle line:84% or their common mechanisms, for being coregulated? 00:42:58.440 --> 00:43:04.590 align:middle line:84% In systems biology, we have these qualitative diagrams that 00:43:04.590 --> 00:43:09.240 align:middle line:84% tell us about the all or none, or Boolean, 00:43:09.240 --> 00:43:13.650 align:middle line:90% which means logical 0's and 1's. 00:43:13.650 --> 00:43:18.030 align:middle line:84% Or we can treat them as continuous differential 00:43:18.030 --> 00:43:21.390 align:middle line:84% equations or stochastic where you 00:43:21.390 --> 00:43:23.940 align:middle line:84% have some of the power of the differential equations, 00:43:23.940 --> 00:43:26.310 align:middle line:84% but you deal with individual molecules 00:43:26.310 --> 00:43:30.960 align:middle line:84% or individual organisms and populations. 00:43:30.960 --> 00:43:33.780 align:middle line:90% Organisms can be stochastic. 00:43:33.780 --> 00:43:36.480 align:middle line:90% Molecules can be stochastic. 00:43:36.480 --> 00:43:39.420 align:middle line:84% And another thing-- again, this theme of optimization-- 00:43:39.420 --> 00:43:45.840 align:middle line:84% we can ask whether a network or network component is optimal. 00:43:45.840 --> 00:43:51.210 align:middle line:84% Is it optimal due to the past history of the organism, 00:43:51.210 --> 00:43:55.230 align:middle line:84% or is it something we can set a goal of a biotech process 00:43:55.230 --> 00:43:57.600 align:middle line:84% to optimize it for a new function? 00:43:57.600 --> 00:44:01.140 align:middle line:84% Linear programming is the mathematical tool 00:44:01.140 --> 00:44:03.190 align:middle line:84% that one can use for studying this. 00:44:03.190 --> 00:44:10.050 align:middle line:84% It's common in economic algorithms. 00:44:10.050 --> 00:44:14.970 align:middle line:84% So we were talking about modeling in this whole realm, 00:44:14.970 --> 00:44:18.390 align:middle line:84% going from minimal life sequences 00:44:18.390 --> 00:44:21.300 align:middle line:84% through the catalysis that involves interactions 00:44:21.300 --> 00:44:24.570 align:middle line:84% of three-dimensional structures, functional genomics, 00:44:24.570 --> 00:44:29.580 align:middle line:84% and the optimality that we get that's required, 00:44:29.580 --> 00:44:31.980 align:middle line:84% as we will see momentarily, for getting 00:44:31.980 --> 00:44:35.860 align:middle line:90% single molecules to work. 00:44:35.860 --> 00:44:37.090 align:middle line:90% What's our parts list? 00:44:37.090 --> 00:44:40.180 align:middle line:84% We're going to try to start simple, 00:44:40.180 --> 00:44:43.320 align:middle line:84% but put the simplicity in the context of the big picture. 00:44:43.320 --> 00:44:48.520 align:middle line:84% The big picture for the atoms is the periodic table 00:44:48.520 --> 00:44:53.430 align:middle line:90% of the more than 100 elements. 00:44:53.430 --> 00:44:55.590 align:middle line:90% We have a very short list here-- 00:44:55.590 --> 00:44:58.830 align:middle line:84% sodium, potassium, ion, chlorine, calcium, magnesium, 00:44:58.830 --> 00:45:02.380 align:middle line:84% molybdenum, manganese, sulfur, selenium, copper, nickel, 00:45:02.380 --> 00:45:08.790 align:middle line:84% cobalt, and silicon, which are useful in many species. 00:45:08.790 --> 00:45:13.230 align:middle line:84% If you have to pick something that's 00:45:13.230 --> 00:45:19.620 align:middle line:84% related to us as a minimal of biochemical system that shows 00:45:19.620 --> 00:45:21.740 align:middle line:84% some of the replicative properties 00:45:21.740 --> 00:45:24.150 align:middle line:84% and evolutionary properties of life, 00:45:24.150 --> 00:45:26.370 align:middle line:90% you would say RNA-based life. 00:45:26.370 --> 00:45:29.820 align:middle line:84% One of the breakthroughs in experimental science 00:45:29.820 --> 00:45:31.650 align:middle line:84% over the last couple of decades is 00:45:31.650 --> 00:45:34.500 align:middle line:84% recognizing that RNA, like proteins, 00:45:34.500 --> 00:45:36.210 align:middle line:90% has catalytic capabilities. 