WEBVTT 00:00:03.217 --> 00:00:06.517 This is the Convergent Science Network podcast. 00:00:08.537 --> 00:00:12.997 Leading researchers in the domain of neuroscience, brain theory and technology 00:00:12.997 --> 00:00:16.757 are interviewed by Paul Vershoor and Tony Prescott. 00:00:17.877 --> 00:00:21.377 Ready to roll? Okay, this is Paul Vershoor with the Convergent Science Network 00:00:21.377 --> 00:00:23.217 together with Tony Prescott. 00:00:24.137 --> 00:00:29.617 And our guest today is Yaki Setti, who was speaking at the BCBT Summer School 00:00:29.617 --> 00:00:35.617 on actually synthetically building different kinds of organs. 00:00:36.197 --> 00:00:41.357 So, Jaki, how do we build a synthetic organ? 00:00:41.817 --> 00:00:46.777 Well, first we have to know the biology. We have to know what God created or 00:00:46.777 --> 00:00:49.857 what biology created to rebuild it. 00:00:50.197 --> 00:00:54.257 We need to know the biology better than the biologist and better than the biology 00:00:54.257 --> 00:00:58.457 knows itself because we need to reconstruct it. So in a way, 00:00:58.497 --> 00:01:04.877 I don't want to sound too sophisticated, but we're playing kind of creator. 00:01:05.497 --> 00:01:09.857 We create something. We use the computer as a platform to play with the data, 00:01:09.997 --> 00:01:12.737 put the data into the computer and play with it. 00:01:13.057 --> 00:01:16.957 And we know what the desired output is. 00:01:17.057 --> 00:01:20.637 We don't know all about it, but we know if something is missing, 00:01:20.837 --> 00:01:23.437 we will see that the output will be wrong. 00:01:24.257 --> 00:01:29.617 So that's the... But no, this doesn't just happen by accident, right? 00:01:29.737 --> 00:01:33.857 It's not that you just have, let's say, a random set of ingredients like some 00:01:33.857 --> 00:01:36.777 big minestrone and now that comes an organ like a pancreas. 00:01:36.937 --> 00:01:44.337 So the first example you discussed was to construct, to really grow in silico a pancreas. 00:01:44.337 --> 00:01:50.797 And then you showed to us that that model of the developed pancreas shared many 00:01:50.797 --> 00:01:56.397 features with the biological pancreas. So the pancreas is actually built from 00:01:56.397 --> 00:01:58.177 different kinds of cells, right? 00:01:58.237 --> 00:02:01.937 And the morphology of the structure will change over time in some developmental process. 00:02:02.357 --> 00:02:08.297 So how do we exactly model that? How do you model such a cell and how do these 00:02:08.297 --> 00:02:11.917 cells then collectively form something that you could call a pancreas? 00:02:11.917 --> 00:02:15.037 So in each cell, you put the biological data from the papers. 00:02:15.437 --> 00:02:17.777 So you have a block diagram. 00:02:18.763 --> 00:02:26.143 Which blocks and arrows, and each arrow is a transit between the two states of the block. 00:02:26.563 --> 00:02:32.403 So you take the cell and you define the state it may be in, and this is correlated to the biology. 00:02:32.643 --> 00:02:38.543 You have to justify every block and every arrow you put in this diagram. 00:02:39.003 --> 00:02:41.543 So for your cell model, how many states can the cell attain? 00:02:41.823 --> 00:02:45.843 Well, it depends which model it is. The pancreas model, it's the most complicated, 00:02:46.023 --> 00:02:51.343 and it had over 150 states all over. 00:02:52.223 --> 00:02:58.863 The number of the differentiation state was, I believe, like 10, 00:02:59.023 --> 00:03:06.063 which is more or less the number of markers that the biologists found. 00:03:06.343 --> 00:03:13.943 So for each state or stage, the biologists have defined we have a state in our 00:03:13.943 --> 00:03:15.643 differentiation component. 00:03:15.923 --> 00:03:19.143 And it's the same for other components in the system. 00:03:19.203 --> 00:03:24.303 For example, the proliferation component has five states, which it is correlated 00:03:24.303 --> 00:03:26.743 with the five stages of cell cycle. 00:03:26.963 --> 00:03:30.643 So we are trying to put into the model and 00:03:30.643 --> 00:03:33.683 to formulate as many biological data as possible 00:03:33.683 --> 00:03:37.023 and to be able to 00:03:37.023 --> 00:03:39.943 justify everything for example if we put an 00:03:39.943 --> 00:03:42.923 assumption or an error or a block in the diagram 00:03:42.923 --> 00:03:45.843 so each one of that have this we 00:03:45.843 --> 00:03:51.543 have to we are able to say okay this data was taken from this paper and that's 00:03:51.543 --> 00:03:57.183 why we constrain the model to be biologically plausible or accurate or at least 00:03:57.183 --> 00:04:03.183 justified right but then so we have the cell the cell can have a large number of states 00:04:03.443 --> 00:04:06.343 then there are certain inputs to that cell that 00:04:06.343 --> 00:04:09.203 will make it actually decide which state it attains now 00:04:09.203 --> 00:04:12.583 that transformation of inputs to the 00:04:12.583 --> 00:04:18.483 state expressed that's more like a logical lookup table like if my inputs have 00:04:18.483 --> 00:04:22.723 the following characteristics then in a deterministic fashion i turn into state 00:04:22.723 --> 00:04:28.503 one or is that transfer function following more the kind of messenger systems 00:04:28.503 --> 00:04:30.183 that you would find within a cell. 00:04:30.423 --> 00:04:33.103 Yeah, it is temporal decision. 00:04:33.403 --> 00:04:35.023 It's based on time and space. 00:04:35.383 --> 00:04:42.003 You will get some kind of message at a certain point if you are exactly at the 00:04:42.003 --> 00:04:47.103 same certain position and has the same specific message. 00:04:47.703 --> 00:04:55.183 So it is deterministic in that kind of perspective if you look at the individual cell. all. 00:04:55.703 --> 00:05:00.523 But if you look at the population level, you'll see that there are a lot of different groups. 00:05:01.461 --> 00:05:06.101 A lot of stochastic decisions are made. For example, if cell number one made 00:05:06.101 --> 00:05:11.041 a decision at one point and cell number two made a decision at another point, 00:05:11.141 --> 00:05:15.441 and they switch roles, so your simulation will be a little bit different. 00:05:15.661 --> 00:05:19.921 It is equivalent to embryos, to human beings. 00:05:20.221 --> 00:05:24.541 All of us have two hands, have two eyes, have one head, two legs, 00:05:24.701 --> 00:05:32.081 but still there is a lot of variety, and part of it is the genetic issue that 00:05:32.081 --> 00:05:35.221 we take with us, and part of it is part of our development. 00:05:35.521 --> 00:05:40.721 So this is why two brothers are not the same, and even identical twins are not identical. 00:05:42.661 --> 00:05:46.941 So, I mean, your model of a cell is a kind of informational model where you 00:05:46.941 --> 00:05:51.061 describe the states it can be in, the kind of switches within it, 00:05:51.081 --> 00:05:54.341 whether the switches can be thrown back, and so on. 00:05:54.341 --> 00:06:02.261 And it's sensitive both to external, I guess, chemicals that can affect the 00:06:02.261 --> 00:06:06.121 membrane and therefore change the internal state, but also internal mechanisms 00:06:06.121 --> 00:06:08.981 that can throw switches. Does that summarize it? 00:06:09.141 --> 00:06:13.601 Yes, that's more or less it. The cell itself, we look at this as a three component, 00:06:13.781 --> 00:06:17.541 the component of the cell and a component of external sensors, 00:06:17.801 --> 00:06:19.961 which we call the membrane or the receptors. 00:06:19.961 --> 00:06:24.321 And there are the receptors and there are the factor, it's not only the genetic 00:06:24.321 --> 00:06:27.241 ones it's not only genes, it can be genes in different states. 00:06:27.601 --> 00:06:29.361 That would summarize it nicely. 00:06:29.821 --> 00:06:34.641 And the specific thing that is exciting about this of course is that you can 00:06:34.641 --> 00:06:42.601 model stem cells which have the exciting property that they can turn into other kinds of cells. 00:06:42.921 --> 00:06:48.321 Yeah, they can differentiate. So perhaps you could because this audience of 00:06:48.321 --> 00:06:52.321 BCBT is mainly interested in neuron cells. 