WEBVTT 00:00:00.000 --> 00:00:00.610 align:middle line:90% 00:00:00.610 --> 00:00:02.440 align:middle line:84% The following content is provided 00:00:02.440 --> 00:00:06.340 align:middle line:84% by MIT OpenCourseWare under a Creative Commons license. 00:00:06.340 --> 00:00:10.120 align:middle line:84% Additional information about our license and MIT OpenCourseWare, 00:00:10.120 --> 00:00:12.730 align:middle line:84% in general, is available at OCW.MIT.edu. 00:00:12.730 --> 00:00:22.510 align:middle line:90% 00:00:22.510 --> 00:00:25.870 align:middle line:84% GEORGE CHURCH: OK, so welcome tonight, 00:00:25.870 --> 00:00:30.050 align:middle line:84% a very special lecture for me tonight in more ways 00:00:30.050 --> 00:00:31.540 align:middle line:90% than I can express. 00:00:31.540 --> 00:00:33.760 align:middle line:84% This is the last of my lectures, but I look forward 00:00:33.760 --> 00:00:36.490 align:middle line:90% to your lectures coming up. 00:00:36.490 --> 00:00:37.190 align:middle line:90% I really do. 00:00:37.190 --> 00:00:39.010 align:middle line:84% I've heard from the TFs, and from what 00:00:39.010 --> 00:00:41.150 align:middle line:84% I've seen from talking to many of you, 00:00:41.150 --> 00:00:44.530 align:middle line:84% this is going to be a really amazing set of projects 00:00:44.530 --> 00:00:47.590 align:middle line:84% this year, more than any other year. 00:00:47.590 --> 00:00:55.360 align:middle line:84% Anyway, last week, we asked what can biology do for computing? 00:00:55.360 --> 00:00:59.290 align:middle line:84% And this week is going to be a little less nuts and bolts 00:00:59.290 --> 00:00:59.980 align:middle line:90% than usual. 00:00:59.980 --> 00:01:01.870 align:middle line:84% The entire course-- as you all understand, 00:01:01.870 --> 00:01:04.269 align:middle line:84% you're all self-selected to life survey 00:01:04.269 --> 00:01:08.050 align:middle line:84% courses, at least this one, because you're still here. 00:01:08.050 --> 00:01:11.906 align:middle line:84% And tonight's going to be more of a survey than ever before, 00:01:11.906 --> 00:01:14.860 align:middle line:84% and it's going to be talking about topics for which we don't 00:01:14.860 --> 00:01:19.450 align:middle line:90% have answers for the most part. 00:01:19.450 --> 00:01:23.080 align:middle line:84% There'll be some interesting details along the way. 00:01:23.080 --> 00:01:27.040 align:middle line:84% But anyway, instead of what biology can do for computing, 00:01:27.040 --> 00:01:29.560 align:middle line:84% this is more about what biology and computing can 00:01:29.560 --> 00:01:30.670 align:middle line:90% do for the world. 00:01:30.670 --> 00:01:32.800 align:middle line:84% That's really what we want to do, 00:01:32.800 --> 00:01:35.410 align:middle line:84% and also, the higher level network 00:01:35.410 --> 00:01:39.250 align:middle line:84% models that we can build since this is network number three. 00:01:39.250 --> 00:01:40.710 align:middle line:90% So we'll start out-- 00:01:40.710 --> 00:01:44.440 align:middle line:84% we've been talking mainly about cellular models as our highest 00:01:44.440 --> 00:01:45.940 align:middle line:84% level of networking, and so we'll 00:01:45.940 --> 00:01:49.160 align:middle line:84% talk about multicellular models, in particular, 00:01:49.160 --> 00:01:52.180 align:middle line:84% the sort of multicellular models that you get that involve 00:01:52.180 --> 00:01:59.110 align:middle line:84% sensory integration and integration of higher levels, 00:01:59.110 --> 00:02:03.220 align:middle line:84% and then all the way up to multi-organ systems, where 00:02:03.220 --> 00:02:07.120 align:middle line:84% we'll take an example from the atomic scale, 00:02:07.120 --> 00:02:11.890 align:middle line:84% all the way up to organ system failure. 00:02:11.890 --> 00:02:14.740 align:middle line:90% And then we'll go from there. 00:02:14.740 --> 00:02:17.500 align:middle line:84% So let's start with multicellular models. 00:02:17.500 --> 00:02:21.880 align:middle line:90% And this is something that I-- 00:02:21.880 --> 00:02:24.310 align:middle line:84% I'll show you a few slides that will remind you 00:02:24.310 --> 00:02:25.592 align:middle line:90% of the first lecture. 00:02:25.592 --> 00:02:27.550 align:middle line:84% They either were taken from the first lecture-- 00:02:27.550 --> 00:02:30.980 align:middle line:84% in this case, this is a wildly different author, and approach, 00:02:30.980 --> 00:02:32.962 align:middle line:84% and so forth, but it's the same lesson 00:02:32.962 --> 00:02:34.420 align:middle line:84% as from the first lecture, which is 00:02:34.420 --> 00:02:36.610 align:middle line:84% we not only have these exponential curves going 00:02:36.610 --> 00:02:38.800 align:middle line:84% through the 1900s, but they're super exponential. 00:02:38.800 --> 00:02:40.480 align:middle line:84% They go up, and the trend lines you get 00:02:40.480 --> 00:02:43.690 align:middle line:84% depend on which set of points you use, the steepest one being 00:02:43.690 --> 00:02:46.000 align:middle line:90% the most recent one, 1995. 00:02:46.000 --> 00:02:47.740 align:middle line:84% And this basically is part of when 00:02:47.740 --> 00:02:50.590 align:middle line:84% it is that we will have computing power 00:02:50.590 --> 00:02:56.260 align:middle line:84% that might be on the order of an integrated human neural net. 00:02:56.260 --> 00:02:59.020 align:middle line:84% We talked about neural nets last time as a metaphor, 00:02:59.020 --> 00:03:02.260 align:middle line:84% or as an algorithm, but when will 00:03:02.260 --> 00:03:05.530 align:middle line:84% we actually have a computer that has 00:03:05.530 --> 00:03:10.930 align:middle line:84% the capability of the human cerebral cortex, 00:03:10.930 --> 00:03:13.120 align:middle line:84% or entire integrated nervous system? 00:03:13.120 --> 00:03:15.520 align:middle line:84% And the basis of those calculations 00:03:15.520 --> 00:03:19.480 align:middle line:84% that Moravec made, and in lecture one 00:03:19.480 --> 00:03:22.030 align:middle line:84% Kurzweil was the super exponential 00:03:22.030 --> 00:03:23.380 align:middle line:90% curve that I showed. 00:03:23.380 --> 00:03:26.710 align:middle line:84% Moravec made them based on his studies 00:03:26.710 --> 00:03:29.540 align:middle line:90% on how the retina works. 00:03:29.540 --> 00:03:33.700 align:middle line:84% And I think those of you who have 00:03:33.700 --> 00:03:36.520 align:middle line:84% done a lot of computer programming, 00:03:36.520 --> 00:03:40.750 align:middle line:84% that understanding how the retina works 00:03:40.750 --> 00:03:42.880 align:middle line:84% is probably one of the more intuitive 00:03:42.880 --> 00:03:48.070 align:middle line:84% of our various human senses, in the sense that you can program 00:03:48.070 --> 00:03:51.920 align:middle line:84% an algorithm that might do in the lower right of slide four 00:03:51.920 --> 00:03:54.370 align:middle line:90% here, edge or motion detection. 00:03:54.370 --> 00:03:59.140 align:middle line:84% When a frog sees a fly, black and white, high contrast, 00:03:59.140 --> 00:04:03.340 align:middle line:84% zoom across its visual field, something goes off. 00:04:03.340 --> 00:04:05.830 align:middle line:84% And you could write an algorithm that could do that, 00:04:05.830 --> 00:04:10.130 align:middle line:84% given a few frames of any kind of form. 00:04:10.130 --> 00:04:13.880 align:middle line:84% So anyway, but there's been literature on the subject. 00:04:13.880 --> 00:04:16.990 align:middle line:84% And when you go through even the best algorithms right now 00:04:16.990 --> 00:04:19.420 align:middle line:84% for doing these kind of calculations 00:04:19.420 --> 00:04:24.460 align:middle line:84% and you figure out how long it takes to do-- 00:04:24.460 --> 00:04:29.080 align:middle line:84% the retina is doing about 10 million detections per second. 00:04:29.080 --> 00:04:35.020 align:middle line:84% And then you scale up from that 0.02 gram chunk of retina 00:04:35.020 --> 00:04:39.070 align:middle line:84% up to the 1,500 gram human brain, 00:04:39.070 --> 00:04:42.680 align:middle line:84% then you get the extrapolation that was in the previous slide. 00:04:42.680 --> 00:04:44.963 align:middle line:84% I think this is interesting both as an introduction 00:04:44.963 --> 00:04:47.380 align:middle line:84% to the kind of algorithms we'd like to think about when we 00:04:47.380 --> 00:04:50.950 align:middle line:84% think about integrating a multicellular system, 00:04:50.950 --> 00:04:55.450 align:middle line:84% but also in terms of where the compute power 00:04:55.450 --> 00:04:58.510 align:middle line:84% both biological and computational is going. 00:04:58.510 --> 00:05:02.500 align:middle line:84% Do we want to engineer biological systems 00:05:02.500 --> 00:05:07.270 align:middle line:84% to keep up with these silicon or whatever other computational 00:05:07.270 --> 00:05:13.040 align:middle line:84% system we're using a decade or so from now. 00:05:13.040 --> 00:05:17.290 align:middle line:84% Now, so I said that the visual system 00:05:17.290 --> 00:05:20.230 align:middle line:84% is a relatively intuitive computational system 00:05:20.230 --> 00:05:21.590 align:middle line:90% to program. 00:05:21.590 --> 00:05:25.370 align:middle line:84% In contrast, I would say the olfactory system is less 00:05:25.370 --> 00:05:27.980 align:middle line:90% intuitive, at least for me. 00:05:27.980 --> 00:05:31.340 align:middle line:84% We know now quite a bit about the molecular biology 00:05:31.340 --> 00:05:31.940 align:middle line:90% of the system. 00:05:31.940 --> 00:05:33.410 align:middle line:90% There are 1,000 receptors. 00:05:33.410 --> 00:05:38.450 align:middle line:84% It's probably one of the largest classes of genomic repeated 00:05:38.450 --> 00:05:40.160 align:middle line:90% proteins. 00:05:40.160 --> 00:05:43.100 align:middle line:84% And these receptors are in the same family 00:05:43.100 --> 00:05:45.800 align:middle line:84% as the receptors that are the major class of drug 00:05:45.800 --> 00:05:49.370 align:middle line:84% receptors, the G protein coupled receptors. 00:05:49.370 --> 00:05:52.100 align:middle line:84% And there is basically one receptor 00:05:52.100 --> 00:05:54.080 align:middle line:84% per cell, which is quite a trick, right? 00:05:54.080 --> 00:05:56.960 align:middle line:84% Because you got one receptor from mom and one from dad. 00:05:56.960 --> 00:05:59.240 align:middle line:84% And you've got thousands of receptors. 00:05:59.240 --> 00:06:02.540 align:middle line:84% But there's basically one per cell. 00:06:02.540 --> 00:06:10.340 align:middle line:84% And they detect olfactory molecule concentrations 00:06:10.340 --> 00:06:21.500 align:middle line:84% over about seven logs, where they can detect a concentration 00:06:21.500 --> 00:06:26.240 align:middle line:84% threshold with a particular standard deviation. 00:06:26.240 --> 00:06:28.040 align:middle line:84% Now this is the beginnings of the model, 00:06:28.040 --> 00:06:30.650 align:middle line:84% just stating the fact that you know about it. 00:06:30.650 --> 00:06:35.225 align:middle line:84% When you look at the neuroanatomy, you've got this-- 00:06:35.225 --> 00:06:37.100 align:middle line:84% the odorant molecules down at the very bottom 00:06:37.100 --> 00:06:41.900 align:middle line:84% of this figure, cilia olfactory neurons, which is 00:06:41.900 --> 00:06:44.630 align:middle line:90% a primary signal transduction. 00:06:44.630 --> 00:06:48.740 align:middle line:84% This thing goes up where it's integrated in the glomeruli. 00:06:48.740 --> 00:06:50.980 align:middle line:84% And you can think of this like the neural nets 00:06:50.980 --> 00:06:52.688 align:middle line:84% that we're talking about last time, where 00:06:52.688 --> 00:06:54.200 align:middle line:90% you had these interneurons. 00:06:54.200 --> 00:07:01.280 align:middle line:84% You had the extra layer that we showed was so important. 00:07:01.280 --> 00:07:03.350 align:middle line:84% And the thing that's amazing about this system 00:07:03.350 --> 00:07:09.290 align:middle line:84% are lodged in these four basic olfactory 00:07:09.290 --> 00:07:13.100 align:middle line:84% facts, which I take from Hopfield's work, which 00:07:13.100 --> 00:07:16.730 align:middle line:84% is really just like this is the same Hopfield that introduced 00:07:16.730 --> 00:07:20.540 align:middle line:84% us to or really championed the concept of neural nets 00:07:20.540 --> 00:07:23.900 align:middle line:84% as a computational algorithmic metaphor. 00:07:23.900 --> 00:07:30.410 align:middle line:84% And here, the idea is you have odor and memory recognition. 