00:45:36.210 --> 00:45:39.420 align:middle line:84% It has the potential to have been one 00:45:39.420 --> 00:45:41.220 align:middle line:90% of the early replicating units. 00:45:41.220 --> 00:45:45.040 align:middle line:84% Just for the sake of describing a simple system, 00:45:45.040 --> 00:45:48.683 align:middle line:84% let's look at something made up of five elements. 00:45:48.683 --> 00:45:50.100 align:middle line:84% It's not necessarily the simplest, 00:45:50.100 --> 00:45:52.290 align:middle line:84% but it's just something to think about. 00:45:52.290 --> 00:45:57.720 align:middle line:84% These five elements can be in the presence 00:45:57.720 --> 00:46:02.370 align:middle line:84% of an environment which can be composed of the same five 00:46:02.370 --> 00:46:02.940 align:middle line:90% elements. 00:46:02.940 --> 00:46:06.000 align:middle line:84% The environment would be very simple, 00:46:06.000 --> 00:46:07.450 align:middle line:84% that complexity versus simplicity. 00:46:07.450 --> 00:46:12.510 align:middle line:84% It would be water, ammonium, positively charged 00:46:12.510 --> 00:46:15.450 align:middle line:84% ions, nucleotide triphosphate, negatively charged 00:46:15.450 --> 00:46:19.200 align:middle line:84% ions, which are precursors to making polymers, 00:46:19.200 --> 00:46:22.500 align:middle line:84% and then possibly lipids, the fatty substances 00:46:22.500 --> 00:46:24.330 align:middle line:90% that bound membranes. 00:46:24.330 --> 00:46:30.000 align:middle line:84% An example of catalyzed RNA polymerization 00:46:30.000 --> 00:46:31.890 align:middle line:84% is in this article, and then I'll 00:46:31.890 --> 00:46:35.040 align:middle line:84% deal with another one that, rather 00:46:35.040 --> 00:46:36.690 align:middle line:84% than using nucleotide triphosphates, 00:46:36.690 --> 00:46:40.530 align:middle line:84% uses slightly larger RNA precursors. 00:46:40.530 --> 00:46:42.750 align:middle line:84% Now let's start doing the toggling back and forth 00:46:42.750 --> 00:46:45.210 align:middle line:84% between living systems and computational systems 00:46:45.210 --> 00:46:47.850 align:middle line:84% so that, whether you're from a computational background 00:46:47.850 --> 00:46:49.680 align:middle line:84% or biological or both, you'll see 00:46:49.680 --> 00:46:52.590 align:middle line:90% some interesting relationships. 00:46:52.590 --> 00:46:58.470 align:middle line:84% Here we talked about a minimal biological system 00:46:58.470 --> 00:46:59.640 align:middle line:90% with five elements. 00:46:59.640 --> 00:47:01.848 align:middle line:84% Here we're going to talk about some minimal problems 00:47:01.848 --> 00:47:03.390 align:middle line:84% with a very small number of elements. 00:47:03.390 --> 00:47:07.350 align:middle line:84% Basically, they're limited to a single line of code each. 00:47:07.350 --> 00:47:11.520 align:middle line:84% And they do something which is related to the topic here. 00:47:11.520 --> 00:47:18.210 align:middle line:84% Replication is an exponential process, exponentially growing. 00:47:18.210 --> 00:47:22.800 align:middle line:84% If something is autocatalytic is another way of describing it. 00:47:22.800 --> 00:47:26.910 align:middle line:84% So we've used that as the theme for our minimal programs. 00:47:26.910 --> 00:47:30.090 align:middle line:84% What is this exponential function? 00:47:30.090 --> 00:47:33.010 align:middle line:84% We give it a very specific argument. 00:47:33.010 --> 00:47:33.870 align:middle line:90% The argument is 1. 00:47:33.870 --> 00:47:38.460 align:middle line:84% So this is e to the 1 power, e being this number here 00:47:38.460 --> 00:47:43.410 align:middle line:84% that's represented the number of times, about 2.7. 00:47:43.410 --> 00:47:47.040 align:middle line:84% The four languages that we're demonstrating here 00:47:47.040 --> 00:47:52.782 align:middle line:84% is Perl, Excel, Fortran 77, and Mathematica. 00:47:52.782 --> 00:47:54.990 align:middle line:84% In this course, we'll mainly use Perl and Mathematica 00:47:54.990 --> 00:47:58.450 align:middle line:84% for reasons that will be evident in the next couple of slides. 