00:06:52.501 --> 00:06:55.541 Perhaps you could fill us in a bit more about the different kinds of stem cells 00:06:55.541 --> 00:06:58.561 and what we know about them. Well, it depends. 00:06:59.541 --> 00:07:01.221 Okay, stem cells in... 00:07:02.441 --> 00:07:06.981 The core of stem cell research is cells that can differentiate and proliferate 00:07:06.981 --> 00:07:09.361 and carry their own data. 00:07:10.761 --> 00:07:15.881 But lately, a lot of research was done and they claimed that stem cells, 00:07:16.101 --> 00:07:19.601 this naïve definition is not enough. 00:07:19.961 --> 00:07:22.921 And there is a notion of stemsness. 00:07:23.361 --> 00:07:27.201 It's kind of how much stem this cell is. 00:07:27.581 --> 00:07:35.201 And it means that even in a growth tree, if it can differentiate to three different cells. 00:07:37.041 --> 00:07:43.101 Or can get any kind of cell like the embryo cell like the Zygote, 00:07:43.101 --> 00:07:45.921 The first cell is like the original cell, 00:07:46.061 --> 00:07:52.841 and this is more or less the new notion of stem cells, but the original one 00:07:52.841 --> 00:07:59.621 is a cell population that can maintain itself and differentiate. 00:08:00.733 --> 00:08:03.673 And in almost any organ, you can find stem cells. 00:08:03.873 --> 00:08:08.273 And even you can, I understand that you can take cells which aren't stem cells, 00:08:08.333 --> 00:08:10.933 and by stressing them, I think I heard this about blood cells, 00:08:11.053 --> 00:08:13.253 was it? You could maybe turn them back into stem cells. 00:08:13.373 --> 00:08:15.593 Well, I don't like the term turn them back. 00:08:15.733 --> 00:08:22.593 You can turn them into new stage. You can turn the stemness mechanism of the cells on. 00:08:23.033 --> 00:08:27.213 They start to behave like a stem cell. Right. And so the behaviors of the cells 00:08:27.213 --> 00:08:29.973 in your model, can you just summarize the things that they can do? 00:08:29.973 --> 00:08:38.513 They can do the two most important mechanisms of differentiating. 00:08:38.673 --> 00:08:44.433 They can go from one state to a variety of states. There is kind of a tree of 00:08:44.433 --> 00:08:46.053 decision that the cell can make. 00:08:46.233 --> 00:08:51.193 And in parallel, orthogonal component would be to proliferate. 00:08:51.673 --> 00:08:59.073 They can create new instance of the same cell. And this is enough to maintain a stem cell population. 00:09:00.133 --> 00:09:04.313 But in our model there is an additional requirement 00:09:04.313 --> 00:09:10.973 and this is their position in space they can move to another place this is how 00:09:10.973 --> 00:09:17.073 we can create structure of organs and we can create different population and 00:09:17.073 --> 00:09:23.433 as we have a front end we visualize the model you can see it, 00:09:23.473 --> 00:09:27.813 the user can view it and can interact with it sometimes you can even say Say, 00:09:27.813 --> 00:09:29.993 okay, I want cell number five, 00:09:30.153 --> 00:09:34.213 six, and seven to be erased from the simulation and see what's going on. 00:09:34.413 --> 00:09:38.793 Or to change the environmental effect at runtime. 00:09:39.033 --> 00:09:45.853 So at the larval stage or when the model is young, it's not an adult yet, 00:09:45.933 --> 00:09:50.553 I want to keep the environment as it should be. 00:09:50.633 --> 00:09:54.313 But at some point, I want to decide that, okay, after three days, 00:09:54.373 --> 00:09:57.493 I want to change the environment. I want to inject a new material. 00:09:57.733 --> 00:10:04.653 Or you want to kill 10% of the cell and then to see what the simulation suggests the result will be. 00:10:05.053 --> 00:10:10.233 Right. So your model consists of a number of initial cells, maybe even just 00:10:10.233 --> 00:10:15.593 one, and then an environment in which the cell lives and it's communicating with the environment. 00:10:15.993 --> 00:10:20.293 And with each other. And with each other, yeah. So in terms of the movement 00:10:20.293 --> 00:10:25.193 of the cell, it's able to do things like move up gradients, chemical gradients. 00:10:25.193 --> 00:10:28.553 Yeah, it senses the gradients in your environment and can move towards it. 00:10:29.213 --> 00:10:32.253 So can you just summarize what the environment looks like from the point of 00:10:32.253 --> 00:10:35.353 view of simulation? So you have chemical gradients in it, for instance. 00:10:35.613 --> 00:10:38.633 Do you have physical structures beyond that? Yeah, okay. The environment is 00:10:38.633 --> 00:10:43.493 divided into voxels, three-dimensional pixels, boxes in the space. 00:10:43.753 --> 00:10:49.193 And inside there is a gradient of chemicals that exist in each part. 00:10:49.193 --> 00:10:55.013 In many cases, the decision what are the values is based on the differential 00:10:55.013 --> 00:10:57.593 equations that determine it. 00:10:58.437 --> 00:11:02.577 Determine the value and it changes over time. 00:11:02.857 --> 00:11:06.417 So the environment itself is a model of its own. 00:11:06.597 --> 00:11:12.417 Then in each one of these voxels, there is a cell or a sphere that senses what 00:11:12.417 --> 00:11:18.757 are the chemicals in the atmosphere or in the surrounding environment and act based on that. 00:11:18.917 --> 00:11:25.977 This concept is inspired by... it's a rather old concept that didn't get a lot 00:11:25.977 --> 00:11:28.497 of attention in science. It's called autonomous agent. 00:11:29.037 --> 00:11:33.997 Autonomous agent is taking out of artificial intelligence. During the 60s, 00:11:34.337 --> 00:11:41.717 they defined this concept as a machine, an object, or agent that interact with this, 00:11:42.317 --> 00:11:48.157 environment and decide on its next move based on its current state and the environment. 00:11:48.457 --> 00:11:52.697 So you consider the whole model to be an agent-based model? It's agent-based 00:11:52.697 --> 00:11:57.257 and it's autonomous agent-based model as adapted to biology. 00:11:57.437 --> 00:12:03.537 This This is like the computer science view of the models, yeah. 00:12:03.857 --> 00:12:09.657 So the first model you described was to replicate the pancreas. 00:12:10.637 --> 00:12:14.197 And historically, that was your first model? Yeah, that was the first model. 00:12:14.357 --> 00:12:18.797 And what was the specific sort of scientific question that led you and the group 00:12:18.797 --> 00:12:22.057 you were in to think that the pancreas was a good place to go for this? 00:12:22.317 --> 00:12:24.697 Well, I wanted to study about the diabetes. 00:12:25.257 --> 00:12:29.117 And I started research about the diabetes. and then I got to the organ that 00:12:29.117 --> 00:12:30.797 is controlling the diabetes. 00:12:31.257 --> 00:12:38.157 And then my view was that you can't study the organ without knowing the process 00:12:38.157 --> 00:12:42.017 that brought him to this state, 00:12:42.157 --> 00:12:46.577 and you need to understand the structure and how it was formed. 00:12:46.717 --> 00:12:52.977 So you need to start by the first day where the organ develops in order to understand 00:12:52.977 --> 00:12:54.497 what went wrong by the end. 00:12:54.497 --> 00:12:59.937 And if you notice during the lecture, I haven't even talked about the diabetes 00:12:59.937 --> 00:13:02.557 because we didn't get this far. 00:13:02.957 --> 00:13:09.937 There were a lot of open questions, a lot of interesting questions that just emerge as we go. 00:13:10.057 --> 00:13:13.877 But I envision that models like that will cover the whole process, 00:13:13.957 --> 00:13:15.837 including the functionality in the end. 00:13:15.837 --> 00:13:21.877 And then you can study diseases and everything and you can track all the way 00:13:21.877 --> 00:13:26.597 back to the history and see if something went wrong on day zero. 00:13:27.477 --> 00:13:31.037 And affected what we see on day 100. 00:13:31.797 --> 00:13:36.597 And I guess many of the diseases that you're interested in have a genetic component, 00:13:36.817 --> 00:13:40.737 which perhaps you'll then be able to recreate in the model once the model is 00:13:40.737 --> 00:13:41.497 sufficiently complete. 