00:07:30.410 --> 00:07:33.110 align:middle line:84% You have background elimination, just 00:07:33.110 --> 00:07:38.540 align:middle line:84% like how your eyes adapt to a mostly red room, 00:07:38.540 --> 00:07:42.410 align:middle line:84% your nose is great, or whole olfactory system 00:07:42.410 --> 00:07:47.210 align:middle line:84% is great at adapting to background odorants. 00:07:47.210 --> 00:07:51.830 align:middle line:84% So you have one known and one unknown thoroughly mixed. 00:07:51.830 --> 00:07:54.560 align:middle line:84% Number three, you can have component separation. 00:07:54.560 --> 00:07:57.680 align:middle line:84% You can have a few different odors at once. 00:07:57.680 --> 00:08:03.215 align:middle line:84% You have odor separation where you have unknowns. 00:08:03.215 --> 00:08:06.530 align:middle line:90% 00:08:06.530 --> 00:08:10.770 align:middle line:84% I mean, the combination of these is quite remarkable. 00:08:10.770 --> 00:08:18.590 align:middle line:84% So the basis of the model that Hopfield has proposed is this. 00:08:18.590 --> 00:08:22.970 align:middle line:84% You have a coverage for each of the i receptors, 00:08:22.970 --> 00:08:25.070 align:middle line:90% where i goes from 1 to 1,000. 00:08:25.070 --> 00:08:27.080 align:middle line:84% This will vary depending on the organism. 00:08:27.080 --> 00:08:30.390 align:middle line:84% Mice have twice as many as we do, probably, and so on. 00:08:30.390 --> 00:08:36.530 align:middle line:84% But the coverage of each of those receptors 00:08:36.530 --> 00:08:45.050 align:middle line:84% is equal to the minimum coverage for the concentration 00:08:45.050 --> 00:08:47.540 align:middle line:84% of the target, where let's say the target is the target 00:08:47.540 --> 00:08:52.010 align:middle line:84% odorant that you're looking for over the concentration that's 00:08:52.010 --> 00:08:55.880 align:middle line:84% the threshold for firing that neuron for that target, 00:08:55.880 --> 00:08:57.500 align:middle line:90% threshold of t. 00:08:57.500 --> 00:09:01.220 align:middle line:84% And you multiply that times either one or some fraction. 00:09:01.220 --> 00:09:05.150 align:middle line:84% It's 1 if you really hit the target. 00:09:05.150 --> 00:09:09.620 align:middle line:84% And it's this fraction for that particular receptor 00:09:09.620 --> 00:09:11.150 align:middle line:90% and that particular target. 00:09:11.150 --> 00:09:12.620 align:middle line:90% You've got some crosstalk. 00:09:12.620 --> 00:09:15.260 align:middle line:84% And the amount of crosstalk you can 00:09:15.260 --> 00:09:21.030 align:middle line:84% think of this as a field of receptors, which 00:09:21.030 --> 00:09:23.090 align:middle line:90% all have variable binding. 00:09:23.090 --> 00:09:27.210 align:middle line:84% And that binding is over the six or seven logs here. 00:09:27.210 --> 00:09:29.135 align:middle line:84% And this just reflects that same thing 00:09:29.135 --> 00:09:30.260 align:middle line:90% that we were talking about. 00:09:30.260 --> 00:09:32.960 align:middle line:90% 00:09:32.960 --> 00:09:37.220 align:middle line:84% So the total number of inputs going 00:09:37.220 --> 00:09:40.190 align:middle line:84% to the next layer in the neural net, the coverage here, 00:09:40.190 --> 00:09:47.520 align:middle line:84% is the sum of this target signal plus the background signal. 00:09:47.520 --> 00:09:48.770 align:middle line:90% So the first term is target. 00:09:48.770 --> 00:09:50.760 align:middle line:84% And the right hand term is the background. 00:09:50.760 --> 00:09:53.720 align:middle line:84% And so you can see they have the same form, the same threshold, 00:09:53.720 --> 00:09:57.425 align:middle line:84% and concentration. c sub b is concentration of background. 00:09:57.425 --> 00:10:03.630 align:middle line:90% 00:10:03.630 --> 00:10:07.290 align:middle line:84% Now, so let's see how this plays out 00:10:07.290 --> 00:10:13.800 align:middle line:84% when you actually try this out in an olfactory processing 00:10:13.800 --> 00:10:14.460 align:middle line:90% problem. 00:10:14.460 --> 00:10:17.490 align:middle line:90% Here you have an odor space. 00:10:17.490 --> 00:10:23.160 align:middle line:84% And what you have here is 80 different neurons 00:10:23.160 --> 00:10:26.790 align:middle line:84% is the y-axis at the top of slide seven. 00:10:26.790 --> 00:10:33.720 align:middle line:84% And then across the x-axis is time in milliseconds going 00:10:33.720 --> 00:10:35.370 align:middle line:90% from 0 to 800 milliseconds. 00:10:35.370 --> 00:10:41.040 align:middle line:84% And this is all modeled on fairly realistic parameters 00:10:41.040 --> 00:10:43.740 align:middle line:90% based on experiments. 00:10:43.740 --> 00:10:46.590 align:middle line:84% And so what you have here is 80 adapting neurons. 00:10:46.590 --> 00:10:49.260 align:middle line:84% And what you're doing is two sniffs of a mixed odor. 00:10:49.260 --> 00:10:52.920 align:middle line:84% That's what this is modeling, OK? 00:10:52.920 --> 00:10:59.490 align:middle line:84% First, you have a 100 to 500 milliseconds, the range from-- 00:10:59.490 --> 00:11:02.820 align:middle line:84% so you can see that something's happening at 100 milliseconds. 00:11:02.820 --> 00:11:05.160 align:middle line:84% And up to 500 you have a mixed odor, 00:11:05.160 --> 00:11:09.330 align:middle line:84% which is 50 parts of x plus 1,000 parts of y, 00:11:09.330 --> 00:11:14.550 align:middle line:90% or 1,000 times y. 00:11:14.550 --> 00:11:18.180 align:middle line:84% And then 500 milliseconds you change this ratio very 00:11:18.180 --> 00:11:19.140 align:middle line:90% slightly, you see? 00:11:19.140 --> 00:11:27.350 align:middle line:84% You just up x by 75 and y up from 1,000 to 1,100. 00:11:27.350 --> 00:11:33.410 align:middle line:84% OK, so that's the paradigm, 50 and 1,000 to 75 and 1,100. 00:11:33.410 --> 00:11:37.340 align:middle line:84% And the sniff at 100 milliseconds, that first one, 00:11:37.340 --> 00:11:42.690 align:middle line:84% that mixed odorant, it activates more than half the neurons. 00:11:42.690 --> 00:11:45.170 align:middle line:84% So this is not like your retina, where 00:11:45.170 --> 00:11:47.180 align:middle line:84% if you have a single point source 00:11:47.180 --> 00:11:52.070 align:middle line:84% it will activate that one broad cell, or maybe a center 00:11:52.070 --> 00:11:54.380 align:middle line:90% surround effect. 00:11:54.380 --> 00:11:57.830 align:middle line:84% But it's really activating half of the neurons, OK? 00:11:57.830 --> 00:12:00.920 align:middle line:84% And then the changed sniff at 500 milliseconds-- remember, 00:12:00.920 --> 00:12:02.990 align:middle line:90% just these two small changes-- 00:12:02.990 --> 00:12:06.350 align:middle line:90% is almost invisible, right? 00:12:06.350 --> 00:12:11.720 align:middle line:84% So that's what you're seeing right here, 00:12:11.720 --> 00:12:16.400 align:middle line:84% OK, is this huge, swamping amount. 00:12:16.400 --> 00:12:21.950 align:middle line:84% But then in B-- so that's what happened in A. B, 00:12:21.950 --> 00:12:26.660 align:middle line:84% you've now plotting instead of individual neurons, 00:12:26.660 --> 00:12:30.470 align:middle line:84% you're doing the instantaneous rate in hertz, OK? 00:12:30.470 --> 00:12:38.210 align:middle line:84% And now, you can see the second sniff here 00:12:38.210 --> 00:12:41.060 align:middle line:84% at the 500 milliseconds, which even though you really 00:12:41.060 --> 00:12:42.842 align:middle line:84% can't see it in the corresponding point 00:12:42.842 --> 00:12:45.050 align:middle line:84% where you're looking at all these individual neurons. 00:12:45.050 --> 00:12:51.320 align:middle line:84% And so you can see that even a 20% spread in one 00:12:51.320 --> 00:12:56.870 align:middle line:84% of these parameters is enough to get this kind of easily 00:12:56.870 --> 00:12:59.270 align:middle line:90% detectable signal. 00:12:59.270 --> 00:13:01.730 align:middle line:84% Now, that's an example of a simulation. 00:13:01.730 --> 00:13:04.850 align:middle line:84% That's not a simulation of a particular experimental 00:13:04.850 --> 00:13:07.910 align:middle line:84% dataset, but it's based on experimental datasets. 00:13:07.910 --> 00:13:09.830 align:middle line:84% And the code, this is not Mathematica, 00:13:09.830 --> 00:13:12.050 align:middle line:84% this is Matlab, which we all wish 00:13:12.050 --> 00:13:14.930 align:middle line:84% we could have interchangeably taught in this course. 00:13:14.930 --> 00:13:17.210 align:middle line:84% And you can see here in green are the comments. 00:13:17.210 --> 00:13:18.680 align:middle line:84% And you can see some of the things 00:13:18.680 --> 00:13:20.138 align:middle line:84% that we've been talking about here. 00:13:20.138 --> 00:13:23.673 align:middle line:90% 00:13:23.673 --> 00:13:25.340 align:middle line:84% The number of receptor types is actually 00:13:25.340 --> 00:13:27.380 align:middle line:90% here 2,000 instead of 1,000. 00:13:27.380 --> 00:13:29.720 align:middle line:84% You see the target, remember, I said 00:13:29.720 --> 00:13:33.715 align:middle line:84% could be some random number that goes over six logarithms. 00:13:33.715 --> 00:13:35.090 align:middle line:84% Well, the way that's done here is 00:13:35.090 --> 00:13:39.718 align:middle line:84% here's a random number generated and you generate 00:13:39.718 --> 00:13:40.760 align:middle line:90% as the log of the target. 00:13:40.760 --> 00:13:43.052 align:middle line:84% And then, you get the target by taking the exponential. 00:13:43.052 --> 00:13:47.880 align:middle line:84% It's very straightforward and the code goes on from there. 00:13:47.880 --> 00:13:52.760 align:middle line:84% OK, now this brings us to a very interesting point. 00:13:52.760 --> 00:13:55.382 align:middle line:84% Since these three lectures on networks, 00:13:55.382 --> 00:13:57.090 align:middle line:84% and in a certain sense, the whole course, 00:13:57.090 --> 00:14:01.250 align:middle line:84% is building towards systems biology network models. 00:14:01.250 --> 00:14:03.170 align:middle line:84% And for these models to be useful, 00:14:03.170 --> 00:14:05.002 align:middle line:84% in the same sense that the models we have 00:14:05.002 --> 00:14:06.710 align:middle line:84% for X-ray crystallography, and the models 00:14:06.710 --> 00:14:09.320 align:middle line:84% we've had for genome interpretation in terms 00:14:09.320 --> 00:14:12.740 align:middle line:84% of homology, and sequence folding, 00:14:12.740 --> 00:14:16.460 align:middle line:84% and so forth needed to be shared. 00:14:16.460 --> 00:14:18.410 align:middle line:84% The models, even more than the data, 00:14:18.410 --> 00:14:22.640 align:middle line:84% need to be shareable because in a certain sense, 00:14:22.640 --> 00:14:25.580 align:middle line:84% even modest manipulation of the data 00:14:25.580 --> 00:14:27.903 align:middle line:84% usually involves some good model behind it. 00:14:27.903 --> 00:14:29.570 align:middle line:84% And so these are some of the work groups 00:14:29.570 --> 00:14:32.840 align:middle line:84% that are working on different modeling schemes. 00:14:32.840 --> 00:14:35.375 align:middle line:84% In particular, I should point out 00:14:35.375 --> 00:14:38.570 align:middle line:84% that BioSpice is now DARPA BioSpice, 00:14:38.570 --> 00:14:41.570 align:middle line:84% and SBML is kind of a growing part of that. 00:14:41.570 --> 00:14:45.230 align:middle line:84% System Biology Markup Language is a way of sharing this data. 00:14:45.230 --> 00:14:48.830 align:middle line:84% And as yet, there's not the kind of convention 00:14:48.830 --> 00:14:52.070 align:middle line:84% that we have in crystallography and DNA 00:14:52.070 --> 00:14:53.888 align:middle line:90% sequencing where you submit-- 00:14:53.888 --> 00:14:55.430 align:middle line:84% at the time that you publish a paper, 00:14:55.430 --> 00:14:59.840 align:middle line:84% you submit your data and model, data/model, 00:14:59.840 --> 00:15:02.420 align:middle line:84% with an accession number to the database or else 00:15:02.420 --> 00:15:03.560 align:middle line:90% you don't get it published. 00:15:03.560 --> 00:15:05.990 align:middle line:84% We're nowhere near that now for the rest of biology. 00:15:05.990 --> 00:15:08.210 align:middle line:84% But these kind of efforts, probably, 00:15:08.210 --> 00:15:09.480 align:middle line:90% are moving in that direction. 00:15:09.480 --> 00:15:12.595 align:middle line:84% And you can see that the features of each of these 00:15:12.595 --> 00:15:13.970 align:middle line:84% are some of the things that we've 00:15:13.970 --> 00:15:15.053 align:middle line:90% been talking about before. 