00:47:58.450 --> 00:48:01.020 align:middle line:84% But I just want to show you these different ones. 00:48:01.020 --> 00:48:06.240 align:middle line:84% And in the theme of accuracy of replication, 00:48:06.240 --> 00:48:12.930 align:middle line:84% how faithfully is a particular string handled, either 00:48:12.930 --> 00:48:16.470 align:middle line:84% the string of nucleotides in a simple life form 00:48:16.470 --> 00:48:19.380 align:middle line:84% or the string of digits in a number-- 00:48:19.380 --> 00:48:21.990 align:middle line:84% and you can see that the internal representation 00:48:21.990 --> 00:48:25.190 align:middle line:84% of the math that goes on inside the computer 00:48:25.190 --> 00:48:26.940 align:middle line:84% has to be done digitally even though these 00:48:26.940 --> 00:48:29.023 align:middle line:84% are transcendental numbers which have an arbitrary 00:48:29.023 --> 00:48:30.930 align:middle line:90% number of digits. 00:48:30.930 --> 00:48:34.950 align:middle line:84% And you can see that some programs are typically limited. 00:48:34.950 --> 00:48:38.010 align:middle line:84% And you can even guess the number of bits representing it 00:48:38.010 --> 00:48:38.700 align:middle line:90% internally. 00:48:38.700 --> 00:48:40.770 align:middle line:84% Here down at the bottom of slide 17, 00:48:40.770 --> 00:48:45.270 align:middle line:84% you can see some of the nitty-gritty detail 00:48:45.270 --> 00:48:46.950 align:middle line:84% of how these things are represented 00:48:46.950 --> 00:48:51.270 align:middle line:84% as 0's and 1's representing the [INAUDIBLE],, 00:48:51.270 --> 00:48:53.500 align:middle line:90% the exponent, and so on. 00:48:53.500 --> 00:48:56.580 align:middle line:84% But you can see some of these programs, when they print, 00:48:56.580 --> 00:49:01.560 align:middle line:84% when they actually print, the program is not aware of 00:49:01.560 --> 00:49:05.400 align:middle line:84% or the programmer was not aware of exactly what 00:49:05.400 --> 00:49:06.930 align:middle line:90% the internal representation was. 00:49:06.930 --> 00:49:10.800 align:middle line:84% And so all these trailing 0's are incorrect. 00:49:10.800 --> 00:49:13.360 align:middle line:90% 00:49:13.360 --> 00:49:16.980 align:middle line:84% The winner in this, of course, is Mathematica, 00:49:16.980 --> 00:49:19.680 align:middle line:84% here to ask for the e to the 1 power. 00:49:19.680 --> 00:49:21.276 align:middle line:90% You say, let's give it-- 00:49:21.276 --> 00:49:23.820 align:middle line:90% 00:49:23.820 --> 00:49:28.620 align:middle line:84% this n bracket says let's start giving arbitrary precision 00:49:28.620 --> 00:49:33.540 align:middle line:84% where the numeric, the n numeric, is 100 digits. 00:49:33.540 --> 00:49:37.080 align:middle line:84% And this seems like a stunt in this particular case. 00:49:37.080 --> 00:49:40.260 align:middle line:84% But actually, when you start doing a series of calculations 00:49:40.260 --> 00:49:43.140 align:middle line:84% where errors can accumulate, the ability 00:49:43.140 --> 00:49:45.900 align:middle line:84% to go into arbitrary precision will 00:49:45.900 --> 00:49:50.940 align:middle line:84% allow you to prevent a catastrophic imprecision. 00:49:50.940 --> 00:49:53.470 align:middle line:84% And what happens is, in a mathematical calculation, 00:49:53.470 --> 00:49:55.990 align:middle line:84% if it needs a little more precision than you initially 00:49:55.990 --> 00:49:59.500 align:middle line:84% anticipated, it will go out and get it 00:49:59.500 --> 00:50:01.240 align:middle line:84% on its own, which is quite remarkable. 00:50:01.240 --> 00:50:04.072 align:middle line:84% Try to do that in any of these other languages. 00:50:04.072 --> 00:50:05.530 align:middle line:84% They will typically give you what's 00:50:05.530 --> 00:50:08.470 align:middle line:84% called an underflow or an overflow error 00:50:08.470 --> 00:50:13.240 align:middle line:84% and either stop or just make mistakes. 00:50:13.240 --> 00:50:14.470 align:middle line:90% OK. 00:50:14.470 --> 00:50:18.070 align:middle line:84% That's my first advertisement for Mathematica. 