00:13:41.817 --> 00:13:44.057 Yes, but it's not only the genetic background. 00:13:45.057 --> 00:13:49.017 It's like the thing that the organ experienced. 00:13:50.489 --> 00:13:56.069 And cause this disease. How far does the model have to develop before you'll 00:13:56.069 --> 00:13:59.609 be able to ask these kinds of questions with it? It still has like... 00:14:00.269 --> 00:14:07.649 Right now it covers all the developmental birth. This is 15 days. 00:14:08.509 --> 00:14:16.709 And we need to add expression of the factors, of the hormones to that, 00:14:16.869 --> 00:14:21.069 and then we can investigate it. I believe in a few years' time. 00:14:21.309 --> 00:14:25.909 So you're close to having something that would be useful in a disease model. 00:14:26.249 --> 00:14:33.489 Yeah, but we are close to having kind of a complete model, but it doesn't necessarily 00:14:33.489 --> 00:14:35.329 mean that we know everything we need. 00:14:35.509 --> 00:14:40.989 Because once you finish modeling the organ and added all the functionality, 00:14:41.389 --> 00:14:44.529 you still need to verify that it is consistent with biology. 00:14:44.529 --> 00:14:50.369 So at each stage when you're developing the model, you're using some data from 00:14:50.369 --> 00:14:52.869 the biology to test the functionality. 00:14:53.509 --> 00:14:57.329 For instance, I think you were showing with the pancreas, you were showing what 00:14:57.329 --> 00:15:02.909 happens with the interaction with blood vessels and how the model produced similar 00:15:02.909 --> 00:15:06.109 outputs to what people observed in experiments. 00:15:06.689 --> 00:15:13.309 And that methodology whereby you take these biological data and use them to 00:15:13.309 --> 00:15:15.389 test the validity of the model. 00:15:16.009 --> 00:15:22.669 Do you consider that the methodology that you have there is finalized, fully refined? 00:15:24.009 --> 00:15:30.049 How do you know when you've done enough tests? Or is it open-ended, I guess? 00:15:30.349 --> 00:15:35.129 Yeah, you never have enough tests because the data keeps streaming into your desktop. 00:15:35.369 --> 00:15:39.729 So there are more papers published and there is new data and you have to refine 00:15:39.729 --> 00:15:41.449 the model as much as you need. 00:15:41.669 --> 00:15:45.969 Okay, I guess there's another way of putting that. What's the minimal set of 00:15:45.969 --> 00:15:52.529 biological constraints that you feel you need in order to say that the model is useful? 00:15:52.789 --> 00:15:59.689 If you manage to reproduce the biology and a few mutations, so it's not enough 00:15:59.689 --> 00:16:01.689 to have the biology reproduced. 00:16:01.689 --> 00:16:09.049 You still need to show that you agree with the known data. And then you can use it for prediction. 00:16:09.369 --> 00:16:15.709 I guess I'm thinking sort of when people are building a machine learning system 00:16:15.709 --> 00:16:17.129 for instance, something that. 00:16:18.837 --> 00:16:23.817 Learn to recognize spoken language and create written language, 00:16:24.717 --> 00:16:29.577 then they might take a data set and use half of it for training, half of it for testing. 00:16:29.917 --> 00:16:35.917 Do you use those kind of methodologies of deliberately leaving out some of the data? 00:16:36.217 --> 00:16:40.737 No, I would include all the data and then start testing. 00:16:41.577 --> 00:16:46.177 But is there a risk then that your model isn't sufficiently... 00:16:46.177 --> 00:16:52.357 Well, you can build the model to fit those data, but then you don't have enough 00:16:52.357 --> 00:16:54.677 test points to really check whether the model is... 00:16:56.237 --> 00:17:00.417 I believe that you should get all the available data into your model, 00:17:00.617 --> 00:17:08.217 then start taking the next step and see if you can get more data once you have the model. 00:17:08.737 --> 00:17:14.237 Would that be because you need what data there is to constrain the model and 00:17:14.237 --> 00:17:17.297 build it? I think that the difference is that I'm working with biology. 00:17:18.097 --> 00:17:22.317 Leaving out some of the biological data makes no sense. 00:17:22.497 --> 00:17:26.237 If you know an evidence on the system, you must agree with it. 00:17:26.377 --> 00:17:34.417 What can be done is that experiments that were done in the past and you are 00:17:34.417 --> 00:17:41.517 not aware of or you prefer to test to keep them as a test case, that can be done. 00:17:41.517 --> 00:17:48.917 But for the wild-type simulation, you can't leave data out. 00:17:49.297 --> 00:17:53.037 So the data set wouldn't be sufficiently rich for this kind of, 00:17:53.057 --> 00:17:56.457 to have two sets of data, one which you used to build the model, 00:17:56.517 --> 00:17:57.837 one which you could use to tell. 00:17:57.997 --> 00:18:04.677 In analogy, we can say that you take the wild-type data, the type of the normal 00:18:04.677 --> 00:18:09.017 growth, normal development, and you test the mutation. 00:18:09.017 --> 00:18:11.777 So that's a separation you can do 00:18:11.777 --> 00:18:14.617 that's what you're doing with the models that's what I'm doing with the models 00:18:14.617 --> 00:18:20.257 but you can take the existing data and say roughly this I keep out and this 00:18:20.257 --> 00:18:25.857 I keep in you have to keep something in mind before you do it and I think this 00:18:25.857 --> 00:18:31.577 separation of mutation and real data can be can be equivalent to what you described, 00:18:32.437 --> 00:18:36.557 so to come back a bit to the cell model that we're using. 00:18:37.897 --> 00:18:40.877 And naively you might think about that you might model a 00:18:40.877 --> 00:18:43.517 cell as let's say an autonomous entity with a 00:18:43.517 --> 00:18:48.857 membrane that sort of is is is motile it moves through some substrate it might 00:18:48.857 --> 00:18:53.217 have interactions with other cells adhere to it and so on but in the simulation 00:18:53.217 --> 00:18:58.697 that's not really what a cell is right in the simulation the cells in the end 00:18:58.697 --> 00:19:02.517 defined operationally as the state of. 00:19:03.849 --> 00:19:08.809 A little volume of space, right? So you take the whole volume that you want 00:19:08.809 --> 00:19:15.329 to simulate and you sort of cut it up in a lot of little cubes and then what 00:19:15.329 --> 00:19:17.869 you call a cell is the state of each of those cubes. 00:19:18.409 --> 00:19:22.269 Well, it depends how you... No, no, I don't like this way of looking. 00:19:22.449 --> 00:19:26.669 I disagree with it because there are a lot of empty cubes. 00:19:27.209 --> 00:19:30.789 What type of cell is that? Is that the null cell? 00:19:31.669 --> 00:19:35.429 You can tell me, right? No, I don't believe in the null cell. 00:19:35.589 --> 00:19:37.529 Cell is an existing entity. 00:19:37.769 --> 00:19:45.349 A null cell or a null object can be a mathematically defined object, 00:19:45.549 --> 00:19:47.849 but it can't be a biological defined object. 00:19:47.909 --> 00:19:53.529 Sure, but in terms of the caricature I gave of the technical approach, 00:19:53.669 --> 00:19:57.049 the implementation is, I think, reasonably accurate. 00:19:57.429 --> 00:20:00.729 Yes. In terms of implementation. The implementation, you can look at it that 00:20:00.729 --> 00:20:08.049 way, but you'll have to add the temporal component into it. Of course. 00:20:08.309 --> 00:20:13.009 So it's kind of, you can think of it as a kind of a three-dimensional space 00:20:13.009 --> 00:20:19.209 or two-dimensional space divided into cubes or pixels and it changes over time. 00:20:19.929 --> 00:20:22.289 So it's like a three-dimensional cellular automaton. 00:20:23.189 --> 00:20:27.629 Yes. You can think about it as a collection of three-dimensional cellular automaton. 00:20:27.789 --> 00:20:33.609 Right. But we have chosen to separate between the environment and the cell. 00:20:34.129 --> 00:20:36.949 And this is because this is how it acts in biology. 00:20:37.569 --> 00:20:43.209 So it's hard to me to, okay, it's like saying you're not a human being. 00:20:43.329 --> 00:20:48.829 You are kind of a pixel and you are just moving from one place to another. 