00:15:15.053 --> 00:15:18.980 align:middle line:84% You've got stochastic modeling, kinetic modeling, 00:15:18.980 --> 00:15:30.350 align:middle line:84% enzyme receptor cell geometry, neural nets, and so on. 00:15:30.350 --> 00:15:32.310 align:middle line:84% This is by no means comprehensive. 00:15:32.310 --> 00:15:34.310 align:middle line:84% And there are hypertext links here if you really 00:15:34.310 --> 00:15:35.690 align:middle line:90% want to dig into them. 00:15:35.690 --> 00:15:37.520 align:middle line:84% And some of the platforms that they've 00:15:37.520 --> 00:15:42.260 align:middle line:84% been thinking about different ways of making it slightly less 00:15:42.260 --> 00:15:43.430 align:middle line:90% system-dependent. 00:15:43.430 --> 00:15:45.530 align:middle line:90% Everyone has their attempts. 00:15:45.530 --> 00:15:50.420 align:middle line:84% Obviously, Windows is not system-independent. 00:15:50.420 --> 00:15:53.420 align:middle line:84% So that's two very brief examples 00:15:53.420 --> 00:15:58.970 align:middle line:84% of multicellular models, both of them being neural models 00:15:58.970 --> 00:16:01.640 align:middle line:84% kind of building from what we were thinking about last time. 00:16:01.640 --> 00:16:03.860 align:middle line:84% But now we're going to a completely different kind 00:16:03.860 --> 00:16:09.050 align:middle line:84% of multicellular model, all the way up from individual effects 00:16:09.050 --> 00:16:12.500 align:middle line:84% that a single nucleotide-- how much effect a single nucleotide 00:16:12.500 --> 00:16:15.980 align:middle line:84% can have on a whole organ system, which in this case 00:16:15.980 --> 00:16:17.940 align:middle line:90% will be cardiovascular. 00:16:17.940 --> 00:16:20.270 align:middle line:90% So we go from--. 00:16:20.270 --> 00:16:26.060 align:middle line:84% And there are a number of physiome and cardiome projects 00:16:26.060 --> 00:16:28.830 align:middle line:90% that are of active interest. 00:16:28.830 --> 00:16:32.440 align:middle line:90% 00:16:32.440 --> 00:16:34.390 align:middle line:84% Just like the previous slides that I 00:16:34.390 --> 00:16:37.330 align:middle line:84% showed showing all the ways of sharing system models, 00:16:37.330 --> 00:16:43.450 align:middle line:84% many of these efforts are kind of loose consortium alliances 00:16:43.450 --> 00:16:47.890 align:middle line:84% where people have put together different models either to deal 00:16:47.890 --> 00:16:52.870 align:middle line:84% with the anatomy, physiology, or sometimes, 00:16:52.870 --> 00:17:00.580 align:middle line:84% a integration of molecular to cellular, cellular to, say, 00:17:00.580 --> 00:17:05.290 align:middle line:84% neural or muscular, and all the way up to fluid dynamics 00:17:05.290 --> 00:17:08.359 align:middle line:90% and so on. 00:17:08.359 --> 00:17:12.310 align:middle line:84% So let's start with a single base-pair change. 00:17:12.310 --> 00:17:14.993 align:middle line:84% And I start with this one because you 00:17:14.993 --> 00:17:16.910 align:middle line:84% should feel comfortable with it at this point. 00:17:16.910 --> 00:17:19.960 align:middle line:90% We mentioned it briefly before. 00:17:19.960 --> 00:17:22.000 align:middle line:84% We'll talk about it in more detail now. 00:17:22.000 --> 00:17:24.589 align:middle line:84% This is the single nucleotide polymorphism, 00:17:24.589 --> 00:17:30.340 align:middle line:84% which causes the change of the beta subunit of hemoglobin, 00:17:30.340 --> 00:17:37.030 align:middle line:84% which is a tetramer, to go from a glutamate, 00:17:37.030 --> 00:17:40.450 align:middle line:84% normally, in most people at position six of the beta chain, 00:17:40.450 --> 00:17:42.970 align:middle line:90% to a valine. 00:17:42.970 --> 00:17:46.180 align:middle line:84% So that's the purple set of tetramers. 00:17:46.180 --> 00:17:51.400 align:middle line:84% Or to tryptophan, and that's the cyan set on the far right. 00:17:51.400 --> 00:17:59.670 align:middle line:84% And this, from combinations of X-ray crystallography 00:17:59.670 --> 00:18:01.860 align:middle line:84% and three-dimensional modeling, and this 00:18:01.860 --> 00:18:04.950 align:middle line:84% is not speculation, just based on primary sequence data, 00:18:04.950 --> 00:18:08.070 align:middle line:84% even though the primary sequence here is very, very close, 00:18:08.070 --> 00:18:12.400 align:middle line:84% just a single nucleotide and a single amino acid change. 00:18:12.400 --> 00:18:14.610 align:middle line:84% Nevertheless, these authors have been 00:18:14.610 --> 00:18:18.840 align:middle line:84% building this from the crystallographic data 00:18:18.840 --> 00:18:22.200 align:middle line:84% where not only do you care about the tetramer itself, 00:18:22.200 --> 00:18:24.810 align:middle line:84% but how the tetramers interact with one another 00:18:24.810 --> 00:18:28.500 align:middle line:84% to make these long, fibrous chains, which are considerably 00:18:28.500 --> 00:18:31.320 align:middle line:90% more stable in the sickle cell. 00:18:31.320 --> 00:18:39.180 align:middle line:84% And what you see here is sort of the valine substitution 00:18:39.180 --> 00:18:41.760 align:middle line:84% is locked in one kind of conformation 00:18:41.760 --> 00:18:44.290 align:middle line:90% and the tryptophan in another. 00:18:44.290 --> 00:18:47.100 align:middle line:84% And you can see how these authors 00:18:47.100 --> 00:18:51.120 align:middle line:84% have decided how a potential fiber can 00:18:51.120 --> 00:18:53.280 align:middle line:90% form in the different cases. 00:18:53.280 --> 00:18:57.110 align:middle line:90% 00:18:57.110 --> 00:18:59.870 align:middle line:84% Now those fibers, in one way or another, 00:18:59.870 --> 00:19:05.390 align:middle line:84% which is not totally mapped out, but you 00:19:05.390 --> 00:19:10.740 align:middle line:84% affect the efficiency of the hemoglobin slightly 00:19:10.740 --> 00:19:13.940 align:middle line:84% and the shape of the cell much more radically. 00:19:13.940 --> 00:19:19.200 align:middle line:84% And this combination results in the cell, 00:19:19.200 --> 00:19:23.090 align:middle line:84% especially under any kind of stress, either oxidative 00:19:23.090 --> 00:19:25.910 align:middle line:84% or other metabolic stress, becoming sickled. 00:19:25.910 --> 00:19:29.330 align:middle line:84% So you can have a combination of cells which have 00:19:29.330 --> 00:19:31.010 align:middle line:90% varying degrees of sickling. 00:19:31.010 --> 00:19:37.430 align:middle line:84% Here in this microscopic slide on the left, 00:19:37.430 --> 00:19:40.130 align:middle line:84% you have both the sickle cell and the normal cell side 00:19:40.130 --> 00:19:42.210 align:middle line:90% by side. 00:19:42.210 --> 00:19:45.260 align:middle line:84% I think when we built up the red blood cell metabolic model, 00:19:45.260 --> 00:19:48.050 align:middle line:84% we talked about some of the non-metabolic considerations 00:19:48.050 --> 00:19:52.280 align:middle line:84% of that, which were that its function is to transport oxygen 00:19:52.280 --> 00:19:54.680 align:middle line:90% and so forth. 00:19:54.680 --> 00:19:58.330 align:middle line:84% And what we're talking about now is more a cell membrane issue. 00:19:58.330 --> 00:20:03.680 align:middle line:84% This is on slide 13, where the internal environment, which 00:20:03.680 --> 00:20:06.050 align:middle line:84% is the hemoglobin, is greatly affecting now 00:20:06.050 --> 00:20:09.170 align:middle line:84% the external environment, which is the hemodynamic flow 00:20:09.170 --> 00:20:12.320 align:middle line:90% in the capillaries. 00:20:12.320 --> 00:20:15.680 align:middle line:90% 00:20:15.680 --> 00:20:17.840 align:middle line:84% So this falls under the heading now 00:20:17.840 --> 00:20:21.560 align:middle line:84% of how is it that we can go from that single nucleotide change 00:20:21.560 --> 00:20:23.720 align:middle line:84% to a very dramatic morphological change? 00:20:23.720 --> 00:20:25.970 align:middle line:84% In this case, it's pathological, but you 00:20:25.970 --> 00:20:30.887 align:middle line:84% could imagine that in the hands of evolutionary adaptation, 00:20:30.887 --> 00:20:32.720 align:middle line:84% an organism could take this and run with it. 00:20:32.720 --> 00:20:35.180 align:middle line:84% Maybe not sickle cell mutation, but some other one that 00:20:35.180 --> 00:20:38.060 align:middle line:84% causes some other morphological change in some other cell 00:20:38.060 --> 00:20:41.120 align:middle line:90% or complex aggregate of cells. 00:20:41.120 --> 00:20:47.750 align:middle line:84% Because here you can see a whole variety of different shapes 00:20:47.750 --> 00:20:49.793 align:middle line:90% of red blood cells. 00:20:49.793 --> 00:20:51.960 align:middle line:84% And remember, these red blood cells are very simple. 00:20:51.960 --> 00:20:56.900 align:middle line:84% They have no macromolecular [INAUDIBLE],, no DNA, no RNA, 00:20:56.900 --> 00:20:59.540 align:middle line:84% so we're really just talking about a bag of proteins which 00:20:59.540 --> 00:21:03.590 align:middle line:84% are greatly affected by these different conditions 00:21:03.590 --> 00:21:05.460 align:middle line:90% or enzyme deficiencies. 00:21:05.460 --> 00:21:07.280 align:middle line:84% So the system models that were built up, 00:21:07.280 --> 00:21:11.450 align:middle line:84% where we had the kinetic parameters for the enzymes, 00:21:11.450 --> 00:21:16.100 align:middle line:84% here can, in principle, be modeled on the impact it would 00:21:16.100 --> 00:21:21.320 align:middle line:84% have on the osmolarity, or the other membrane properties that 00:21:21.320 --> 00:21:22.910 align:middle line:84% were listed in the previous slide, 00:21:22.910 --> 00:21:25.250 align:middle line:84% or the sickling that was in two slides back. 00:21:25.250 --> 00:21:36.680 align:middle line:90% 00:21:36.680 --> 00:21:41.810 align:middle line:84% So from a single nucleotide polymorphism, 00:21:41.810 --> 00:21:44.240 align:middle line:84% we've gone from this three-dimensional fiber 00:21:44.240 --> 00:21:50.000 align:middle line:84% of hemoglobin to a change in the three-dimensional structure 00:21:50.000 --> 00:21:57.020 align:middle line:84% of this biconcave disk to a sickle disk. 00:21:57.020 --> 00:21:59.540 align:middle line:84% But you can also go, sort of in parallel, 00:21:59.540 --> 00:22:02.960 align:middle line:84% the same single-nucleotide polymorphism, or some like it, 00:22:02.960 --> 00:22:05.750 align:middle line:84% can take you up to pathogen resistance. 00:22:05.750 --> 00:22:09.500 align:middle line:84% We've already mentioned that sickle cell hemoglobin can take 00:22:09.500 --> 00:22:11.690 align:middle line:90% you to malarial resistance. 00:22:11.690 --> 00:22:19.010 align:middle line:84% But in addition, you can get components 00:22:19.010 --> 00:22:22.860 align:middle line:84% of the enzymatic metabolic pathway, 00:22:22.860 --> 00:22:28.940 align:middle line:84% such as the one that we modeled a few lectures back, 00:22:28.940 --> 00:22:30.630 align:middle line:90% such as glutathione peroxidase. 00:22:30.630 --> 00:22:35.150 align:middle line:84% This is part of the redox components. 00:22:35.150 --> 00:22:39.110 align:middle line:84% And here, erythrocytes that are heterozygous 00:22:39.110 --> 00:22:46.230 align:middle line:84% for this particular allele should 00:22:46.230 --> 00:22:48.390 align:middle line:84% be more efficient in sheltering the cell membrane 00:22:48.390 --> 00:22:51.090 align:middle line:84% from irreversible oxidation in binding hemoglobin caused 00:22:51.090 --> 00:22:53.130 align:middle line:84% by the oxidant stress that's exerted 00:22:53.130 --> 00:22:57.660 align:middle line:90% by the malarial parasite. 00:22:57.660 --> 00:22:59.910 align:middle line:84% And what they observe here is actually 00:22:59.910 --> 00:23:07.470 align:middle line:84% an interaction between these two possible haplotypes. 00:23:07.470 --> 00:23:10.080 align:middle line:84% So it's actually, you can think of it as a pair of haplotypes. 00:23:10.080 --> 00:23:14.880 align:middle line:84% It's the one is the hemoglobin AS, 00:23:14.880 --> 00:23:18.000 align:middle line:84% the one we've been talking about for sickling, 00:23:18.000 --> 00:23:20.880 align:middle line:84% and then, this glutathione peroxidase. 