00:50:18.070 --> 00:50:23.020 align:middle line:84% You can see all of these are very simple programs that belie 00:50:23.020 --> 00:50:27.010 align:middle line:84% and hide underneath the very complicated electronics 00:50:27.010 --> 00:50:31.048 align:middle line:84% and algorithms that are built in to accomplish this. 00:50:31.048 --> 00:50:32.590 align:middle line:84% I'll give you a feeling for what that 00:50:32.590 --> 00:50:34.480 align:middle line:90% is in just a couple of slides. 00:50:34.480 --> 00:50:36.220 align:middle line:90% So back to self-replication. 00:50:36.220 --> 00:50:38.890 align:middle line:84% We're toggling back and forth between simple biological and 00:50:38.890 --> 00:50:40.810 align:middle line:90% computational systems. 00:50:40.810 --> 00:50:48.160 align:middle line:84% We have-- here we've represented tri and hexanucleotides, three 00:50:48.160 --> 00:50:49.840 align:middle line:90% nucleotides and six. 00:50:49.840 --> 00:50:53.170 align:middle line:84% Again, remember, these letters are crude-- 00:50:53.170 --> 00:50:55.420 align:middle line:84% or I should say simple representations 00:50:55.420 --> 00:50:57.850 align:middle line:84% of electron density involving dozens 00:50:57.850 --> 00:51:01.870 align:middle line:84% of atoms in the shape of a cytosine, or a C, or guanine, 00:51:01.870 --> 00:51:04.665 align:middle line:84% or a G. It's important that you know 00:51:04.665 --> 00:51:09.610 align:middle line:84% that there are two strands in DNA, typically, 00:51:09.610 --> 00:51:11.110 align:middle line:90% or even many RNAs-- 00:51:11.110 --> 00:51:14.380 align:middle line:84% in this particular example, whether this is RNA or DNA. 00:51:14.380 --> 00:51:17.470 align:middle line:84% And in order to indicate the orientation of the strand, 00:51:17.470 --> 00:51:19.720 align:middle line:90% it has a directionality. 00:51:19.720 --> 00:51:22.090 align:middle line:84% We indicate 5-prime, which is the name 00:51:22.090 --> 00:51:27.610 align:middle line:84% of an atom in the ribose, but the details 00:51:27.610 --> 00:51:30.220 align:middle line:84% are not important from a computational standpoint. 00:51:30.220 --> 00:51:32.650 align:middle line:84% It's just a way of indicating this is one end of the RNA. 00:51:32.650 --> 00:51:36.660 align:middle line:84% And a CCG is not the same as a GCC. 00:51:36.660 --> 00:51:39.190 align:middle line:84% The 5-prime indicates the 5-prime end of that. 00:51:39.190 --> 00:51:42.250 align:middle line:84% We take two of these, which are basically identical, 00:51:42.250 --> 00:51:45.280 align:middle line:84% and they will ligate together spontaneously 00:51:45.280 --> 00:51:48.070 align:middle line:84% if you set up the chemistry right 00:51:48.070 --> 00:51:55.000 align:middle line:84% to make these two trinucleotides into a hexanucleotide, CCG CCG. 00:51:55.000 --> 00:51:57.410 align:middle line:84% In the presence of a complement here, 00:51:57.410 --> 00:52:01.270 align:middle line:84% which is CGG CGG, a different sequence-- 00:52:01.270 --> 00:52:04.180 align:middle line:84% I've tried to emphasize that by making it capital and green-- 00:52:04.180 --> 00:52:08.890 align:middle line:90% that would bind to-- 00:52:08.890 --> 00:52:10.630 align:middle line:84% these two trinucleotides would bind 00:52:10.630 --> 00:52:13.420 align:middle line:84% to this, aligned by Watson-Crick base pairs. 00:52:13.420 --> 00:52:16.960 align:middle line:84% The rules for Watson-Crick base pairs, many of you may know. 00:52:16.960 --> 00:52:21.310 align:middle line:84% A pairs with T, and C pairs with G. 00:52:21.310 --> 00:52:23.650 align:middle line:84% And that will catalyze this process. 00:52:23.650 --> 00:52:27.797 align:middle line:84% It will speed up the kinetics in which these trinucleotides 00:52:27.797 --> 00:52:29.380 align:middle line:84% [? turn ?] [? into ?] hexanucleotides. 00:52:29.380 --> 00:52:30.770 align:middle line:90% Now, here it gets interesting. 00:52:30.770 --> 00:52:33.460 align:middle line:84% This hexanucleotide now drops down here 00:52:33.460 --> 00:52:36.790 align:middle line:84% and speeds up the process of this different trinucleotide 00:52:36.790 --> 00:52:40.540 align:middle line:84% [INAUDIBLE] and forming the original one that 00:52:40.540 --> 00:52:41.840 align:middle line:90% catalyzes the first reaction. 