00:20:49.009 --> 00:20:50.169 But no, that's not the truth. 00:20:50.449 --> 00:20:54.709 That's not the case. If you're doing something, if we are talking or something, 00:20:54.789 --> 00:20:58.029 there is something behind this action. 00:20:58.229 --> 00:21:02.349 We're not just talking, we're having conversation. You say something, I reply. 00:21:03.749 --> 00:21:09.709 So to think about it just as movement and things in the space, 00:21:09.929 --> 00:21:13.349 I think it will be simplification of reality. 00:21:13.869 --> 00:21:19.829 But yes, if you prefer to see the technology challenge or the implementation. 00:21:19.829 --> 00:21:21.869 Just to define what we're doing, right? 00:21:21.909 --> 00:21:25.289 Because we're interpreting the outcome of an algorithm, right? 00:21:25.349 --> 00:21:31.089 And the algorithm is just, if you want, deciding what state a certain voxel 00:21:31.089 --> 00:21:36.989 in that simulation should take, a certain little bit of this three-dimensional volume. 00:21:37.649 --> 00:21:42.369 And then with that, you model these systems, these organs, or C. 00:21:42.449 --> 00:21:44.909 Elegans in the brain. These are the three examples we looked at. 00:21:45.549 --> 00:21:49.769 So then the question becomes, okay, when is then such a simulation, 00:21:49.989 --> 00:21:53.169 let's say, defendable? When is it plausible? 00:21:53.769 --> 00:21:57.209 When is it not more our interpretation? 00:21:58.673 --> 00:22:01.373 As opposed to what's really going on in the simulation, right? 00:22:01.433 --> 00:22:07.033 So for the pancreas, you showed us that we have to look at a system that essentially 00:22:07.033 --> 00:22:11.333 at the stage in development, your modeling expresses three cell types, right? 00:22:11.393 --> 00:22:15.213 And these three cell types cluster in very specific forms, no? 00:22:15.293 --> 00:22:18.693 And then what you want to recover in your simulation is, again, 00:22:18.733 --> 00:22:21.413 the generation of these three cell types and this kind of clustering. 00:22:21.533 --> 00:22:22.893 Is that correct? That's correct. 00:22:23.073 --> 00:22:28.633 Okay. And then I could argue, well, and then the structuring of this pancreas 00:22:28.633 --> 00:22:33.413 was also very much predicated of the role as you told us on the role of the 00:22:33.413 --> 00:22:38.193 blood vessels right so the blood vessels providing some sort of scaffold, 00:22:38.793 --> 00:22:43.933 in which that very much guides the coagulation of the cells yeah, 00:22:44.793 --> 00:22:49.213 so but then but this is a well-known biological fact it's not something that 00:22:49.213 --> 00:22:54.413 we sure but you could still argue the biologists were not really able to tell 00:22:54.413 --> 00:22:57.353 you yet how this then happened in all its details, right? 00:22:57.433 --> 00:23:01.153 How sort of, let's say, specific parameters might affect this. 00:23:01.433 --> 00:23:04.613 But then you could, as we saw earlier, right? Your cells, as you said, 00:23:04.733 --> 00:23:07.073 have dozens, hundreds of states. 00:23:09.753 --> 00:23:13.853 Large populations of them give rise to specific structures. 00:23:14.093 --> 00:23:18.033 But what we try to explain are the three cell types of the pancreas forming, 00:23:18.233 --> 00:23:23.133 okay, like cauliflower-type structures. So, I could say, well, 00:23:23.293 --> 00:23:25.193 this is a super powerful model. 00:23:25.293 --> 00:23:29.493 You have so many parameters you can play with, this can never go wrong. 00:23:29.713 --> 00:23:35.633 You should always hit the jackpot with that, right? Yes, but all the parameters 00:23:35.633 --> 00:23:38.473 in the model are constrained by biology. 00:23:38.993 --> 00:23:42.733 So I believe that what is in the model, it's in the biology. 00:23:42.953 --> 00:23:45.193 So maybe the biology is a superpower model. 00:23:45.513 --> 00:23:50.653 You can think about it. It's right because you can find the pathways in cellular 00:23:50.653 --> 00:23:52.613 decision that you can bypass. 00:23:52.893 --> 00:23:57.113 You have a few pathways that leads to the same result. 00:23:57.413 --> 00:24:01.453 So let's ignore the model and think about the biology. 00:24:02.113 --> 00:24:05.873 Why the biology of two paths that goes to the same destination. 00:24:06.193 --> 00:24:12.793 So another way of validating the model is to create variants of it which don't 00:24:12.793 --> 00:24:13.733 fit the biological data. 00:24:13.893 --> 00:24:18.613 So just take the one which you think is accurate and flip a few of the arrows 00:24:18.613 --> 00:24:23.673 or whatever and see if they do anything similar. And that would tell you how. 00:24:24.860 --> 00:24:31.560 How valid it is to say that that specific model is needed to generate the data that you see. 00:24:31.680 --> 00:24:34.180 Now, in some sense, that's what you're doing when you get your mutations, 00:24:34.300 --> 00:24:38.480 but then you're saying this is an actual mutation that happens in nature, 00:24:38.620 --> 00:24:40.000 and I can see what happens in the model. 00:24:40.140 --> 00:24:45.020 But have you done experiments where you just flip components or leave components 00:24:45.020 --> 00:24:47.120 out and see how that works? Yes, of course. 00:24:48.620 --> 00:24:55.040 One of the first in-silico experiments we did was just to take the nucleus, 00:24:55.180 --> 00:25:01.280 the inner decision of the cell, and just make a big spaghetti, 00:25:01.480 --> 00:25:08.320 like Erdos way of testing, taking the arrows and connect them randomly to other 00:25:08.320 --> 00:25:12.740 states and to see what's going on, like the craziest stuff. 00:25:12.860 --> 00:25:22.620 It's just like taking the DNA and just take the operon and point it to express 00:25:22.620 --> 00:25:26.160 a different gene, just like a crazy scientist did. 00:25:26.400 --> 00:25:32.020 And yes, this disabled everything in the simulation, but at least it's proven 00:25:32.020 --> 00:25:36.000 that that data is really necessary for the modeling. 00:25:36.780 --> 00:25:42.240 But what I'm trying to say, and maybe I'm not clear enough, we are trying that 00:25:42.240 --> 00:25:48.040 the models will be realistic as possible and to take off a pathway or take off 00:25:48.040 --> 00:25:50.600 genes just because we don't understand what they do, 00:25:50.740 --> 00:25:55.100 or there are other genes that do the same, that does the same. 00:25:55.420 --> 00:25:59.380 I wouldn't recommend that, because then you might miss point. 00:26:00.726 --> 00:26:02.766 In the model. The idea is to put 00:26:02.766 --> 00:26:07.066 into the model as many biological data as possible, not the minimal one. 00:26:07.286 --> 00:26:10.566 Well, there are two issues for me on this, right? On the one hand, 00:26:10.646 --> 00:26:14.646 how do you really quantify a cauliflower, right? 00:26:14.726 --> 00:26:19.746 So in some sense, in the presentation, it was a bit more sort of by eye, 00:26:19.866 --> 00:26:22.086 like, well, so it looks sort of similar, right? 00:26:22.686 --> 00:26:26.646 And I think we would all agree that we can do better than that, right? 00:26:26.726 --> 00:26:29.366 And also on top of that, that cauliflowers have fractal-like structures, 00:26:29.526 --> 00:26:33.966 which raises interesting questions about development, right, and morphogenesis. 00:26:34.886 --> 00:26:40.726 So how then do you really quantify that similarity between the structure that 00:26:40.726 --> 00:26:45.506 you see in your simulation and that you find in biology? It's not easy. 00:26:46.106 --> 00:26:52.246 Life is hard. Life is hard, yeah. And one of the hurdles in this issue is that 00:26:52.246 --> 00:26:56.266 there are not that many data, biological data. 00:26:56.266 --> 00:27:00.546 It's not that I can ask you, ask a biologist, please give me how many cells 00:27:00.546 --> 00:27:03.586 there are in this cauliflower and what are the fractals. 00:27:03.906 --> 00:27:07.806 This will be a very naive way to see the experiments. 00:27:08.346 --> 00:27:11.526 Usually, that's what we get. That's the data. No, but still, 00:27:11.606 --> 00:27:14.566 in the literature, you will find these characterized. 