00:23:20.880 --> 00:23:23.640 align:middle line:84% And so you can imagine, that since both of these things 00:23:23.640 --> 00:23:27.030 align:middle line:84% interact with malaria parasite, that your genotype 00:23:27.030 --> 00:23:27.900 align:middle line:90% will depend-- 00:23:27.900 --> 00:23:30.060 align:middle line:84% I'm sorry, your phenotype in respect to malaria 00:23:30.060 --> 00:23:33.577 align:middle line:84% will depend on the alleles that you have at both cases. 00:23:33.577 --> 00:23:35.910 align:middle line:84% So you can-- here, you can have an A over S heterozygote 00:23:35.910 --> 00:23:39.247 align:middle line:90% and a two over one heterozygote. 00:23:39.247 --> 00:23:41.580 align:middle line:84% And so that's something to take into account when you're 00:23:41.580 --> 00:23:46.170 align:middle line:84% trying to do any kind of predictive modeling or modeling 00:23:46.170 --> 00:23:49.200 align:middle line:84% to explain the functional genomics that you 00:23:49.200 --> 00:23:51.840 align:middle line:84% have in a patient who's a compound heterozygote 00:23:51.840 --> 00:23:53.380 align:middle line:90% like that. 00:23:53.380 --> 00:23:55.590 align:middle line:90% So now, the third pathway. 00:23:55.590 --> 00:23:57.960 align:middle line:84% Now, from single-- the same set of single nucleotide 00:23:57.960 --> 00:23:59.790 align:middle line:84% polymorphisms that affect either hemoglobin 00:23:59.790 --> 00:24:02.800 align:middle line:84% or one of the major enzymes in the red blood cell. 00:24:02.800 --> 00:24:06.300 align:middle line:90% Now, on to cell morphology. 00:24:06.300 --> 00:24:09.690 align:middle line:90% On to pathogen interaction. 00:24:09.690 --> 00:24:11.670 align:middle line:84% And finally, to interaction with drugs. 00:24:11.670 --> 00:24:13.290 align:middle line:84% We talked about pharmacogenomics. 00:24:13.290 --> 00:24:19.650 align:middle line:84% Here's another example where you have the drug-induced oxidative 00:24:19.650 --> 00:24:24.330 align:middle line:84% hemolysis that you occur with certain enzymopathies 00:24:24.330 --> 00:24:28.470 align:middle line:84% like glucose-6-phosphate dehydrogenase. 00:24:28.470 --> 00:24:32.550 align:middle line:84% This enzyme, glucose-6-phosphate dehydrogenase, 00:24:32.550 --> 00:24:36.660 align:middle line:84% interacts with drugs such as primaquine. 00:24:36.660 --> 00:24:42.700 align:middle line:84% And it disrupts mitochondrial function, heme biosynthesis, 00:24:42.700 --> 00:24:45.700 align:middle line:90% and so forth, and so on. 00:24:45.700 --> 00:24:51.720 align:middle line:84% This is a very significant consideration 00:24:51.720 --> 00:24:54.420 align:middle line:84% with a long list of 20-some drugs 00:24:54.420 --> 00:24:56.460 align:middle line:84% that has to be taken to account when 00:24:56.460 --> 00:25:03.490 align:middle line:84% you have any of a variety of red blood cell enzyme changes. 00:25:03.490 --> 00:25:08.160 align:middle line:84% So we've got these three effects of a limited number 00:25:08.160 --> 00:25:11.310 align:middle line:84% of single nucleotide polymorphisms. 00:25:11.310 --> 00:25:16.860 align:middle line:84% How do you make this transition from-- 00:25:16.860 --> 00:25:18.570 align:middle line:84% and they're all somewhat interconnected. 00:25:18.570 --> 00:25:20.403 align:middle line:84% Part of the reason it's resistant to malaria 00:25:20.403 --> 00:25:22.350 align:middle line:84% is because it's less effective in its cell 00:25:22.350 --> 00:25:27.750 align:middle line:84% shape, and its ability to transport oxygen, and do 00:25:27.750 --> 00:25:29.450 align:middle line:90% its metabolism. 00:25:29.450 --> 00:25:32.460 align:middle line:84% And same thing with drug sensitivity. 00:25:32.460 --> 00:25:39.750 align:middle line:84% How are these changes in the hemoglobin-- 00:25:39.750 --> 00:25:42.690 align:middle line:84% might they be reflected in the three-dimensional shape 00:25:42.690 --> 00:25:45.360 align:middle line:84% of the erythrocyte, which is a kind 00:25:45.360 --> 00:25:47.940 align:middle line:90% of a membrane-bound compartment? 00:25:47.940 --> 00:25:50.520 align:middle line:84% And here's a model that struck me as interesting. 00:25:50.520 --> 00:25:56.140 align:middle line:84% It's not been that extensively tested since 1998. 00:25:56.140 --> 00:25:58.830 align:middle line:90% But the idea is that Band3-- 00:25:58.830 --> 00:26:01.260 align:middle line:84% this is one of those names that just comes out 00:26:01.260 --> 00:26:04.320 align:middle line:84% of the molecular biologist literature, 00:26:04.320 --> 00:26:06.330 align:middle line:84% that they ran a gel, and they count 00:26:06.330 --> 00:26:08.100 align:middle line:90% the bands, and number three. 00:26:08.100 --> 00:26:13.380 align:middle line:84% And it turns out it's about 10% of the red blood cell membrane. 00:26:13.380 --> 00:26:16.650 align:middle line:90% It's a very abundant protein. 00:26:16.650 --> 00:26:20.670 align:middle line:84% And it's responsible for the equilibration 00:26:20.670 --> 00:26:24.480 align:middle line:84% of anions such as carbonate and chloride across the red blood 00:26:24.480 --> 00:26:25.740 align:middle line:90% cell membrane. 00:26:25.740 --> 00:26:29.160 align:middle line:84% It's not a pump, it just kind of allows these anions 00:26:29.160 --> 00:26:31.200 align:middle line:84% to go across because the pump here 00:26:31.200 --> 00:26:34.170 align:middle line:84% is the pumps that are involved in proton 00:26:34.170 --> 00:26:37.740 align:middle line:84% and there's also sodium potassium. 00:26:37.740 --> 00:26:40.320 align:middle line:84% And that's what your ATP in the red blood cell is going for. 00:26:40.320 --> 00:26:42.112 align:middle line:84% And these are just kind of following along. 00:26:42.112 --> 00:26:44.460 align:middle line:90% 00:26:44.460 --> 00:26:54.190 align:middle line:84% The idea here is that if you change the degree-- 00:26:54.190 --> 00:26:57.438 align:middle line:84% let's see, I think this is covered in more detail. 00:26:57.438 --> 00:26:58.770 align:middle line:90% No, sorry. 00:26:58.770 --> 00:27:02.405 align:middle line:90% 00:27:02.405 --> 00:27:05.670 align:middle line:84% The mechanism of action here is greatly 00:27:05.670 --> 00:27:08.850 align:middle line:84% affected by its disulfide effects. 00:27:08.850 --> 00:27:13.590 align:middle line:84% And if you change the redox components in the red blood 00:27:13.590 --> 00:27:17.400 align:middle line:84% cell, you get just a slight change in conformation 00:27:17.400 --> 00:27:18.840 align:middle line:90% of this molecule. 00:27:18.840 --> 00:27:22.140 align:middle line:84% And then that can result in a net difference 00:27:22.140 --> 00:27:26.040 align:middle line:84% between the cross-sectional area of this protein on the outside 00:27:26.040 --> 00:27:29.820 align:middle line:84% and the inside, which translates into a net change-- 00:27:29.820 --> 00:27:34.277 align:middle line:84% since the phospholipids don't exchange, 00:27:34.277 --> 00:27:36.360 align:middle line:84% or the rate constant for the phospholipid exchange 00:27:36.360 --> 00:27:38.760 align:middle line:84% is slow and known, then this results 00:27:38.760 --> 00:27:45.900 align:middle line:90% in a conformational change. 00:27:45.900 --> 00:27:48.840 align:middle line:90% Anyway, that's the model. 00:27:48.840 --> 00:27:51.900 align:middle line:84% That connects the single nucleotide to-- 00:27:51.900 --> 00:27:54.900 align:middle line:84% could connect it to the three-dimensional structure 00:27:54.900 --> 00:27:55.770 align:middle line:90% of the cell. 00:27:55.770 --> 00:27:58.103 align:middle line:84% Then you want to connect the three-dimensional structure 00:27:58.103 --> 00:28:01.560 align:middle line:84% of the cell to its ability to carry out its function 00:28:01.560 --> 00:28:02.820 align:middle line:90% in the capillaries. 00:28:02.820 --> 00:28:06.390 align:middle line:84% And its function there is to allow the fusion 00:28:06.390 --> 00:28:12.570 align:middle line:90% of oxygen and carbon dioxide. 00:28:12.570 --> 00:28:22.520 align:middle line:90% 00:28:22.520 --> 00:28:28.550 align:middle line:84% So here, each of these cells is shaped by a mechanical process. 00:28:28.550 --> 00:28:30.350 align:middle line:84% A mechanical process here, you can 00:28:30.350 --> 00:28:36.020 align:middle line:84% take the known or the measurable mechanical properties 00:28:36.020 --> 00:28:39.230 align:middle line:84% of a red blood cell and subject them 00:28:39.230 --> 00:28:45.170 align:middle line:84% to a method called finite elements analysis where 00:28:45.170 --> 00:28:49.790 align:middle line:84% you're solving these partial differentiable equations. 00:28:49.790 --> 00:28:55.500 align:middle line:84% And you can calculate the exchange of the oxygen-- 00:28:55.500 --> 00:28:59.270 align:middle line:90% 00:28:59.270 --> 00:29:01.725 align:middle line:84% alveolar just refers to-- this would be in the lungs. 00:29:01.725 --> 00:29:03.600 align:middle line:84% This would be a small capillary in the lungs. 00:29:03.600 --> 00:29:05.017 align:middle line:84% And you can see the capillary now, 00:29:05.017 --> 00:29:07.670 align:middle line:84% the endothelial walls are close enough so that the red blood 00:29:07.670 --> 00:29:10.828 align:middle line:84% cells are basically deformed in their shape 00:29:10.828 --> 00:29:11.870 align:middle line:90% as they go through there. 00:29:11.870 --> 00:29:14.690 align:middle line:84% And the exact shapes that are compatible with this 00:29:14.690 --> 00:29:17.240 align:middle line:84% will determine the rate that these little arrows that 00:29:17.240 --> 00:29:21.470 align:middle line:84% go from the oxygen on the outside the lung epithelium 00:29:21.470 --> 00:29:24.360 align:middle line:90% get through to the blood cell. 00:29:24.360 --> 00:29:31.130 align:middle line:84% So that's a reference that deals with that kind of problem. 00:29:31.130 --> 00:29:39.580 align:middle line:84% Now, as we start thinking about building up a larger system 00:29:39.580 --> 00:29:42.100 align:middle line:84% model, the other parts of the system, each of them-- 00:29:42.100 --> 00:29:43.600 align:middle line:84% so we've got a red blood cell model, 00:29:43.600 --> 00:29:44.642 align:middle line:90% it's the metabolic model. 00:29:44.642 --> 00:29:49.000 align:middle line:84% In principle, it could also be a shape model and a diffusion 00:29:49.000 --> 00:29:50.260 align:middle line:90% model. 00:29:50.260 --> 00:29:53.360 align:middle line:84% But the other parts of the system 00:29:53.360 --> 00:29:58.010 align:middle line:90% typically involve muscle cells. 00:29:58.010 --> 00:30:07.830 align:middle line:84% They have the smooth muscles throughout the entire arterial 00:30:07.830 --> 00:30:12.820 align:middle line:84% and venous system, and most significantly, the heart 00:30:12.820 --> 00:30:13.750 align:middle line:90% muscles. 00:30:13.750 --> 00:30:19.870 align:middle line:84% And so here you have an example of action potential in one 00:30:19.870 --> 00:30:23.380 align:middle line:84% of the ventricular cells of the dog heart. 00:30:23.380 --> 00:30:27.730 align:middle line:84% And it includes all the components that at least we, 00:30:27.730 --> 00:30:34.460 align:middle line:84% since this is not an entire course on ventricular heart 00:30:34.460 --> 00:30:36.890 align:middle line:90% modeling. 00:30:36.890 --> 00:30:39.230 align:middle line:84% This is the sort of parameters, in terms 00:30:39.230 --> 00:30:42.620 align:middle line:84% of all the I to the currents for each of the different potassium 00:30:42.620 --> 00:30:45.980 align:middle line:84% channels to K, all of-- each of the ones at I 00:30:45.980 --> 00:30:48.680 align:middle line:90% sub K was all potassium. 00:30:48.680 --> 00:30:52.550 align:middle line:84% And then there's some in sodium, and sodium and calcium, calcium 00:30:52.550 --> 00:30:54.607 align:middle line:90% potassium, sodium alone. 00:30:54.607 --> 00:30:56.690 align:middle line:84% And so each of these things, that internal storage 00:30:56.690 --> 00:30:59.900 align:middle line:84% of calcium [INAUDIBLE] from one subcellular compartment 00:30:59.900 --> 00:31:03.210 align:middle line:90% to another and so on. 00:31:03.210 --> 00:31:03.920 align:middle line:90% You get the idea. 00:31:03.920 --> 00:31:07.130 align:middle line:84% Each of these things has to have the parameters measured. 00:31:07.130 --> 00:31:10.160 align:middle line:84% And any of that are absent, you have 00:31:10.160 --> 00:31:16.670 align:middle line:84% to have reasonable ways of getting surrogates. 00:31:16.670 --> 00:31:20.330 align:middle line:84% And so this is just the sort of data 00:31:20.330 --> 00:31:24.470 align:middle line:84% that would go into the other major type of cell that 00:31:24.470 --> 00:31:29.390 align:middle line:84% comes into this cardiovascular modeling. 