00:52:41.840 --> 00:52:47.055 align:middle line:84% So this catalyzes, speeds up the first reaction, which 00:52:47.055 --> 00:52:48.430 align:middle line:84% produces a product that speeds up 00:52:48.430 --> 00:52:52.010 align:middle line:84% the second, which produces the first catalyst, and so on. 00:52:52.010 --> 00:52:55.150 align:middle line:84% This is called a hypercycle, championed 00:52:55.150 --> 00:52:58.450 align:middle line:90% by Eigen and Schuster. 00:52:58.450 --> 00:53:01.810 align:middle line:84% And this is probably one of the simplest examples of it. 00:53:01.810 --> 00:53:04.110 align:middle line:84% And I think it gives you a feeling for how 00:53:04.110 --> 00:53:06.640 align:middle line:84% a variety of interdependent biochemical processes 00:53:06.640 --> 00:53:09.040 align:middle line:84% can result in autocatalysis where 00:53:09.040 --> 00:53:13.060 align:middle line:84% you get these exponential cycles which we recognize as something 00:53:13.060 --> 00:53:15.100 align:middle line:90% very similar to replication. 00:53:15.100 --> 00:53:17.530 align:middle line:90% This would have a very-- 00:53:17.530 --> 00:53:20.140 align:middle line:84% this could have a very high probability of replication. 00:53:20.140 --> 00:53:23.980 align:middle line:84% It could be very faithful, but its complexity is low. 00:53:23.980 --> 00:53:26.590 align:middle line:84% The input complexity is low, and the output complexity is low. 00:53:26.590 --> 00:53:30.490 align:middle line:90% That's how we would qualify it. 00:53:30.490 --> 00:53:37.930 align:middle line:84% Toggling back to computing and simple examples, 00:53:37.930 --> 00:53:40.780 align:middle line:84% I've given some simple examples already of Perl in Mathematica. 00:53:40.780 --> 00:53:45.490 align:middle line:84% But in order to give a few more here on slide 19, 00:53:45.490 --> 00:53:48.040 align:middle line:84% why did we choose these two for this course? 00:53:48.040 --> 00:53:52.660 align:middle line:84% It's been my experience, both in our laboratory and research 00:53:52.660 --> 00:53:54.670 align:middle line:84% environment, and also in this class, 00:53:54.670 --> 00:53:57.430 align:middle line:84% that these are two of the easiest languages to learn. 00:53:57.430 --> 00:53:59.710 align:middle line:84% That doesn't mean that they're absolutely easy, 00:53:59.710 --> 00:54:03.320 align:middle line:84% but the learning curve is very simple. 00:54:03.320 --> 00:54:04.990 align:middle line:90% It's very fast. 00:54:04.990 --> 00:54:10.630 align:middle line:84% And you can, by example, change programs 00:54:10.630 --> 00:54:12.070 align:middle line:84% that are working into things that 00:54:12.070 --> 00:54:14.500 align:middle line:90% do what you want very quickly. 00:54:14.500 --> 00:54:16.858 align:middle line:84% It's a high-level language in a sense-- 00:54:16.858 --> 00:54:18.400 align:middle line:84% the higher the level of the language, 00:54:18.400 --> 00:54:20.303 align:middle line:84% the closer you are to English conversation. 00:54:20.303 --> 00:54:21.970 align:middle line:84% The lower the level of language you are, 00:54:21.970 --> 00:54:27.100 align:middle line:84% the closer you are to bits, 0's and 1's, 00:54:27.100 --> 00:54:32.120 align:middle line:84% or the actual electronics inside of computers. 00:54:32.120 --> 00:54:34.180 align:middle line:84% So in the hierarchy, Perl and Mathematica 00:54:34.180 --> 00:54:36.040 align:middle line:90% are very high up there. 00:54:36.040 --> 00:54:38.275 align:middle line:90% Yeah. 00:54:38.275 --> 00:54:39.900 align:middle line:84% AUDIENCE: Where do you get the password 00:54:39.900 --> 00:54:42.302 align:middle line:90% to download Mathematica? 00:54:42.302 --> 00:54:44.760 align:middle line:84% GEORGE CHURCH: You will get that from your section teaching 00:54:44.760 --> 00:54:45.260 align:middle line:90% fellow. 00:54:45.260 --> 00:54:48.380 align:middle line:90% 00:54:48.380 --> 00:54:49.652 align:middle line:90% Yeah. 00:54:49.652 --> 00:54:51.110 align:middle line:84% So the question is where do you get 00:54:51.110 --> 00:54:53.510 align:middle line:84% the password for Mathematica for this course, 00:54:53.510 --> 00:54:55.830 align:middle line:84% and the teaching fellows are responsible for that. 00:54:55.830 --> 00:54:56.880 align:middle line:90% And that's been tested. 