00:27:14.826 --> 00:27:19.886 There will be images. There will be microscopic analysis, electromicroscopy. 00:27:20.166 --> 00:27:23.546 There will be different levels of analysis. You will hardly find it in mice. 00:27:23.546 --> 00:27:27.486 You will find it in, I think, I didn't come across too many. 00:27:27.666 --> 00:27:30.566 Or you might have the inter-cauliflower spacing. 00:27:31.506 --> 00:27:37.086 Yes. What I did, for example, is doing some of image analysis, 00:27:37.226 --> 00:27:44.626 trying to compare not exactly the population, but to split between the color 00:27:44.626 --> 00:27:47.826 of the domains of the color. 00:27:47.826 --> 00:27:54.806 Or how many pixels carry red color and how many green colored pixels there are 00:27:54.806 --> 00:27:59.206 in histology picture and to compare it with a cell count in the simulation. 00:27:59.826 --> 00:28:04.586 So it's kind of making... Okay, but you're saying you couldn't really do that 00:28:04.586 --> 00:28:06.986 yet because the data wasn't there. The data is not there. 00:28:07.686 --> 00:28:13.246 The data is very dirty in this... Okay, but which parameter in your model would, 00:28:13.346 --> 00:28:15.606 for instance, control the inter-cauliflower spacing? 00:28:16.948 --> 00:28:21.848 There is no single parameter. There are a few parameters and all related to the biology. 00:28:21.968 --> 00:28:34.968 For example, the flow of the gradient of the blood vessels secreting the factors in the environment. 00:28:35.228 --> 00:28:40.728 That would be a very important parameter. And indeed, when we played with this 00:28:40.728 --> 00:28:49.628 parameter, we see that we generated a lot of different structures that may and may not be correct. 00:28:49.788 --> 00:28:53.888 But this is something to be tested in the future when we can control the blood 00:28:53.888 --> 00:28:59.408 vessel, the way the blood vessel secretes factors to the environment. Right. 00:28:59.828 --> 00:29:05.788 So then after the pancreas, we looked at the C. elegans, right? 00:29:05.868 --> 00:29:12.708 And there in particular, you looked at the hermaphrodites who are basically producing other C. 00:29:12.728 --> 00:29:15.208 Elegans by cell fertilization. 00:29:16.388 --> 00:29:20.688 And that C. elegans, the way you described it, just works a bit like a sausage machine, right? 00:29:20.748 --> 00:29:25.808 So you have sort of stem cells generating the cells that move through the body. 00:29:27.028 --> 00:29:31.928 They then sort of get transformed into eggs on a new stage, and then they pass 00:29:31.928 --> 00:29:35.888 through a sperm phase, or they got merged with sperm, and then there you have… 00:29:35.888 --> 00:29:38.168 They get fertilized. Then you have your fertilized egg, right? 00:29:38.448 --> 00:29:43.528 But it is a big part of the C. elegans, this reproductive organ. 00:29:43.788 --> 00:29:49.148 Sure. But it's not all of it. It still have a neuron system and locomotive system. 00:29:49.588 --> 00:29:53.928 But you sort of, you modeled this sausage machine that sort of is spitting out, 00:29:53.948 --> 00:29:56.448 if you want, these new C. elegans, right? Yes, this is the granite. 00:29:56.488 --> 00:30:01.768 This is the productive organ of the C. elegans. And so which elements of the 00:30:01.768 --> 00:30:07.228 assimilation of the pancreas could you carry over to this C. elegans system? 00:30:08.099 --> 00:30:12.119 By means of platform, I used the same platform. Principles were. 00:30:12.319 --> 00:30:14.619 And the same principle was implemented. 00:30:15.099 --> 00:30:22.619 Which is? And the way we construct the cell and the cell, different components, 00:30:22.819 --> 00:30:28.799 the nucleus and the membrane and all the autonomous agent concept we discussed 00:30:28.799 --> 00:30:32.759 earlier was just adopted to this model. 00:30:33.039 --> 00:30:35.259 We had to change the name of the genes. 00:30:35.879 --> 00:30:40.259 But this is basically it. and the connections between the states. Or the blocks. 00:30:40.459 --> 00:30:44.859 Yeah, and the way they affect each other. So if we think about it as a graph, 00:30:45.039 --> 00:30:50.339 we just had a different graph between the nodes and the neurons. 00:30:50.359 --> 00:30:53.859 But I would assume that those are the number of states of the cells who have changed. 00:30:54.339 --> 00:31:00.099 Yes, yes. If in the pancreas we had like 120 states in the cell against, 00:31:00.239 --> 00:31:04.539 we had like 20, which is like five times less. 00:31:05.479 --> 00:31:11.419 So, in the environment in which the C. elegans is growing, what is that environment? 00:31:11.599 --> 00:31:15.639 Because the pancreas is growing in a very different kind of environment from 00:31:15.639 --> 00:31:24.779 a... Yes, for example, in the pancreas model, we had a very complicated network of blood vessels. 00:31:25.019 --> 00:31:30.439 In the C. elegans, it's very easy. At the tip cell, there is one cell that secretes 00:31:30.439 --> 00:31:32.699 a factor to the environment. 00:31:32.899 --> 00:31:37.739 That's it. So it was much easier to get a more accurate description of it. 00:31:38.039 --> 00:31:48.179 And it was even modeled as a using order and order differential equation at the stationary phase. 00:31:48.559 --> 00:31:53.339 So it's just like a chemical that is being secreted in a gradient. 00:31:53.699 --> 00:31:55.719 It is much easier to capture. 00:31:56.119 --> 00:32:00.239 So it's like one cell of the blood vessels of the pancreas model. 00:32:00.919 --> 00:32:09.899 It's much easier. And this This is why it was much easier to generate good prediction 00:32:09.899 --> 00:32:13.439 or testable predictions while in the pancreas. 00:32:13.439 --> 00:32:18.379 Every prediction I had had to wait like five years until the biologists find 00:32:18.379 --> 00:32:20.159 the right technique to test it. 00:32:20.419 --> 00:32:22.779 In C. elegans it was tested within weeks. 00:32:23.179 --> 00:32:29.159 But in the pancreas, I think the benchmark was very much the morphology of the structure. 00:32:29.319 --> 00:32:33.219 That's correct, right? So what would be that benchmark for C elegance? 00:32:34.251 --> 00:32:41.731 I think it's, okay, it's more accuracy because most of the, if in the pancreas 00:32:41.731 --> 00:32:47.231 model, I had to use a lot of hand waving and say, okay, look, this is similar. 00:32:47.351 --> 00:32:54.371 In the C. elegans model, I can talk exactly about the lengths of areas and the 00:32:54.371 --> 00:32:57.911 lengths of zone and number of cells and cell cycle. 00:32:57.911 --> 00:33:02.491 And indeed, out of all this refinement and fine-tuning of the model, 00:33:03.011 --> 00:33:09.511 it turns out, for example, even in the small details, what is the different 00:33:09.511 --> 00:33:15.351 cell cycle between the adult and the young C. 00:33:16.131 --> 00:33:23.611 Elegans? During development, what is the ratio between the cell cycle at the two faces? 00:33:23.831 --> 00:33:30.251 And we suggested in the model 1 to 5. And then biologists think it's 1 to 3.5. 00:33:31.051 --> 00:33:34.811 So it's the same range. And we didn't take it into account. 00:33:35.111 --> 00:33:42.151 We just try to make it as realistic as possible and to collaborate it with time. 00:33:42.371 --> 00:33:48.091 And this all tiny detail just emerged. So that would mean for C. 00:33:48.111 --> 00:33:54.131 Elegans, the benchmark is more like what kind of cells do you have in C. 00:33:54.191 --> 00:33:56.911 Elegans and what position at what point in time? 00:33:58.151 --> 00:34:02.831 And the number of cells. Sure, yeah, exactly. So you can start talking about 00:34:02.831 --> 00:34:08.631 the numbers and quantity aspects while in the pancreas it will just look, 00:34:08.671 --> 00:34:09.551 it looks like a cauliflower. 00:34:10.071 --> 00:34:16.471 It's similar. But in the first C. elegans, we talk in the end about three or 00:34:16.471 --> 00:34:22.111 four cell types, the stages, right? So why is that a hard problem? 00:34:23.367 --> 00:34:29.227 Why this is a hard problem? Because it's not all about the cell type. 00:34:30.487 --> 00:34:33.787 What controls them? What is the network? This is harder. 