00:31:29.390 --> 00:31:33.260 align:middle line:84% And finally, you can integrate this up 00:31:33.260 --> 00:31:36.020 align:middle line:90% to a fairly complete system. 00:31:36.020 --> 00:31:37.580 align:middle line:84% It's obviously not complete until you 00:31:37.580 --> 00:31:40.460 align:middle line:84% get to the whole organism, but higher level, 00:31:40.460 --> 00:31:44.260 align:middle line:84% and how we're talking about whole body recirculation. 00:31:44.260 --> 00:31:47.540 align:middle line:84% This NSR group, this is a hypertext link 00:31:47.540 --> 00:31:50.930 align:middle line:84% for this particular model which you can download. 00:31:50.930 --> 00:31:52.500 align:middle line:84% And it has a four-chambered heart. 00:31:52.500 --> 00:31:54.620 align:middle line:84% We were just talking about one ventricular 00:31:54.620 --> 00:31:56.300 align:middle line:90% cell on the previous slide. 00:31:56.300 --> 00:31:58.100 align:middle line:90% But that would be-- 00:31:58.100 --> 00:32:00.980 align:middle line:84% if we look at the second box from the top here, 00:32:00.980 --> 00:32:05.130 align:middle line:84% this is where the heart and lung model would fit in there. 00:32:05.130 --> 00:32:08.300 align:middle line:84% It's got seven different organs, which include kidney, liver, 00:32:08.300 --> 00:32:11.140 align:middle line:84% lower limbs, and so forth down at the bottom part 00:32:11.140 --> 00:32:12.990 align:middle line:90% of this diagram. 00:32:12.990 --> 00:32:14.180 align:middle line:90% And you get the picture. 00:32:14.180 --> 00:32:23.690 align:middle line:84% Each of these things, you're modeling the volumes and flows, 00:32:23.690 --> 00:32:25.790 align:middle line:90% Now, this is systems biology. 00:32:25.790 --> 00:32:29.540 align:middle line:84% This is really kind of a renaissance of interest 00:32:29.540 --> 00:32:32.330 align:middle line:84% in physiology, which is what it would have been called before. 00:32:32.330 --> 00:32:36.800 align:middle line:84% And there are some really interesting mysteries out there 00:32:36.800 --> 00:32:41.570 align:middle line:84% to be solved that are really only solvable, or possibly only 00:32:41.570 --> 00:32:42.920 align:middle line:90% solvable, at the system level. 00:32:42.920 --> 00:32:45.140 align:middle line:84% You never know, somebody could come up 00:32:45.140 --> 00:32:46.730 align:middle line:84% and say, oh, yeah, this is really 00:32:46.730 --> 00:32:50.340 align:middle line:84% well understood immunology and that answers everything. 00:32:50.340 --> 00:32:55.307 align:middle line:84% But even if it's some kind of immunology, 00:32:55.307 --> 00:32:57.140 align:middle line:84% you still have to say, how does it play out? 00:32:57.140 --> 00:32:59.473 align:middle line:84% And what happens is we've been talking about sickle cell 00:32:59.473 --> 00:33:02.960 align:middle line:84% disease, and you can have fairly mild pain 00:33:02.960 --> 00:33:05.130 align:middle line:90% as one of the major symptoms. 00:33:05.130 --> 00:33:06.980 align:middle line:84% But then all of a sudden, out of nowhere, 00:33:06.980 --> 00:33:09.470 align:middle line:84% you'll get this multi-organ system failure 00:33:09.470 --> 00:33:13.850 align:middle line:84% where almost every major organ in that previous slide, 00:33:13.850 --> 00:33:15.590 align:middle line:84% lungs and so forth, fill up with fluid 00:33:15.590 --> 00:33:19.082 align:middle line:84% and you have a very high chance of dying. 00:33:19.082 --> 00:33:21.290 align:middle line:84% And this could happen to every one of us in this room 00:33:21.290 --> 00:33:23.840 align:middle line:84% because it's not restricted to sickle cell. 00:33:23.840 --> 00:33:30.980 align:middle line:84% If any of you have severe burn, or car crash, major bone 00:33:30.980 --> 00:33:33.590 align:middle line:84% injuries, and so forth, you have a very good chance 00:33:33.590 --> 00:33:36.030 align:middle line:84% of getting into multi-organ system failure. 00:33:36.030 --> 00:33:39.500 align:middle line:84% And it really isn't known how that plays out. 00:33:39.500 --> 00:33:42.290 align:middle line:84% I mean, not even that-- not just a quantitative model, 00:33:42.290 --> 00:33:45.020 align:middle line:90% but even qualitatively. 00:33:45.020 --> 00:33:49.220 align:middle line:84% So I think this is something-- there are projects now 00:33:49.220 --> 00:33:54.620 align:middle line:84% to get genomic data collection, both genetic and expression 00:33:54.620 --> 00:34:00.470 align:middle line:84% data, and time will tell whether that is actually the best 00:34:00.470 --> 00:34:02.770 align:middle line:90% route to solving that mystery. 00:34:02.770 --> 00:34:04.340 align:middle line:90% So that was multi-organ. 00:34:04.340 --> 00:34:07.430 align:middle line:84% Now we're talking about multi-organism. 00:34:07.430 --> 00:34:11.690 align:middle line:84% After you've built up an entire organism, 00:34:11.690 --> 00:34:13.670 align:middle line:84% then you want to know how it acts 00:34:13.670 --> 00:34:18.560 align:middle line:84% at the next level of network analysis, which is how does it 00:34:18.560 --> 00:34:22.489 align:middle line:90% fit in with other organisms? 00:34:22.489 --> 00:34:26.330 align:middle line:84% None of us, almost no organism, really 00:34:26.330 --> 00:34:29.690 align:middle line:84% belongs in ecological niche all by itself. 00:34:29.690 --> 00:34:35.659 align:middle line:84% And some of the modeling we are taught in this course 00:34:35.659 --> 00:34:37.730 align:middle line:90% could be called simulation. 00:34:37.730 --> 00:34:43.820 align:middle line:84% And probably one of the longest-lasting computer 00:34:43.820 --> 00:34:48.440 align:middle line:84% games in the world, and one of the most successful, 00:34:48.440 --> 00:34:52.550 align:middle line:90% I've heard, is The Sims. 00:34:52.550 --> 00:34:54.530 align:middle line:90% It started with SimCity in-- 00:34:54.530 --> 00:34:57.440 align:middle line:90% 00:34:57.440 --> 00:34:58.520 align:middle line:90% SimCity in '87. 00:34:58.520 --> 00:35:00.995 align:middle line:90% 00:35:00.995 --> 00:35:02.870 align:middle line:84% And then this particular one illustrated here 00:35:02.870 --> 00:35:06.470 align:middle line:84% is SimLife, which is all about ecological modeling. 00:35:06.470 --> 00:35:08.870 align:middle line:84% And it's not entirely different from what 00:35:08.870 --> 00:35:12.530 align:middle line:84% a serious ecological-modeling program would do. 00:35:12.530 --> 00:35:14.870 align:middle line:84% It certainly has a lot of the interesting parameters 00:35:14.870 --> 00:35:18.410 align:middle line:90% such as the lifespan here. 00:35:18.410 --> 00:35:21.890 align:middle line:84% You're doing demographics here so you have the lifespan, 00:35:21.890 --> 00:35:28.850 align:middle line:84% the amount of food needed, the size, kind of vision, 00:35:28.850 --> 00:35:32.300 align:middle line:84% roaming, so forth, of all these different kinds of animals. 00:35:32.300 --> 00:35:35.090 align:middle line:90% 00:35:35.090 --> 00:35:38.960 align:middle line:84% You can have plants, you have herbivores, carnivores, 00:35:38.960 --> 00:35:39.930 align:middle line:90% and so on. 00:35:39.930 --> 00:35:42.560 align:middle line:84% And then you can set up the population sizes, 00:35:42.560 --> 00:35:45.980 align:middle line:84% and then it will do simulations based 00:35:45.980 --> 00:35:51.440 align:middle line:84% on that, where each individual animal is tracked in terms 00:35:51.440 --> 00:35:53.570 align:middle line:90% of its position and quantity. 00:35:53.570 --> 00:35:59.360 align:middle line:84% And as you would expect, as the carnivores build up, 00:35:59.360 --> 00:36:00.860 align:middle line:84% then they knock down the herbivores, 00:36:00.860 --> 00:36:04.160 align:middle line:84% and the plants go up, and so forth. 00:36:04.160 --> 00:36:06.680 align:middle line:84% And it goes through these cycles. 00:36:06.680 --> 00:36:08.990 align:middle line:84% And so this is basically a stochastic model, 00:36:08.990 --> 00:36:13.730 align:middle line:84% just like the stochastic models that we had for molecules, 00:36:13.730 --> 00:36:16.010 align:middle line:90% but here, at the organism level. 00:36:16.010 --> 00:36:19.160 align:middle line:84% And what's happening here in the slower right-hand section 00:36:19.160 --> 00:36:22.580 align:middle line:84% is you've got little green plants getting 00:36:22.580 --> 00:36:25.160 align:middle line:84% eaten by herbivores, and the herbivores getting 00:36:25.160 --> 00:36:26.360 align:middle line:90% eaten by carnivores. 00:36:26.360 --> 00:36:29.050 align:middle line:90% 00:36:29.050 --> 00:36:35.210 align:middle line:84% Now, hopefully, this course has already 00:36:35.210 --> 00:36:40.100 align:middle line:84% or will inspire you to really think globally. 00:36:40.100 --> 00:36:42.500 align:middle line:84% To think not only globally in terms 00:36:42.500 --> 00:36:48.020 align:middle line:84% of how little pieces of molecular tools 00:36:48.020 --> 00:36:50.750 align:middle line:84% fit together into systems and networks, 00:36:50.750 --> 00:36:53.600 align:middle line:84% but how our systems and networks we model 00:36:53.600 --> 00:36:58.430 align:middle line:84% fit into the big picture of what are really important problems. 00:36:58.430 --> 00:37:04.550 align:middle line:84% Maybe a slightly improved Viagra is not 00:37:04.550 --> 00:37:08.330 align:middle line:84% in the same category as a new tuberculosis drug 00:37:08.330 --> 00:37:09.080 align:middle line:90% or maybe it is. 00:37:09.080 --> 00:37:13.130 align:middle line:84% I mean, you decide for yourself what is thinking globally. 00:37:13.130 --> 00:37:14.870 align:middle line:84% But you have to act locally, and that's 00:37:14.870 --> 00:37:16.162 align:middle line:90% what we're doing in this class. 00:37:16.162 --> 00:37:19.070 align:middle line:90% 00:37:19.070 --> 00:37:20.780 align:middle line:84% When we're thinking globally, and we're 00:37:20.780 --> 00:37:23.240 align:middle line:84% thinking about ecological systems, 00:37:23.240 --> 00:37:28.490 align:middle line:84% we're thinking about the lithosphere, for the most part, 00:37:28.490 --> 00:37:32.150 align:middle line:84% and its interaction with the hydrosphere. 00:37:32.150 --> 00:37:40.030 align:middle line:84% The lithosphere is mostly silicon dioxide, 00:37:40.030 --> 00:37:44.880 align:middle line:84% a tiny bit of carbon, and it gets very hot, very quickly. 00:37:44.880 --> 00:37:52.760 align:middle line:84% So only the top 0.1%, that's the top four kilometers of it, 00:37:52.760 --> 00:37:56.510 align:middle line:84% is survivable in any type of organism we know of. 00:37:56.510 --> 00:37:59.720 align:middle line:84% About 110 degrees centigrade is where organisms start 00:37:59.720 --> 00:38:01.940 align:middle line:90% having a great deal of trouble. 00:38:01.940 --> 00:38:03.740 align:middle line:84% You and I would have trouble before that. 00:38:03.740 --> 00:38:07.570 align:middle line:90% 00:38:07.570 --> 00:38:11.010 align:middle line:84% The biosphere here is about 3 times 10 00:38:11.010 --> 00:38:21.150 align:middle line:84% to the 15th grams, my estimate for marine organisms. 00:38:21.150 --> 00:38:27.210 align:middle line:84% And maybe 10 to the 18th grams for all land plants and animals 00:38:27.210 --> 00:38:30.150 align:middle line:90% and microorganisms. 00:38:30.150 --> 00:38:36.425 align:middle line:84% The microbial hydrosphere here is about 10 00:38:36.425 --> 00:38:41.580 align:middle line:84% to the 21st milliliters, which works out about 10 00:38:41.580 --> 00:38:42.960 align:middle line:90% to the 27th cells. 00:38:42.960 --> 00:38:44.930 align:middle line:84% A phenomenal number of these cells, about 10 00:38:44.930 --> 00:38:47.640 align:middle line:84% to the 26th of those cells is a single species, 00:38:47.640 --> 00:38:52.710 align:middle line:84% which is prochlorococcus, which is responsible for maybe 00:38:52.710 --> 00:38:56.670 align:middle line:84% about 50% of the Earth's photosynthesis. 00:38:56.670 --> 00:39:00.660 align:middle line:84% A lot of that photosynthesis does not end up in fixed carbon 00:39:00.660 --> 00:39:03.570 align:middle line:84% because it is immediately consumed 00:39:03.570 --> 00:39:11.070 align:middle line:84% by one of its other predators, basically. 00:39:11.070 --> 00:39:18.097 align:middle line:84% But there's really quite a lot of cells out there which, 00:39:18.097 --> 00:39:19.680 align:middle line:84% to the extent that they're well-mixed, 00:39:19.680 --> 00:39:21.930 align:middle line:84% and they're not-- of course, it's not perfectly mixed, 00:39:21.930 --> 00:39:23.770 align:middle line:84% all the Pacific and Atlantic and so forth. 