00:54:56.880 --> 00:54:57.672 align:middle line:90% I think that works. 00:54:57.672 --> 00:55:04.780 align:middle line:90% 00:55:04.780 --> 00:55:06.580 align:middle line:90% So Perl is freely-- 00:55:06.580 --> 00:55:07.750 align:middle line:90% it's open source. 00:55:07.750 --> 00:55:09.670 align:middle line:90% It's interesting in that regard. 00:55:09.670 --> 00:55:13.540 align:middle line:84% I think many of you will find the concept of open source 00:55:13.540 --> 00:55:16.060 align:middle line:90% where anyone can-- 00:55:16.060 --> 00:55:18.790 align:middle line:84% when software breaks or needs to be expanded, 00:55:18.790 --> 00:55:22.570 align:middle line:84% or it needs to be understood, it's there for your inspection 00:55:22.570 --> 00:55:25.960 align:middle line:84% rather than hidden behind some corporate doors. 00:55:25.960 --> 00:55:27.760 align:middle line:90% Very interesting aspect of Perl. 00:55:27.760 --> 00:55:34.240 align:middle line:84% Mathematica is not open source as far as I know. 00:55:34.240 --> 00:55:36.490 align:middle line:84% It has some interesting redeeming features, though. 00:55:36.490 --> 00:55:39.400 align:middle line:90% It's very strong on math. 00:55:39.400 --> 00:55:43.030 align:middle line:90% No surprise given its name. 00:55:43.030 --> 00:55:46.750 align:middle line:84% In particular, it's both symbolic and numeric 00:55:46.750 --> 00:55:48.250 align:middle line:90% and in graphics. 00:55:48.250 --> 00:55:50.930 align:middle line:84% And we'll have some examples of that in just a moment. 00:55:50.930 --> 00:55:52.830 align:middle line:84% It can do things that are hard to do 00:55:52.830 --> 00:55:56.170 align:middle line:90% in Perl and other languages. 00:55:56.170 --> 00:55:58.840 align:middle line:84% Perl is also very strong for web applications. 00:55:58.840 --> 00:56:02.980 align:middle line:84% Many of the most amazing things which are done on the web 00:56:02.980 --> 00:56:06.620 align:middle line:90% have Perl behind the scenes. 00:56:06.620 --> 00:56:12.460 align:middle line:84% Now, when we-- a further comparison of-- 00:56:12.460 --> 00:56:16.840 align:middle line:84% this is the dark side of computation and biology, 00:56:16.840 --> 00:56:21.190 align:middle line:84% showing another analogy, which is parasites. 00:56:21.190 --> 00:56:24.550 align:middle line:84% We have parasitic computer and viral-- 00:56:24.550 --> 00:56:26.350 align:middle line:90% biological viruses. 00:56:26.350 --> 00:56:33.550 align:middle line:84% This one-- in past years, I had this little bit of a computer 00:56:33.550 --> 00:56:34.733 align:middle line:90% virus-- 00:56:34.733 --> 00:56:35.900 align:middle line:90% this is not the whole thing. 00:56:35.900 --> 00:56:37.317 align:middle line:84% This is just a little piece of it. 00:56:37.317 --> 00:56:38.620 align:middle line:90% It's fairly short code. 00:56:38.620 --> 00:56:42.790 align:middle line:84% And this was a very nasty one at the time, quite a while ago. 00:56:42.790 --> 00:56:46.162 align:middle line:84% And I had it in the PowerPoint presentation. 00:56:46.162 --> 00:56:47.620 align:middle line:84% When people would download it, they 00:56:47.620 --> 00:56:48.787 align:middle line:90% would then send me an email. 00:56:48.787 --> 00:56:51.590 align:middle line:90% 00:56:51.590 --> 00:56:57.190 align:middle line:84% I'm sorry to inform you you have a virus in your PowerPoint. 00:56:57.190 --> 00:56:59.620 align:middle line:84% And so this year, I've upgraded the PowerPoint. 00:56:59.620 --> 00:57:03.160 align:middle line:84% So this is actually an image rather than the actual text. 00:57:03.160 --> 00:57:05.440 align:middle line:84% And so your viral detectors will not 00:57:05.440 --> 00:57:08.860 align:middle line:84% detect this unless they're a lot smarter than I think they are. 00:57:08.860 --> 00:57:13.100 align:middle line:84% This, on the other hand, is the costs of these viruses. 00:57:13.100 --> 00:57:14.920 align:middle line:90% This is not a laughing matter. 