00:34:34.027 --> 00:34:39.147 How the external signal affects the development. That's a hard problem. 00:34:39.747 --> 00:34:43.847 It's hard to know what happens there. And not all is known. 00:34:44.647 --> 00:34:52.827 Many, many issues there are open. For example, what causes cell death? 00:34:53.367 --> 00:34:55.587 Is it autonomous? Is it inner decision? 00:34:56.147 --> 00:34:58.947 Is there an external signal that kills them? 00:34:59.407 --> 00:35:04.187 We know that 30% of the cell do not make it to death. 00:35:05.607 --> 00:35:08.207 30%, it's a lot of them. It's like third. 00:35:10.167 --> 00:35:14.687 So we know that they don't make it. What happens when they go to apoptosis? 00:35:14.907 --> 00:35:18.787 You mentioned this also in your talk, that this was one of the insights that 00:35:18.787 --> 00:35:25.307 you had. But apoptosis of cells, so the cell death, is an active process, 00:35:25.447 --> 00:35:27.227 right? This is really a regulated process. 00:35:27.427 --> 00:35:33.547 There are receptors sitting at the outside of cells that are directly driving apoptosis. 00:35:33.927 --> 00:35:39.867 So it's not like, well, in your simulation, you were telling us you were inducing cell death. 00:35:39.987 --> 00:35:44.347 But you're saying, okay, if cells pass through this zone, there's a 30% chance they won't come out. 00:35:44.547 --> 00:35:49.407 That's because apoptosis is very active in the adult. Thank you for watching! 00:35:50.446 --> 00:35:53.586 In the developing organ, it's less active. 00:35:54.006 --> 00:36:00.346 Usually, you know, how many cells a human, an adult is losing in his lifetime, 00:36:00.606 --> 00:36:02.666 the embryo keep growing. 00:36:03.606 --> 00:36:07.366 Usually, developing systems has less apoptosis than the adult. 00:36:07.586 --> 00:36:09.646 Well, they better, or else it won't develop, right? 00:36:10.646 --> 00:36:15.186 No, what I'm trying to get to here is that in C. 00:36:15.266 --> 00:36:19.046 Elegans, we would assume that also their apoptosis is actively regulated. 00:36:19.046 --> 00:36:23.526 For instance, depending on environmental factors, apoptosis might be more or 00:36:23.526 --> 00:36:26.126 less, right? Depending on, let's say, the stress on the animal. 00:36:26.986 --> 00:36:29.286 Well, in those elements, you sort 00:36:29.286 --> 00:36:34.246 of, you are not really including in your simulation. No, we left it out. 00:36:34.406 --> 00:36:37.086 We just flipped a coin at that stage. Exactly. 00:36:37.386 --> 00:36:43.326 This is an extension to the model. I'm not sure there is enough data about apoptosis 00:36:43.326 --> 00:36:44.746 and C. elegans at this stage. 00:36:44.886 --> 00:36:47.806 I'm not sure about it. I don't want to commit. it but if there 00:36:47.806 --> 00:36:51.566 is it can be formulated and added to the model okay i'm 00:36:51.566 --> 00:36:55.026 still trying to understand how you build and tune these models so um in 00:36:55.026 --> 00:36:57.986 the case of c elegans you have what you described a 00:36:57.986 --> 00:37:00.806 very simple environment with a chemical gradient and then 00:37:00.806 --> 00:37:03.526 you have your uh cell model for your 00:37:03.526 --> 00:37:06.866 initial stem cell which uh presumably 00:37:06.866 --> 00:37:09.806 when you set it up has some of the relevant details from 00:37:09.806 --> 00:37:12.546 the biology but not all and then 00:37:12.546 --> 00:37:15.446 i imagine you drive the model you see what happens 00:37:15.446 --> 00:37:18.586 and you see well it's not building a worm for me yet and 00:37:18.586 --> 00:37:21.406 then you say okay what else is what's wrong with this what 00:37:21.406 --> 00:37:26.366 can i do different what you go back to the library and with much more many more 00:37:26.366 --> 00:37:32.366 uh so so it's a very iterative process where you you run models and all the 00:37:32.366 --> 00:37:36.126 time the models aren't working and gradually they get closer and closer to what 00:37:36.126 --> 00:37:39.526 you would like to see yes okay and at that point you You say, 00:37:39.566 --> 00:37:46.886 well, I think I've got a reasonably good working model of the wild-type worm. 00:37:48.646 --> 00:37:51.626 Then you go to a different part of the library and say, let's look at mutants 00:37:51.626 --> 00:37:55.846 and say, let's see what happens if we now flip the circuit. Do we get mutants? 00:37:56.946 --> 00:38:03.286 So the validation stage is quite a strong one, really, because I think you already 00:38:03.286 --> 00:38:09.806 have something that captures the development of the living natural animal. 00:38:09.946 --> 00:38:15.346 And then you can mutate it in ways that you'd be fairly confident match real 00:38:15.346 --> 00:38:20.126 mutations. Yeah, but that's a very important part of the modeling process. 00:38:20.466 --> 00:38:23.946 But once you complete it, this is the really interesting part. 00:38:24.326 --> 00:38:28.706 You start to play with it. You start to risk it. You can play the crazy scientist. 00:38:29.286 --> 00:38:33.406 You can do whatever you have in your mind. And then you can get predictions. 00:38:33.726 --> 00:38:35.406 And some of them are testable. 00:38:36.206 --> 00:38:39.026 And for some of them you can tell whether they are plausible 00:38:39.026 --> 00:38:41.766 and where whether they are testable some of 00:38:41.766 --> 00:38:48.186 them are just crazy enough too crazy to be tested so some of the predictions 00:38:48.186 --> 00:38:53.166 you make from the model turn out to be wrong yeah and then very often then you 00:38:53.166 --> 00:38:59.026 go to stage one go back to the library so so the the methodology doesn't change 00:38:59.026 --> 00:39:01.006 that much i guess you get your basic working model, 00:39:01.546 --> 00:39:04.766 You try it out with some mutants, some it works with, some it doesn't. 00:39:04.766 --> 00:39:08.226 You say, well, it didn't work with that mutant. I need to refine the model again. Yes. 00:39:08.526 --> 00:39:11.906 There are a version of models all the time. 00:39:12.106 --> 00:39:18.426 And this is a never-ending process because data flows in and as the data came 00:39:18.426 --> 00:39:22.246 to your bench, you just need to rethink all the model. 00:39:22.546 --> 00:39:25.006 Sometimes a small paper can change the whole model. 00:39:25.486 --> 00:39:29.886 What I find interesting is that apparently you live in a field of developmental 00:39:29.886 --> 00:39:34.126 biology where the literature is pretty clear because apparently you can just 00:39:34.126 --> 00:39:38.606 go to the library and pick out the papers and get the parameters for your simulation and reality. 00:39:39.977 --> 00:39:43.637 At least the field in which we exist, it's never like that, right? 00:39:43.717 --> 00:39:45.897 You find the papers, they're contradictory, they're incomplete, 00:39:46.257 --> 00:39:47.457 you have to meditate on this stuff. 00:39:47.657 --> 00:39:51.637 They have to talk to the specialist to actually understand what the hell they're writing about, right? 00:39:51.697 --> 00:39:55.697 So is it really fair to just say, okay, then I go to the library, 00:39:55.757 --> 00:39:57.137 get the papers, I know my parameters. 00:39:57.357 --> 00:40:00.257 I cannot believe that's how it works. That's not fair at all. 00:40:00.277 --> 00:40:03.817 So how does it really happen? You know, every three papers have four different options. 00:40:04.197 --> 00:40:09.577 Exactly right, yeah. But the modeling approach allows you to change between hypotheses. 00:40:09.977 --> 00:40:12.697 And try to check different cellular mechanisms. 00:40:13.117 --> 00:40:15.737 So you can play out scenarios, right? That's the whole point. 00:40:15.877 --> 00:40:18.357 That's the whole point of modeling in general, I think. 00:40:18.477 --> 00:40:26.497 But what the modeling, my modeling or the modeling I presented offer is kind 00:40:26.497 --> 00:40:33.257 of a nicer visuals way for developmental systems, 00:40:33.657 --> 00:40:37.417 in particular at the population level. 00:40:37.417 --> 00:40:40.557 It's not a single pathway that we are testing. 00:40:40.657 --> 00:40:43.297 It's not one cell. It's the whole population. 00:40:43.717 --> 00:40:47.357 And it's the whole population over time and space. 