00:39:23.770 --> 00:39:25.860 align:middle line:84% But it's a considerably more well-mixed, 00:39:25.860 --> 00:39:27.480 align:middle line:84% say, than the lithosphere organisms, 00:39:27.480 --> 00:39:31.390 align:middle line:84% where you have organisms down a kilometer into the ground. 00:39:31.390 --> 00:39:34.900 align:middle line:84% They don't move around very rapidly. 00:39:34.900 --> 00:39:36.650 align:middle line:84% So when you have a population of that size 00:39:36.650 --> 00:39:38.610 align:middle line:84% as you remember from one of our first lectures 00:39:38.610 --> 00:39:41.520 align:middle line:84% on population size, the effect of population size 00:39:41.520 --> 00:39:49.578 align:middle line:84% determines the rate of drift and the optimality of the organism. 00:39:49.578 --> 00:39:51.120 align:middle line:84% So one of the things that we do, when 00:39:51.120 --> 00:39:54.660 align:middle line:84% we've been talking about mining the biosphere, one 00:39:54.660 --> 00:39:56.160 align:middle line:84% of the things that we're looking for 00:39:56.160 --> 00:40:03.540 align:middle line:84% are new tools that we can use for nanoengineering. 00:40:03.540 --> 00:40:07.170 align:middle line:84% And kind of one of my pet ones that I'm becoming interested 00:40:07.170 --> 00:40:08.670 align:middle line:84% in, we don't work on the same thing, 00:40:08.670 --> 00:40:12.870 align:middle line:84% but it is kind of interesting, is a set of-- we mentioned this 00:40:12.870 --> 00:40:18.715 align:middle line:84% briefly in the drug protein interaction lecture where-- 00:40:18.715 --> 00:40:21.090 align:middle line:84% but I don't think I mentioned this particular one-- where 00:40:21.090 --> 00:40:24.300 align:middle line:84% you have polyketides that go together. 00:40:24.300 --> 00:40:28.050 align:middle line:84% It's another polymer that has certain similarities 00:40:28.050 --> 00:40:29.670 align:middle line:84% to the basic polymers we talked about, 00:40:29.670 --> 00:40:33.990 align:middle line:84% but each step in it had to have a protein enzyme-- 00:40:33.990 --> 00:40:36.550 align:middle line:84% an enzymatic domain to accomplish that. 00:40:36.550 --> 00:40:38.550 align:middle line:84% And so one of these is tetracycline. 00:40:38.550 --> 00:40:41.040 align:middle line:84% And this is one of the more aromatic kind 00:40:41.040 --> 00:40:42.900 align:middle line:84% of coupled-aromatic compounds that's 00:40:42.900 --> 00:40:44.940 align:middle line:84% made because you have these [INAUDIBLE] 00:40:44.940 --> 00:40:50.250 align:middle line:84% and aromatases in addition to the polymerization steps. 00:40:50.250 --> 00:40:52.350 align:middle line:84% And so you make this thing that looks 00:40:52.350 --> 00:40:55.950 align:middle line:84% kind of like a poly aromatic hydrocarbon. 00:40:55.950 --> 00:41:01.200 align:middle line:84% And also in nature, you will find, in soot, in forest fires, 00:41:01.200 --> 00:41:02.970 align:middle line:84% in all kinds of natural phenomena, 00:41:02.970 --> 00:41:07.410 align:middle line:84% you will find polycyclic aromatic hydrocarbons, which 00:41:07.410 --> 00:41:10.710 align:middle line:84% some of you may know about as the potent carcinogens, 00:41:10.710 --> 00:41:14.670 align:middle line:84% but they're also just natural components. 00:41:14.670 --> 00:41:17.040 align:middle line:84% But they also-- you can start to see this 00:41:17.040 --> 00:41:19.650 align:middle line:84% looks a little bit like the buckyballs and buckytubes 00:41:19.650 --> 00:41:21.025 align:middle line:84% that we talked about when we were 00:41:21.025 --> 00:41:24.570 align:middle line:84% talking about molecular-type transistors. 00:41:24.570 --> 00:41:29.790 align:middle line:84% So the possibility of mining the biosphere for enzymes 00:41:29.790 --> 00:41:31.950 align:middle line:84% that act to synthesize or degrade 00:41:31.950 --> 00:41:33.510 align:middle line:84% this class of compounds, I think, 00:41:33.510 --> 00:41:34.930 align:middle line:90% would be just an example. 00:41:34.930 --> 00:41:36.130 align:middle line:90% We could list many others. 00:41:36.130 --> 00:41:39.150 align:middle line:84% But the part of it is just to dream, to imagine, 00:41:39.150 --> 00:41:41.590 align:middle line:84% what it is would like to find out there. 00:41:41.590 --> 00:41:46.950 align:middle line:84% And if there's an abundant source of it, 00:41:46.950 --> 00:41:48.720 align:middle line:90% then you should be able to-- 00:41:48.720 --> 00:41:51.010 align:middle line:84% abundant source of, say, the compound, in this case, 00:41:51.010 --> 00:41:55.030 align:middle line:84% you need to be able to find a microorganism and an enzyme 00:41:55.030 --> 00:41:57.060 align:middle line:84% that goes with that because there's a truly 00:41:57.060 --> 00:41:59.190 align:middle line:90% phenomenal amount of diversity. 00:41:59.190 --> 00:42:02.370 align:middle line:90% 00:42:02.370 --> 00:42:05.580 align:middle line:84% Another very important global consideration, 00:42:05.580 --> 00:42:08.610 align:middle line:84% rather than just mining it for new tools, 00:42:08.610 --> 00:42:11.520 align:middle line:84% is thinking about ways that either we could engineer 00:42:11.520 --> 00:42:16.890 align:middle line:84% or accidentally mess up our entire planet, 00:42:16.890 --> 00:42:22.800 align:middle line:84% as we could be doing with global warming or could do. 00:42:22.800 --> 00:42:25.860 align:middle line:84% And perhaps it's naive to think that we actually 00:42:25.860 --> 00:42:27.690 align:middle line:84% are having such a big effect because we 00:42:27.690 --> 00:42:32.100 align:middle line:84% know that global warming changes periodically over millennia. 00:42:32.100 --> 00:42:35.110 align:middle line:90% 00:42:35.110 --> 00:42:40.030 align:middle line:84% But it is very clear from the record 00:42:40.030 --> 00:42:42.410 align:middle line:84% just exactly how much carbon dioxide we are releasing 00:42:42.410 --> 00:42:46.258 align:middle line:84% and it makes sense, that it is consistent with the kind 00:42:46.258 --> 00:42:47.800 align:middle line:84% of temperature changes that have been 00:42:47.800 --> 00:42:50.530 align:middle line:84% observed since the Industrial Revolution. 00:42:50.530 --> 00:42:55.790 align:middle line:84% In any case, when you look at, in particular, 00:42:55.790 --> 00:42:57.850 align:middle line:90% the Southern Ocean-- 00:42:57.850 --> 00:42:59.958 align:middle line:90% up at the top of slide 27-- 00:42:59.958 --> 00:43:02.000 align:middle line:84% you can see there is a few places, in particular, 00:43:02.000 --> 00:43:04.330 align:middle line:84% Southern Ocean seems like it's a prime candidate where 00:43:04.330 --> 00:43:07.510 align:middle line:84% you have high nutrients but low chlorophyll. 00:43:07.510 --> 00:43:09.730 align:middle line:84% And you say, well, why would you have high nutrients 00:43:09.730 --> 00:43:14.930 align:middle line:84% but very little of the chlorophyll around, 00:43:14.930 --> 00:43:20.170 align:middle line:84% which is a tipoff that you have the photosynthetic bacteria. 00:43:20.170 --> 00:43:23.500 align:middle line:84% And the reason is that you have a limitation 00:43:23.500 --> 00:43:24.670 align:middle line:90% of some micronutrient. 00:43:24.670 --> 00:43:29.350 align:middle line:84% Micronutrient means it's not needed in the vast quantities 00:43:29.350 --> 00:43:32.270 align:middle line:84% that you need nitrogen, phosphorus, carbon, oxygen, 00:43:32.270 --> 00:43:33.470 align:middle line:90% and so on. 00:43:33.470 --> 00:43:36.070 align:middle line:84% So iron, typically, is the limiting micronutrient 00:43:36.070 --> 00:43:37.330 align:middle line:90% in the Southern Ocean. 00:43:37.330 --> 00:43:40.810 align:middle line:84% And so there actually have been little pilot experiments 00:43:40.810 --> 00:43:44.890 align:middle line:84% to drop iron off the backs of huge tankers. 00:43:44.890 --> 00:43:47.260 align:middle line:84% And there are now at least seven patents 00:43:47.260 --> 00:43:49.720 align:middle line:84% that have been filed on doing this in order 00:43:49.720 --> 00:43:53.740 align:middle line:84% to balance out the carbon credits for different nations. 00:43:53.740 --> 00:43:57.700 align:middle line:84% This is potentially a very big bit of terrestrial engineering 00:43:57.700 --> 00:43:58.840 align:middle line:90% that might happen. 00:43:58.840 --> 00:44:04.240 align:middle line:84% It involves trying to change phytoplankton. 00:44:04.240 --> 00:44:08.450 align:middle line:84% And as this line here begins to point out, 00:44:08.450 --> 00:44:11.050 align:middle line:84% even though phytoplankton are only 1% 00:44:11.050 --> 00:44:13.750 align:middle line:84% of the total global biomass, it's about 50% 00:44:13.750 --> 00:44:16.940 align:middle line:90% of the carbon fixation. 00:44:16.940 --> 00:44:19.570 align:middle line:84% What is the source of this 50-fold anomaly? 00:44:19.570 --> 00:44:24.530 align:middle line:84% Why is it that even though it's doing-- 00:44:24.530 --> 00:44:26.410 align:middle line:84% how is it doing 50% of the carbon fixation? 00:44:26.410 --> 00:44:28.750 align:middle line:84% And what's happening is a lot of the carbon 00:44:28.750 --> 00:44:32.443 align:middle line:84% is going right back out after being fixed. 00:44:32.443 --> 00:44:34.360 align:middle line:84% Instead of settling to the bottom of the ocean 00:44:34.360 --> 00:44:40.030 align:middle line:84% and never bothering us again, it gets returned. 00:44:40.030 --> 00:44:42.280 align:middle line:84% And the exact modeling of that requires 00:44:42.280 --> 00:44:45.460 align:middle line:84% a much deeper knowledge of exactly what organisms 00:44:45.460 --> 00:44:46.810 align:middle line:90% are present in the ocean. 00:44:46.810 --> 00:44:48.590 align:middle line:84% One of the problems with this however, 00:44:48.590 --> 00:44:52.750 align:middle line:84% is a genomic problem, in a certain sense, which 00:44:52.750 --> 00:44:56.680 align:middle line:84% is that very few of these organisms 00:44:56.680 --> 00:44:58.840 align:middle line:84% can be cultivated in the laboratory. 00:44:58.840 --> 00:45:02.230 align:middle line:84% On the order of 99.9% of the organisms that you sample 00:45:02.230 --> 00:45:04.180 align:middle line:84% from the ocean, or from the soil, 00:45:04.180 --> 00:45:07.090 align:middle line:84% a number of different environments, 00:45:07.090 --> 00:45:09.350 align:middle line:84% do not grow well in the laboratory. 00:45:09.350 --> 00:45:13.120 align:middle line:84% So if you were to study them, some people 00:45:13.120 --> 00:45:15.880 align:middle line:84% are feeling that the best route to studying them 00:45:15.880 --> 00:45:19.030 align:middle line:90% is going directly for-- 00:45:19.030 --> 00:45:21.100 align:middle line:84% looking at their genomes, without necessarily 00:45:21.100 --> 00:45:23.632 align:middle line:84% being able to grow them in the laboratory. 00:45:23.632 --> 00:45:26.410 align:middle line:90% 00:45:26.410 --> 00:45:30.850 align:middle line:84% Now, when we talk about this problem of-- 00:45:30.850 --> 00:45:34.780 align:middle line:84% even if we fertilize the oceans, and we 00:45:34.780 --> 00:45:37.240 align:middle line:84% can model this whole procedure, I won't go into details. 00:45:37.240 --> 00:45:47.390 align:middle line:84% But if there are predators there that take the fixed carbon 00:45:47.390 --> 00:45:49.820 align:middle line:84% and turn it back into carbon dioxide, 00:45:49.820 --> 00:45:52.190 align:middle line:84% you want to be able to monitor that. 00:45:52.190 --> 00:45:55.670 align:middle line:84% Now, of course, the ocean is a large, complex set 00:45:55.670 --> 00:45:59.540 align:middle line:84% of prokaryotic autotrophs that are, say, 00:45:59.540 --> 00:46:02.450 align:middle line:84% photosynthetic bacteria, and eukaryotic photosynthetic 00:46:02.450 --> 00:46:08.210 align:middle line:84% bacteria, and prokaryotic and eukaryotic predators 00:46:08.210 --> 00:46:09.570 align:middle line:90% of various sorts. 00:46:09.570 --> 00:46:12.470 align:middle line:84% And this is, I think, one of the more interesting 00:46:12.470 --> 00:46:18.560 align:middle line:84% predator-prey differential equation models. 00:46:18.560 --> 00:46:21.380 align:middle line:84% It's almost the same as many of the other differential 00:46:21.380 --> 00:46:26.880 align:middle line:84% equations that we did from the first lecture on growth. 