00:57:14.920 --> 00:57:18.040 align:middle line:84% This is 4 times the cost of the entire Human Genome 00:57:18.040 --> 00:57:20.380 align:middle line:90% Project per year. 00:57:20.380 --> 00:57:23.140 align:middle line:84% The Genome Project took us about 20 years to get going. 00:57:23.140 --> 00:57:25.090 align:middle line:84% Every year, we spend four times that amount 00:57:25.090 --> 00:57:27.370 align:middle line:84% on computer viruses, which are, as far as I can tell, 00:57:27.370 --> 00:57:29.890 align:middle line:90% completely frivolous. 00:57:29.890 --> 00:57:32.290 align:middle line:90% Even more serious is this. 00:57:32.290 --> 00:57:33.790 align:middle line:90% Now, this is real text. 00:57:33.790 --> 00:57:35.110 align:middle line:90% This is not an image. 00:57:35.110 --> 00:57:37.480 align:middle line:90% It looked like an image to you. 00:57:37.480 --> 00:57:38.860 align:middle line:90% It's very serious. 00:57:38.860 --> 00:57:43.180 align:middle line:84% And I sincerely hope that some of the people in this class 00:57:43.180 --> 00:57:49.840 align:middle line:84% make a contribution to the intellectual avenues which 00:57:49.840 --> 00:57:53.260 align:middle line:84% will ultimately lead to the defeat of the AIDS virus 00:57:53.260 --> 00:57:56.830 align:middle line:84% and various other viruses and bacteria like them that cause 00:57:56.830 --> 00:58:01.360 align:middle line:90% so much suffering in the world. 00:58:01.360 --> 00:58:03.100 align:middle line:90% 20 million have died. 00:58:03.100 --> 00:58:08.620 align:middle line:84% This is worse than the Black Plague and the 1918 influenza 00:58:08.620 --> 00:58:10.990 align:middle line:90% epidemic. 00:58:10.990 --> 00:58:14.890 align:middle line:84% And the analogy here between this part of the virus-- 00:58:14.890 --> 00:58:19.180 align:middle line:84% this is a little piece of the viral code in symbols of the 20 00:58:19.180 --> 00:58:21.580 align:middle line:90% amino acids single-letter code. 00:58:21.580 --> 00:58:25.330 align:middle line:84% This is a little piece of the computer virus. 00:58:25.330 --> 00:58:28.270 align:middle line:84% And I've highlighted here the command copy 00:58:28.270 --> 00:58:29.800 align:middle line:90% in the computer virus. 00:58:29.800 --> 00:58:34.720 align:middle line:84% This is just part of what it does to get this particular VBS 00:58:34.720 --> 00:58:36.700 align:middle line:84% script, the thing you're seeing here, 00:58:36.700 --> 00:58:39.190 align:middle line:84% into some other part of your directory. 00:58:39.190 --> 00:58:44.450 align:middle line:84% And this, essentially, is part of the copy command in the AIDS 00:58:44.450 --> 00:58:44.950 align:middle line:90% virus. 00:58:44.950 --> 00:58:46.360 align:middle line:90% This is the polymerase. 00:58:46.360 --> 00:58:49.600 align:middle line:84% The protein is responsible for making copies 00:58:49.600 --> 00:58:52.660 align:middle line:90% of the code for the virus. 00:58:52.660 --> 00:58:54.430 align:middle line:84% And highlighted in red here are some 00:58:54.430 --> 00:58:56.860 align:middle line:84% of the mutations that make this particular AIDS 00:58:56.860 --> 00:59:00.610 align:middle line:84% virus resistant to the drugs which have been instrumental, 00:59:00.610 --> 00:59:05.680 align:middle line:84% in some cases, in taming AIDS temporarily, and at great cost. 00:59:05.680 --> 00:59:10.240 align:middle line:84% Clearly, vaccines, and hopefully other public health measures, 00:59:10.240 --> 00:59:12.940 align:middle line:90% will be the answer. 00:59:12.940 --> 00:59:14.840 align:middle line:84% Here are some other conceptual connections. 00:59:14.840 --> 00:59:17.770 align:middle line:84% These are not meant to be dogma or to limit you in any way 00:59:17.770 --> 00:59:21.400 align:middle line:84% but hopefully to expand the way you think about these things. 00:59:21.400 --> 00:59:23.230 align:middle line:84% Obviously, as with other analogies, 00:59:23.230 --> 00:59:24.200 align:middle line:90% they will break down. 00:59:24.200 --> 00:59:26.690 align:middle line:84% But let's just follow for a moment. 