00:40:47.917 --> 00:40:52.477 So you can get a lot of new insight and ask a lot of new questions. 00:40:53.437 --> 00:40:56.377 So that means also the model can summarize a lot of data. 00:40:56.497 --> 00:40:59.777 It can help you to play out different scenarios to see, okay, 00:40:59.857 --> 00:41:03.877 what if all this missing information would have this characteristic? 00:41:04.937 --> 00:41:09.157 It could also highlight what you don't know. Because it summarizes it. Sure. 00:41:09.517 --> 00:41:13.557 But now, in the end, if you want to turn that model into a theory, 00:41:13.697 --> 00:41:16.877 you must explain something and you must make predictions, right? 00:41:17.057 --> 00:41:21.077 So at this stage, if we look at the model of C. elegans, what do you think have 00:41:21.077 --> 00:41:24.617 you really explained and what is a testable prediction of that? 00:41:24.917 --> 00:41:31.057 Yeah, we explained a lot of the interplay between cell cycle and differentiation. 00:41:31.937 --> 00:41:39.537 And one of the insights I presented is that if you change the ratio of cell 00:41:39.537 --> 00:41:41.517 cycle between the larva and the adult, 00:41:42.354 --> 00:41:49.314 then you get different behaviors. And one of them is that your stem cell population 00:41:49.314 --> 00:41:53.854 gets shorter or smaller once you reduce it. 00:41:54.014 --> 00:42:01.234 It seems a bit naive and it seems a bit straightforward, but the scientific 00:42:01.234 --> 00:42:08.314 community never thought of looking at the consequence of the cell cycle at the 00:42:08.314 --> 00:42:12.034 larva stage on the adult stage. 00:42:12.834 --> 00:42:16.134 So this is kind of question that we can easily answer 00:42:16.134 --> 00:42:19.174 or ask what happens if you change something 00:42:19.174 --> 00:42:22.634 at day zero what happened in day five right and it's 00:42:22.634 --> 00:42:29.674 not that easy to to test in the in the lab because the you change something 00:42:29.674 --> 00:42:34.574 and you need to wait a few days it's better to have kind of what where to look 00:42:34.574 --> 00:42:40.654 at not not what the answer is but where where to look and search and research. 00:42:41.514 --> 00:42:44.474 And in the system right and what that 00:42:44.474 --> 00:42:47.454 may give us an answer about it so then 00:42:47.454 --> 00:42:50.354 the last example you you analyzed and presented was 00:42:50.354 --> 00:42:55.054 neurogenesis right the development of the nervous system and again you roughly 00:42:55.054 --> 00:42:59.994 took the same modeling framework the same cell model but now again changing 00:42:59.994 --> 00:43:02.014 again the states changing the 00:43:02.014 --> 00:43:07.554 the expression path of the control pathways ways to build a piece of brain, 00:43:08.294 --> 00:43:11.134 it's not a piece of brain it's a piece of the nerve. 00:43:12.194 --> 00:43:19.174 System because what you see there what will be later the nerves so what did 00:43:19.174 --> 00:43:21.774 you exactly learn from that exercise what did you find. 00:43:22.684 --> 00:43:27.084 Before we go there, in terms of actually building that model, 00:43:27.244 --> 00:43:34.884 from the previous two models, were you able to take little networks of the blocks 00:43:34.884 --> 00:43:36.764 from within those models and use them again? 00:43:36.764 --> 00:43:43.844 Because we're interested this week in how evolution has reused some basic processes, 00:43:44.024 --> 00:43:48.684 perhaps in very different ways, sort of regulatory dream networks that might 00:43:48.684 --> 00:43:50.544 be involved in body patterning and inversion, 00:43:50.744 --> 00:43:53.044 but reused again in building bits of brain. 00:43:53.404 --> 00:43:58.484 And I wonder if there are examples of that in your work that you can say there 00:43:58.484 --> 00:44:04.844 was a network in one part of my cell and I was just able to use it in the pancreas, in the C. 00:44:04.844 --> 00:44:11.284 Elegans and in the mouse brain. First, I'd like to give you some statistics that may explain it. 00:44:11.384 --> 00:44:15.004 The first model, the pancreas, it took five years to develop. 00:44:15.544 --> 00:44:20.864 The second model, the C. elegans, it took two years. 00:44:21.464 --> 00:44:27.144 And the model of the brain, of the neuronal migration, took three months. 00:44:27.544 --> 00:44:32.204 So we short the time of developing the model itself. off. 00:44:32.304 --> 00:44:39.084 The analysis keeps taking a long time and looking for the prediction is a long period, 00:44:39.324 --> 00:44:46.224 but to develop, since we have a good view of what the modeling framework and 00:44:46.224 --> 00:44:50.804 the approach is, it takes less time to develop the model. 00:44:51.804 --> 00:44:58.224 So the basic concept and the basic approach, as time goes, as the research continues, 00:44:58.564 --> 00:45:01.384 it's much easier to implement it. 00:45:01.880 --> 00:45:07.180 And this concept of autonomous agent and the concept of having this sensor unit 00:45:07.180 --> 00:45:12.680 and the internal switch unit and the differentiation and proliferation are being 00:45:12.680 --> 00:45:16.980 controlled by these two components seems to work. 00:45:17.220 --> 00:45:21.480 So this seems like a general concept in developmental biology. 00:45:21.820 --> 00:45:25.060 I'm very, very careful here because I don't want to claim, hey, 00:45:25.260 --> 00:45:28.720 I have the holy grail, I'm having the way to model. 00:45:28.920 --> 00:45:32.220 But it seems very beneficial. we have we are 00:45:32.220 --> 00:45:35.200 encouraged by the results we don't it 00:45:35.200 --> 00:45:37.940 doesn't mean we hit the jackpot you know it's kind 00:45:37.940 --> 00:45:41.240 of we still have to we still have to to do 00:45:41.240 --> 00:45:44.680 a lot but what was the benchmark what was the benchmark in this case what did 00:45:44.680 --> 00:45:49.960 you replicate and how accurate was that replication and in the neuron we replicate 00:45:49.960 --> 00:45:55.300 the way neurons migrate from their birthplace to the place where they are going 00:45:55.300 --> 00:45:59.840 to going to become part of the nerve system. 00:46:00.100 --> 00:46:04.240 So it's from the core of the brain to the surface. 00:46:05.140 --> 00:46:08.320 It is guided by fibers, by the gallial fibers. 00:46:08.820 --> 00:46:14.960 And at the first stage, they use the gallial fiber to track the road. 00:46:15.100 --> 00:46:21.980 Then something happens and they stop following this gallium fiber, 00:46:22.520 --> 00:46:23.980 moving randomly in space. 00:46:23.980 --> 00:46:29.800 Then they reattach to the gallium fiber and just putting layer on top of the 00:46:29.800 --> 00:46:32.660 layer of neurons on the surface of the brain. 00:46:33.240 --> 00:46:36.940 But now, so in your sim, so what we're really talking about is also really the 00:46:36.940 --> 00:46:44.260 migration of cells along these guiding axes of glia or certain kinds of gradients 00:46:44.260 --> 00:46:46.100 that might attract them or repel them. 00:46:47.060 --> 00:46:49.340 So that means in that setup, 00:46:50.277 --> 00:46:54.017 collisions is an issue right that's cells if 00:46:54.017 --> 00:46:57.377 these migration patterns get disrupted cells might not 00:46:57.377 --> 00:47:02.397 be able to migrate across certain other cell population it's in all cases they 00:47:02.397 --> 00:47:07.557 do not they do not migrate one towards the other and one across the other they 00:47:07.557 --> 00:47:12.677 if they see that there is a cell in a neighboring in a neighboring pixel they 00:47:12.677 --> 00:47:15.357 won't move to it no way but in reality. 00:47:16.457 --> 00:47:23.277 Cells have to go through layers of cells that might have embedded themselves 00:47:23.277 --> 00:47:26.017 earlier, right? Because you always go to the outside. 00:47:26.417 --> 00:47:31.417 Yeah, but you need to think, okay, it's not exactly three-dimensional, 00:47:31.657 --> 00:47:34.657 not exactly two-dimensional. It's somewhere in the middle. 00:47:34.837 --> 00:47:38.417 So there is a layer of cells. They climb above it. 00:47:38.557 --> 00:47:44.017 They're creating another layer and they are walking on top of them and then 00:47:44.017 --> 00:47:48.277 they will place themselves on the same layer. So it's kind of two slices. 