00:46:26.880 --> 00:46:29.180 align:middle line:84% We did the logistic equation, some 00:46:29.180 --> 00:46:32.018 align:middle line:84% of the ones we did on repressor function, and so on. 00:46:32.018 --> 00:46:34.310 align:middle line:84% But the thing that's kind of interesting about this one 00:46:34.310 --> 00:46:40.160 align:middle line:84% is it's not simply a single-cell bacteria 00:46:40.160 --> 00:46:43.310 align:middle line:84% being eaten by a single-celled heterotroph, 00:46:43.310 --> 00:46:45.020 align:middle line:84% let's say, a blue-green algae being eaten 00:46:45.020 --> 00:46:46.490 align:middle line:90% by a single-cell heterotroph. 00:46:46.490 --> 00:46:48.940 align:middle line:84% Here, they are both multicellular. 00:46:48.940 --> 00:46:50.690 align:middle line:84% They're sort of the minimal multicellular. 00:46:50.690 --> 00:46:53.240 align:middle line:84% But because they are-- so Chlorella 00:46:53.240 --> 00:46:56.030 align:middle line:84% is one of the smallest plants, multicellular plants. 00:46:56.030 --> 00:47:03.920 align:middle line:84% And Brachionus is a small rotifer, a multicellular 00:47:03.920 --> 00:47:04.515 align:middle line:90% animal. 00:47:04.515 --> 00:47:06.140 align:middle line:84% Because they're multi-cellular, though, 00:47:06.140 --> 00:47:11.690 align:middle line:84% now you have the demographic mortality and fecundity 00:47:11.690 --> 00:47:14.120 align:middle line:84% of each of these things that has to be modeled in. 00:47:14.120 --> 00:47:18.950 align:middle line:84% That's the m and the lambda for Brachionus. 00:47:18.950 --> 00:47:25.400 align:middle line:84% And so this plays into these equations where you have here, 00:47:25.400 --> 00:47:28.700 align:middle line:84% in the upper-right of slide 28, the rate 00:47:28.700 --> 00:47:34.130 align:middle line:84% of change of nitrogen with respect to time, 00:47:34.130 --> 00:47:36.950 align:middle line:90% and concentration of Chlorella. 00:47:36.950 --> 00:47:40.370 align:middle line:84% So nitrogen is N, Chlorella is C. 00:47:40.370 --> 00:47:47.960 align:middle line:84% The R is the Brachionus that are reproducing 00:47:47.960 --> 00:47:51.390 align:middle line:84% and B are the Brachionus which are the total. 00:47:51.390 --> 00:47:54.410 align:middle line:84% And you can see each of these equations and how it plays out. 00:47:54.410 --> 00:47:55.940 align:middle line:90% Now this is how-- 00:47:55.940 --> 00:47:58.640 align:middle line:84% and the way they play it out here is you do it 00:47:58.640 --> 00:48:01.460 align:middle line:90% by dilution rate, delta, here. 00:48:01.460 --> 00:48:12.950 align:middle line:84% And the dilution rate is shown along the horizontal x-axis 00:48:12.950 --> 00:48:14.990 align:middle line:84% for each of these three plots here. 00:48:14.990 --> 00:48:19.430 align:middle line:84% And what they're modeling is on the far left, 00:48:19.430 --> 00:48:26.790 align:middle line:84% the nitrogen concentration as a function of dilution rate. 00:48:26.790 --> 00:48:34.640 align:middle line:84% And you can see you get these different sectors of behavior. 00:48:34.640 --> 00:48:36.380 align:middle line:84% When you look at the Chlorella, this 00:48:36.380 --> 00:48:40.970 align:middle line:84% is a green, photosynthetic, multicellular organism, 00:48:40.970 --> 00:48:43.760 align:middle line:84% in units of millions of cells per milliliter, 00:48:43.760 --> 00:48:47.620 align:middle line:90% that can take different-- 00:48:47.620 --> 00:48:51.770 align:middle line:84% it take different pathways here where, 00:48:51.770 --> 00:48:53.420 align:middle line:84% as a function of dilution rate, you 00:48:53.420 --> 00:48:55.610 align:middle line:84% get this bifurcation where you get 00:48:55.610 --> 00:49:03.380 align:middle line:84% a huge split in the possible concentrations of Chlorella 00:49:03.380 --> 00:49:07.220 align:middle line:84% is the one curve and the Brachionus, the predator, 00:49:07.220 --> 00:49:08.315 align:middle line:90% is the other curve. 00:49:08.315 --> 00:49:10.190 align:middle line:84% This kind of bifurcation is the sort of thing 00:49:10.190 --> 00:49:14.810 align:middle line:84% we talked about earlier, both in the logistic equation 00:49:14.810 --> 00:49:24.910 align:middle line:84% and in the self-exclusion modeling. 00:49:24.910 --> 00:49:29.290 align:middle line:84% And in the upper-right-hand, here's our friend 00:49:29.290 --> 00:49:31.320 align:middle line:84% the coefficient of variation in percent, 00:49:31.320 --> 00:49:36.160 align:middle line:84% ranging from 0 to 80 or so on the vertical axis 00:49:36.160 --> 00:49:37.730 align:middle line:90% and the dilution rate. 00:49:37.730 --> 00:49:42.550 align:middle line:84% And again, you can see that you get a peak in the dilution rate 00:49:42.550 --> 00:49:44.470 align:middle line:90% at around 0.75. 00:49:44.470 --> 00:49:49.550 align:middle line:84% The bottom two are models, and the upper right is the data. 00:49:49.550 --> 00:49:51.880 align:middle line:84% You can actually see, when you run real Chlorella 00:49:51.880 --> 00:49:55.150 align:middle line:84% and Brachionus in here, that you do get a peak just as you would 00:49:55.150 --> 00:49:57.295 align:middle line:84% predict in this kind of bifurcation analysis 00:49:57.295 --> 00:49:59.680 align:middle line:90% in kinetic modeling. 00:49:59.680 --> 00:50:04.960 align:middle line:84% That's due to-- where you get interesting behavior 00:50:04.960 --> 00:50:09.040 align:middle line:84% in this complicated-- or this fairly simple model ecosystem 00:50:09.040 --> 00:50:10.960 align:middle line:90% where you just have two species. 00:50:10.960 --> 00:50:15.070 align:middle line:84% In the ocean, of course, you have a lot more. 00:50:15.070 --> 00:50:17.230 align:middle line:84% Now, that's out in the wide world 00:50:17.230 --> 00:50:20.890 align:middle line:84% where you have 10 to the 26 cells. 00:50:20.890 --> 00:50:24.310 align:middle line:84% But inside every one of us in this room, no offense, 00:50:24.310 --> 00:50:29.170 align:middle line:84% but there are about 10 times as many non-human cells 00:50:29.170 --> 00:50:32.050 align:middle line:84% in each of us as are human cells. 00:50:32.050 --> 00:50:35.020 align:middle line:90% And we have our own ecology. 00:50:35.020 --> 00:50:39.970 align:middle line:84% And there is a literature where certain organisms 00:50:39.970 --> 00:50:41.740 align:middle line:84% we know cause diseases, and they're 00:50:41.740 --> 00:50:44.980 align:middle line:84% put into one category, infectious diseases, 00:50:44.980 --> 00:50:46.450 align:middle line:84% but then there's another set that 00:50:46.450 --> 00:50:48.140 align:middle line:90% takes some time to sort out. 00:50:48.140 --> 00:50:50.110 align:middle line:84% And you can see here are candidate diseases 00:50:50.110 --> 00:50:51.520 align:middle line:90% and candidate organisms. 00:50:51.520 --> 00:50:53.980 align:middle line:84% And all these little bars that are going in here 00:50:53.980 --> 00:50:57.310 align:middle line:84% are references that you can look up 00:50:57.310 --> 00:51:00.520 align:middle line:84% where maybe it's not an airtight case yet, 00:51:00.520 --> 00:51:02.320 align:middle line:84% or maybe it's not even-- maybe there's 00:51:02.320 --> 00:51:05.110 align:middle line:84% controversial or discredited [INAUDIBLE],, 00:51:05.110 --> 00:51:07.990 align:middle line:84% but these are links where, eventually, this will be 00:51:07.990 --> 00:51:10.000 align:middle line:90% called an infectious disease. 00:51:10.000 --> 00:51:12.520 align:middle line:84% Helicobacter pylori causes stomach ulcers. 00:51:12.520 --> 00:51:14.935 align:middle line:84% I think that fact's basically already accepted. 00:51:14.935 --> 00:51:17.980 align:middle line:90% 00:51:17.980 --> 00:51:20.260 align:middle line:84% Some Australian doctor named Marshall 00:51:20.260 --> 00:51:22.093 align:middle line:84% decided to drink a bunch of Helicobacter 00:51:22.093 --> 00:51:24.010 align:middle line:84% to prove to his colleagues, who didn't believe 00:51:24.010 --> 00:51:26.830 align:middle line:84% any stomach ulcers were caused by Helicobacter, much less all 00:51:26.830 --> 00:51:27.640 align:middle line:90% of them. 00:51:27.640 --> 00:51:31.960 align:middle line:84% And he drank it, he cured himself, he drank it again. 00:51:31.960 --> 00:51:37.420 align:middle line:84% And he caused stomach ulcers in himself. 00:51:37.420 --> 00:51:41.560 align:middle line:84% And hopefully, none of you will volunteer 00:51:41.560 --> 00:51:45.190 align:middle line:84% to do similar things with some of these other nasty guys 00:51:45.190 --> 00:51:47.890 align:middle line:84% here because I think some of the list of things 00:51:47.890 --> 00:51:50.320 align:middle line:84% that they cause are a little more serious than stomach 00:51:50.320 --> 00:51:52.990 align:middle line:84% ulcers, and the cures are a little less obvious 00:51:52.990 --> 00:51:55.240 align:middle line:84% than the cures for a gram-negative bacterium 00:51:55.240 --> 00:51:58.840 align:middle line:90% like Helicobacter. 00:51:58.840 --> 00:52:01.600 align:middle line:84% There have been various efforts started 00:52:01.600 --> 00:52:05.950 align:middle line:84% to actually mine what's the connection with genomics 00:52:05.950 --> 00:52:07.210 align:middle line:90% and computational biology? 00:52:07.210 --> 00:52:12.340 align:middle line:84% To mine the transcriptome for evidence of these. 00:52:12.340 --> 00:52:14.140 align:middle line:84% You can look for bacteria very easily 00:52:14.140 --> 00:52:19.660 align:middle line:84% because they have ribosomal RNA components which you can PCR. 00:52:19.660 --> 00:52:25.600 align:middle line:84% Viruses are more complicated because they're not conserved, 00:52:25.600 --> 00:52:28.390 align:middle line:84% or they don't have any universally conserved element 00:52:28.390 --> 00:52:29.980 align:middle line:90% like ribosomal RNA. 00:52:29.980 --> 00:52:33.910 align:middle line:84% But there have been efforts to look 00:52:33.910 --> 00:52:36.523 align:middle line:84% for what's present in the human transcriptome that's not 00:52:36.523 --> 00:52:37.690 align:middle line:90% present in the human genome. 00:52:37.690 --> 00:52:42.070 align:middle line:84% The human genome is something highly purified and cloned 00:52:42.070 --> 00:52:44.170 align:middle line:84% and so forth, while the transcriptome 00:52:44.170 --> 00:52:48.550 align:middle line:84% is whatever cells they ground up that day and they. 00:52:48.550 --> 00:52:52.810 align:middle line:84% And indeed, lurking in there are many of these hepatitis, 00:52:52.810 --> 00:52:57.730 align:middle line:84% papilloma, Epstein-Barr virus, retroviral-like elements which 00:52:57.730 --> 00:53:00.280 align:middle line:84% are not present in the human genome as yet. 00:53:00.280 --> 00:53:02.390 align:middle line:84% Human genome is not completely finished. 00:53:02.390 --> 00:53:04.330 align:middle line:84% So that's still an escape clause. 00:53:04.330 --> 00:53:10.660 align:middle line:84% But basically, these are smoking guns for at least commensals. 00:53:10.660 --> 00:53:15.760 align:middle line:84% Microorganisms and viruses that are living in tissues, some 00:53:15.760 --> 00:53:17.620 align:middle line:90% of them may be tissue-specific. 00:53:17.620 --> 00:53:20.343 align:middle line:84% Some of them, subset, may actually cause disease. 00:53:20.343 --> 00:53:21.760 align:middle line:84% And that's the whole problem then, 00:53:21.760 --> 00:53:23.510 align:middle line:90% is sorting out cause and effect. 00:53:23.510 --> 00:53:27.460 align:middle line:84% We know how it's been done, heroically, 00:53:27.460 --> 00:53:28.960 align:middle line:84% with Helicobacter pylori, but how do 00:53:28.960 --> 00:53:30.168 align:middle line:90% we do it with the other ones? 00:53:30.168 --> 00:53:32.125 align:middle line:84% Some of them will have tissue-culture models. 00:53:32.125 --> 00:53:36.430 align:middle line:90% 00:53:36.430 --> 00:53:41.110 align:middle line:84% At the time this slide was done, but so-- 00:53:41.110 --> 00:53:45.790 align:middle line:84% this is just illustrating kind of the flow 00:53:45.790 --> 00:53:52.030 align:middle line:84% of nucleic acid information that comes in about-- 00:53:52.030 --> 00:53:55.420 align:middle line:84% that was capable of being mined for new microorganisms, 00:53:55.420 --> 00:53:58.420 align:middle line:84% new viruses that might be present in a tissue-specific 00:53:58.420 --> 00:54:01.250 align:middle line:90% fashion, might be pathogens. 00:54:01.250 --> 00:54:03.370 align:middle line:84% These are some of the most popular sequences 00:54:03.370 --> 00:54:04.810 align:middle line:90% from which we're mining. 