00:59:26.690 --> 00:59:28.660 align:middle line:84% What we're calling the instruction set, 00:59:28.660 --> 00:59:34.030 align:middle line:84% the concept of an instruction set, in computers is a program, 00:59:34.030 --> 00:59:37.030 align:middle line:90% in organisms is a genome. 00:59:37.030 --> 00:59:40.450 align:middle line:84% We've already said bits, so 0's and 1's, A, C, G and T. 00:59:40.450 --> 00:59:43.100 align:middle line:84% The stable memory, the thing that you can depend on 00:59:43.100 --> 00:59:45.100 align:middle line:84% but is a little bit slow for access, 00:59:45.100 --> 00:59:48.790 align:middle line:84% are disks and tapes in computers, and they're DNA, 00:59:48.790 --> 00:59:51.220 align:middle line:90% or in some cases RNA genomes. 00:59:51.220 --> 00:59:55.515 align:middle line:84% That then-- you take it out of the slow memory, stable memory, 00:59:55.515 --> 00:59:56.890 align:middle line:84% and move it into something that's 00:59:56.890 --> 00:59:59.410 align:middle line:84% more active, more volatile, which 00:59:59.410 --> 01:00:02.330 align:middle line:84% is random-access memory in computers 01:00:02.330 --> 01:00:06.920 align:middle line:90% and the RNA in organisms. 01:00:06.920 --> 01:00:11.570 align:middle line:84% The environment for computers tends to be complex-- 01:00:11.570 --> 01:00:14.690 align:middle line:84% it's the internet sockets and people 01:00:14.690 --> 01:00:18.500 align:middle line:84% banging on the keyboard-- while organisms can be very simple. 01:00:18.500 --> 01:00:20.565 align:middle line:84% I gave some examples earlier where it's complex, 01:00:20.565 --> 01:00:22.940 align:middle line:84% but it can be as simple as water and salts, in which they 01:00:22.940 --> 01:00:25.490 align:middle line:84% can replicate very complicated structures. 01:00:25.490 --> 01:00:29.600 align:middle line:90% The input and output can be-- 01:00:29.600 --> 01:00:31.670 align:middle line:84% the input can be analog and converted 01:00:31.670 --> 01:00:37.310 align:middle line:84% to digital in this process, digital analog for output-- 01:00:37.310 --> 01:00:39.770 align:middle line:84% the analog, say, of your screens. 01:00:39.770 --> 01:00:43.430 align:middle line:84% And the input-output is governed by these proteins 01:00:43.430 --> 01:00:48.050 align:middle line:84% at the end of the central dogma in organisms. 01:00:48.050 --> 01:00:53.640 align:middle line:90% 01:00:53.640 --> 01:00:57.150 align:middle line:84% When we make these complicated systems from simple things, 01:00:57.150 --> 01:01:03.210 align:middle line:84% they go from so-called monomers, which really combine 01:01:03.210 --> 01:01:06.090 align:middle line:84% into polymers, which need not be linear polymers, 01:01:06.090 --> 01:01:08.400 align:middle line:84% although they often are in biological systems-- 01:01:08.400 --> 01:01:12.310 align:middle line:84% basically, you go from minerals to chips in computers. 01:01:12.310 --> 01:01:14.130 align:middle line:84% These are replicated in factories. 01:01:14.130 --> 01:01:17.730 align:middle line:84% The factories or cells can be as small as femtolitre. 01:01:17.730 --> 01:01:20.700 align:middle line:84% Remember, this prefix, this femto, is 10 to the -15. 01:01:20.700 --> 01:01:23.340 align:middle line:90% That's about 1 cubic micron. 01:01:23.340 --> 01:01:28.110 align:middle line:84% Very small factories, very amazing productivity 01:01:28.110 --> 01:01:30.450 align:middle line:90% and complexity. 01:01:30.450 --> 01:01:32.370 align:middle line:90% Again, input-output as above. 01:01:32.370 --> 01:01:36.180 align:middle line:84% And communication is extremely fast in computers, 01:01:36.180 --> 01:01:39.480 align:middle line:84% slower but very rich in organisms. 01:01:39.480 --> 01:01:41.820 align:middle line:90% After a very short break-- 01:01:41.820 --> 01:01:44.670 align:middle line:84% we can stretch-- we'll come back and talk 01:01:44.670 --> 01:01:48.630 align:middle line:84% in more detail about how computers actually are made 01:01:48.630 --> 01:01:51.240 align:middle line:84% and how biological systems are made. 01:01:51.240 --> 01:01:53.090 align:middle line:90% Thank you. 01:01:53.090 --> 01:02:36.000 align:middle line:90%