00:54:04.810 --> 00:54:09.520 align:middle line:84% Obviously, human was popular at times [INAUDIBLE] 00:54:09.520 --> 00:54:11.240 align:middle line:90% the human genome was in place. 00:54:11.240 --> 00:54:13.000 align:middle line:84% Here's Brachionus, our friend, which 00:54:13.000 --> 00:54:15.730 align:middle line:84% was pretty high up on this list, considering most of you 00:54:15.730 --> 00:54:18.200 align:middle line:84% probably didn't hear about Brachionus before this lecture. 00:54:18.200 --> 00:54:19.990 align:middle line:90% Now you've heard of it twice. 00:54:19.990 --> 00:54:23.440 align:middle line:84% There just was a spurt in sequencing interest. 00:54:23.440 --> 00:54:25.930 align:middle line:84% And of course, HIV is the winner year 00:54:25.930 --> 00:54:29.080 align:middle line:84% after year because of the importance of resequencing it 00:54:29.080 --> 00:54:31.400 align:middle line:90% for new mutations. 00:54:31.400 --> 00:54:35.920 align:middle line:84% And it really has a record number of new mutations. 00:54:35.920 --> 00:54:43.120 align:middle line:84% And we've talked about HIV from a variety 00:54:43.120 --> 00:54:45.400 align:middle line:90% of different standpoints. 00:54:45.400 --> 00:54:48.280 align:middle line:84% One of them is polymerase, is protease, 00:54:48.280 --> 00:54:52.720 align:middle line:84% is drug targets, and as a source of drug resistance. 00:54:52.720 --> 00:54:56.550 align:middle line:84% And that you can follow up that model, that sort 00:54:56.550 --> 00:54:59.470 align:middle line:84% of atomic-scale protein modeling, 00:54:59.470 --> 00:55:04.090 align:middle line:84% where you're looking for new drugs 00:55:04.090 --> 00:55:06.250 align:middle line:84% where the polymerase, the mutant polymerase, 00:55:06.250 --> 00:55:08.260 align:middle line:84% will not be resistant to the new drug. 00:55:08.260 --> 00:55:10.600 align:middle line:84% But you can also monitor at the population scale. 00:55:10.600 --> 00:55:13.990 align:middle line:84% As HIV goes to a particular patient, 00:55:13.990 --> 00:55:20.530 align:middle line:84% or to a population of patients, how does the drug resistance 00:55:20.530 --> 00:55:22.180 align:middle line:90% change as a function of time? 00:55:22.180 --> 00:55:25.840 align:middle line:84% Here, the horizontal axis is in days, up to 30 days, 00:55:25.840 --> 00:55:35.430 align:middle line:84% and then the vertical axis is the titers 00:55:35.430 --> 00:55:40.120 align:middle line:84% of the viruses [INAUDIBLE] as a function of time. 00:55:40.120 --> 00:55:48.770 align:middle line:84% And what you have here is rates of exponential increase. 00:55:48.770 --> 00:55:51.500 align:middle line:84% These are all on logarithmic scales. 00:55:51.500 --> 00:55:54.020 align:middle line:84% And where you're modeling such a thing 00:55:54.020 --> 00:55:57.643 align:middle line:84% as the clearance to the immune system and so forth. 00:55:57.643 --> 00:55:59.060 align:middle line:84% Originally, this virus was thought 00:55:59.060 --> 00:56:06.020 align:middle line:84% to be a very slowly replicating virus, almost cryptic. 00:56:06.020 --> 00:56:07.400 align:middle line:84% And then later was found out this 00:56:07.400 --> 00:56:10.640 align:middle line:84% is actually a very rapidly replicating, but immune system 00:56:10.640 --> 00:56:13.817 align:middle line:90% is very rapidly responding. 00:56:13.817 --> 00:56:15.650 align:middle line:84% Just each of these models we've gone through 00:56:15.650 --> 00:56:17.910 align:middle line:84% has a set of parameters that it goes along with. 00:56:17.910 --> 00:56:19.610 align:middle line:84% In a certain sense, there's quite a bit 00:56:19.610 --> 00:56:22.670 align:middle line:84% you can learn about a model just by seeing what parameters are 00:56:22.670 --> 00:56:26.300 align:middle line:84% present or absent, whether they're experimentally 00:56:26.300 --> 00:56:28.820 align:middle line:84% determined, how accurately, and so on. 00:56:28.820 --> 00:56:33.680 align:middle line:84% And model parameters here are mortality rate of uninfected 00:56:33.680 --> 00:56:37.915 align:middle line:84% CD4+ T-cells, same thing for infected cells, 00:56:37.915 --> 00:56:40.040 align:middle line:84% and they can have different [INAUDIBLE] you can see 00:56:40.040 --> 00:56:44.300 align:middle line:84% on the far right that there's a huge difference between a diet 00:56:44.300 --> 00:56:50.250 align:middle line:84% of 1/4 per day versus [INAUDIBLE] and so on. 00:56:50.250 --> 00:56:55.250 align:middle line:84% There's rates for getting infected, rates of virus loss, 00:56:55.250 --> 00:57:05.900 align:middle line:84% production, threshold value for remission, so on and so forth. 00:57:05.900 --> 00:57:09.250 align:middle line:84% And these are important parameters 00:57:09.250 --> 00:57:13.540 align:middle line:84% for how the virus population changes 00:57:13.540 --> 00:57:15.808 align:middle line:84% within a person or the population. 00:57:15.808 --> 00:57:17.350 align:middle line:84% How it spreads through the population 00:57:17.350 --> 00:57:24.560 align:middle line:84% depends on, in many cases, on the herd immunity and so forth. 00:57:24.560 --> 00:57:27.550 align:middle line:84% And we get into issues of public health. 00:57:27.550 --> 00:57:30.870 align:middle line:90% 00:57:30.870 --> 00:57:40.110 align:middle line:84% It's almost a truism that most of the additional quality 00:57:40.110 --> 00:57:43.770 align:middle line:84% years of life, quality-adjusted life years here, 00:57:43.770 --> 00:57:48.960 align:middle line:84% QALY, that we have in the world the fact that our life 00:57:48.960 --> 00:57:50.730 align:middle line:84% expectancy is so much longer and so forth, 00:57:50.730 --> 00:57:55.260 align:middle line:84% are mainly public health decisions that have been made, 00:57:55.260 --> 00:57:57.060 align:middle line:90% not so much pharmaceutical. 00:57:57.060 --> 00:58:01.040 align:middle line:90% 00:58:01.040 --> 00:58:08.100 align:middle line:84% Such things as clean water have made a huge impact. 00:58:08.100 --> 00:58:09.560 align:middle line:90% And so we need to think-- 00:58:09.560 --> 00:58:12.455 align:middle line:84% but even when you think about the pharmaceutical aspects, 00:58:12.455 --> 00:58:14.580 align:middle line:84% it needs to be thought of in a public health sense. 00:58:14.580 --> 00:58:18.050 align:middle line:84% And a surprising amount of public health officials 00:58:18.050 --> 00:58:23.300 align:middle line:84% in even the most developed nations, maybe most of them 00:58:23.300 --> 00:58:25.640 align:middle line:84% do not have formal education in public health. 00:58:25.640 --> 00:58:30.320 align:middle line:84% And so I urge you all to get at least some education in this. 00:58:30.320 --> 00:58:32.480 align:middle line:90% Because in this, we have-- 00:58:32.480 --> 00:58:36.050 align:middle line:84% here, they're trying to build and have 00:58:36.050 --> 00:58:41.220 align:middle line:84% built quantitative models, not only for the kinetics 00:58:41.220 --> 00:58:43.880 align:middle line:84% which a disease will go through a population or even 00:58:43.880 --> 00:58:50.400 align:middle line:84% through an individual, but the way that you make decisions. 00:58:50.400 --> 00:58:53.060 align:middle line:84% And this is having an impact on the way decisions are-- 00:58:53.060 --> 00:58:55.580 align:middle line:84% how these different projects are prioritized. 00:58:55.580 --> 00:58:57.320 align:middle line:84% At any given time, you might have 00:58:57.320 --> 00:59:01.820 align:middle line:84% hundreds of different candidate vaccines, which 00:59:01.820 --> 00:59:05.810 align:middle line:84% is one of the major, very effective public health 00:59:05.810 --> 00:59:08.300 align:middle line:90% strategies. 00:59:08.300 --> 00:59:10.350 align:middle line:84% And this is the way they prioritize it. 00:59:10.350 --> 00:59:14.030 align:middle line:84% So the level one, saves money and it improves life. 00:59:14.030 --> 00:59:17.750 align:middle line:90% So that's almost a no-brainer. 00:59:17.750 --> 00:59:19.140 align:middle line:90% Then you have different levels. 00:59:19.140 --> 00:59:24.350 align:middle line:84% We have it might cost a $10,000, or $100,000 00:59:24.350 --> 00:59:26.915 align:middle line:84% per quality-adjusted life year saved. 00:59:26.915 --> 00:59:30.200 align:middle line:90% 00:59:30.200 --> 00:59:33.860 align:middle line:84% And so the level one candidates are cytomegalovirus, 00:59:33.860 --> 00:59:36.260 align:middle line:90% and therapeutic vaccines. 00:59:36.260 --> 00:59:38.720 align:middle line:84% These are not aimed at an infectious disease, 00:59:38.720 --> 00:59:42.500 align:middle line:84% they're for diabetes, rheumatoid arthritis, multiple sclerosis, 00:59:42.500 --> 00:59:44.195 align:middle line:90% various bacterial ones. 00:59:44.195 --> 00:59:46.070 align:middle line:84% And of course, HIV was not even in this study 00:59:46.070 --> 00:59:50.030 align:middle line:84% because it was such a high priority already 00:59:50.030 --> 00:59:50.750 align:middle line:90% within the NIH. 00:59:50.750 --> 00:59:55.773 align:middle line:90% 00:59:55.773 --> 00:59:58.190 align:middle line:84% So is there a role for genomics and computational biology, 00:59:58.190 --> 01:00:01.280 align:middle line:84% the title of this course, in this vaccine research 01:00:01.280 --> 01:00:02.840 align:middle line:90% and development? 01:00:02.840 --> 01:00:05.090 align:middle line:84% I think the answer is yes, but it 01:00:05.090 --> 01:00:07.940 align:middle line:84% requires a great deal of creativity and resourcefulness 01:00:07.940 --> 01:00:10.130 align:middle line:90% on all of your parts. 01:00:10.130 --> 01:00:13.730 align:middle line:84% There are new opportunities with DNA vaccines where 01:00:13.730 --> 01:00:18.020 align:middle line:84% you can have one or more DNAs shot intramuscularly 01:00:18.020 --> 01:00:20.750 align:middle line:90% or a variety of other ways. 01:00:20.750 --> 01:00:23.000 align:middle line:84% What is delivered by the DNA vaccine 01:00:23.000 --> 01:00:27.740 align:middle line:84% can be various so-called intracellular vaccines, 01:00:27.740 --> 01:00:30.110 align:middle line:84% which can either act through cell-mediated immunity 01:00:30.110 --> 01:00:32.990 align:middle line:84% or in some other way intracellularly. 01:00:32.990 --> 01:00:41.480 align:middle line:84% RNAi is a rapidly emerging way of delivering things 01:00:41.480 --> 01:00:44.300 align:middle line:84% that may not be classically considered 01:00:44.300 --> 01:00:47.270 align:middle line:90% therapeutics or vaccines. 01:00:47.270 --> 01:00:49.760 align:middle line:90% The concept of multiplexing-- 01:00:49.760 --> 01:00:57.440 align:middle line:84% many of us have received multiple vaccines at a time. 01:00:57.440 --> 01:00:59.330 align:middle line:84% Typically, every year, your flu vaccine 01:00:59.330 --> 01:01:01.170 align:middle line:84% will have two or three strains in it. 01:01:01.170 --> 01:01:06.020 align:middle line:84% But as you can see from this middle article, 01:01:06.020 --> 01:01:07.880 align:middle line:84% the diversity of certain diseases, 01:01:07.880 --> 01:01:13.340 align:middle line:84% like HIV and influenza, almost demand attention from genomics. 01:01:13.340 --> 01:01:16.760 align:middle line:84% The diversity that can occur, either from year to year 01:01:16.760 --> 01:01:25.050 align:middle line:84% or throughout the entire global set of viruses. 01:01:25.050 --> 01:01:27.740 align:middle line:84% Another opportunity is when we have 01:01:27.740 --> 01:01:32.120 align:middle line:84% vectors for, say, arthropod vectors, 01:01:32.120 --> 01:01:34.070 align:middle line:90% insect vectors like mosquito. 01:01:34.070 --> 01:01:40.070 align:middle line:84% And we have opportunities to hit malaria and all 01:01:40.070 --> 01:01:41.417 align:middle line:90% the different life stages. 01:01:41.417 --> 01:01:43.250 align:middle line:84% There are multiple life stages within humans 01:01:43.250 --> 01:01:48.440 align:middle line:84% and multiple stages within the insect vector. 01:01:48.440 --> 01:01:50.720 align:middle line:84% And now, the genomics of these two just 01:01:50.720 --> 01:01:56.750 align:middle line:84% came out recently, in the same week, in Nature and Science. 01:01:56.750 --> 01:02:01.517 align:middle line:84% This provides a whole new set of inspirations for work 01:02:01.517 --> 01:02:02.600 align:middle line:90% that can be done on these. 01:02:02.600 --> 01:02:05.000 align:middle line:84% I think we should take a little break now, 01:02:05.000 --> 01:02:09.250 align:middle line:84% and we'll wrap up the talk after that break.