00:47:48.797 --> 00:47:55.797 So they go over and... What I was trying to get at, Zoltan Molniar was here 00:47:55.797 --> 00:47:58.457 early this week talking also about development of the brain. 00:47:58.977 --> 00:48:03.637 And a typical pattern there is that you have certain subpopulations moving from 00:48:03.637 --> 00:48:08.457 these neural plates at different points in time and also having to cross through. 00:48:08.797 --> 00:48:12.477 They cross through other populations. And they cross through other populations. 00:48:12.477 --> 00:48:15.897 And what I wanted to ask you about is that I would... 00:48:16.637 --> 00:48:21.037 My claim would be that with your simulation technique, you cannot handle that because. 00:48:22.109 --> 00:48:26.089 One location in space has just one specific state. 00:48:27.069 --> 00:48:33.149 So the dynamics of, let's say, migration and possible collision and obstruction 00:48:33.149 --> 00:48:36.669 and so on is not something that you can capture that way, or would you think 00:48:36.669 --> 00:48:38.009 that's too negative interpretation? 00:48:38.589 --> 00:48:44.189 I think you've just defined a research question. Okay. It's enough to be a PhD thesis. 00:48:44.669 --> 00:48:48.669 This is something to investigate. So I might have a chance to get my PhD done. 00:48:48.809 --> 00:48:51.989 That might work. You're accepted. Okay, great. 00:48:52.949 --> 00:48:57.369 But you still didn't fully answer my question, and maybe I don't know if it was a bad question. 00:48:57.569 --> 00:49:01.729 But you said that for the brain model, 00:49:01.849 --> 00:49:05.029 you keep the same structure of the previous two models for the cell. 00:49:05.149 --> 00:49:09.249 But what I was interested in, all these cells are doing similar behaviors of 00:49:09.249 --> 00:49:10.929 differentiating and migrating. 00:49:11.529 --> 00:49:18.729 Are there some bits of network within the decision-making parts of the cell that you could reuse? 00:49:18.729 --> 00:49:27.169 Yeah, the proliferation, the cell cycle is the same to all the models. 00:49:28.089 --> 00:49:32.349 There is the decision about asymmetric and symmetric proliferation, 00:49:32.429 --> 00:49:34.849 which is different between them. 00:49:34.949 --> 00:49:41.569 But the proliferation component is the same, and the differentiation component is very similar. 00:49:41.569 --> 00:49:47.649 So this is exciting in terms of when we think about evolution and we think, 00:49:47.669 --> 00:49:50.789 well, brain neurons are very different from C. 00:49:50.849 --> 00:49:54.969 Elegans neurons, but actually they share a lot of the same chemical machinery 00:49:54.969 --> 00:49:56.529 for doing what they do. Yeah. 00:49:57.317 --> 00:50:02.137 You have to change the specificities of what they're doing, but say networks can work internally. 00:50:02.397 --> 00:50:05.457 Yeah, I think this is the discussion that we had at the end of the talk, 00:50:05.577 --> 00:50:11.217 that you think that you're surprised that the biology is simple building blocks, 00:50:11.297 --> 00:50:13.097 and I see it as the reality. 00:50:13.637 --> 00:50:20.417 It's kind of, yeah, it's surprising that the biology carries simple building 00:50:20.417 --> 00:50:22.417 blocks, but I think this is the way it is. 00:50:22.597 --> 00:50:26.777 I don't think it is as complicated as we'd like to. Not wanting to agree with 00:50:26.777 --> 00:50:32.437 either of you, I could argue, wait, you guys are both deeply confused because 00:50:32.437 --> 00:50:35.397 we're scientists. Of course we're deeply confused. 00:50:36.557 --> 00:50:43.597 Right. But the point is that in your case, what carried over between these phenomena 00:50:43.597 --> 00:50:47.357 was a modeling strategy, an algorithm, if you want. 00:50:47.357 --> 00:50:53.117 But that technology as such doesn't tell you much about guiding principles because, 00:50:53.217 --> 00:50:58.877 as you said yourself, you don't have all these principles explicitly defined. 00:50:59.377 --> 00:51:05.137 And Tony is seeking, let's say, common principles across all these different things. 00:51:05.157 --> 00:51:10.977 Like the way you would grow an organ, a pancreas would include principles that 00:51:10.977 --> 00:51:13.117 you would carry over to having a C. 00:51:13.577 --> 00:51:17.077 Elegans sausage machine. But I think we said that there were some intrinsic 00:51:17.077 --> 00:51:23.177 cellular networks that you could copy whole chunks of the network across from 00:51:23.177 --> 00:51:25.157 these two different cell types. 00:51:25.237 --> 00:51:30.137 If you prefer to think of it as a mechanism of tears, that this is a motive 00:51:30.137 --> 00:51:33.877 of a mechanism of a cell, this is another way to look at it. 00:51:34.717 --> 00:51:38.977 Because, okay, but then we have to, if you gentlemen are so optimistic about 00:51:38.977 --> 00:51:41.937 this, you must declare from me. We're scientists, we must be optimistic. 00:51:41.937 --> 00:51:47.637 Right here and right now, what the common principles are between pancreas, 00:51:47.637 --> 00:51:49.437 C. elegans, and this piece of brain. 00:51:50.457 --> 00:51:55.137 So they both have proliferation, differentiation. They both have a kind of component 00:51:55.137 --> 00:52:01.817 of external sensor recognition or signaling, and they both have kind of internal 00:52:01.817 --> 00:52:03.737 inner cellular decision. 00:52:03.937 --> 00:52:08.157 And the decision of the mechanism must be defined by all the three components. 00:52:08.797 --> 00:52:13.217 It will be wrong to say that the environment and the extracellular signaling 00:52:13.217 --> 00:52:21.297 is not important, And it will be even more wrong to say that the internal genes is not effective. 00:52:21.677 --> 00:52:26.097 And it will be wrong to say that this has no proliferation or no differentiation. 00:52:26.517 --> 00:52:33.297 So this is necessary components for all the three elements, whatever it is, 00:52:33.337 --> 00:52:36.677 whether it is the pancreas in mice or the C. 00:52:36.757 --> 00:52:40.777 Elegans, germline development, or the neural development. Yeah, 00:52:40.777 --> 00:52:42.857 but the risk is, of course, that you say, look, E. 00:52:42.897 --> 00:52:47.357 Coli can have flagella and it can sort of flop around in some gradient and humans 00:52:47.357 --> 00:52:50.457 walk and they share common principles because they navigate. 00:52:50.817 --> 00:52:55.457 So the point here is that, Fritz, I could argue that in case of the pancreas 00:52:55.457 --> 00:52:59.597 and the brain, we might have all sorts of regulatory genes that really tightly 00:52:59.597 --> 00:53:01.157 orchestrate this process, 00:53:01.397 --> 00:53:06.417 while the same might not hold for the complete pathway of the C. 00:53:06.437 --> 00:53:09.057 Elegans or the trajectory that cells go through in C. 00:53:09.097 --> 00:53:14.277 Elegans. you will not have single regulatory genes guiding and orchestrating that process. 00:53:14.657 --> 00:53:18.897 Well, it's never a single gene. It's kind of a combination of many genes. 00:53:19.357 --> 00:53:21.397 It's between one and many. 00:53:22.237 --> 00:53:28.797 But with the bacteria, with the E. coli flagella, it's a different system. 00:53:29.137 --> 00:53:32.537 It's not developing. It's hardly proliferating. If you want to see it. 00:53:32.817 --> 00:53:38.617 I meant something else with this. It was like seeing a similarity between the flagella of E. 00:53:38.677 --> 00:53:42.117 Coli in our legs because they're both used for navigation and say, 00:53:42.177 --> 00:53:46.057 ah, this is now possible to have a common principle. Maybe they are. 00:53:47.226 --> 00:53:52.966 I do, roll it out. It seems a long shot, though. So those abilities are probably 00:53:52.966 --> 00:53:54.666 in our genome. They're just not expressed. 00:53:55.006 --> 00:53:59.486 And I think what you're saying is it's not surprising because we're all descended 00:53:59.486 --> 00:54:04.786 from a common ancestor and C. elegans and the mouse and ourselves. 00:54:05.126 --> 00:54:09.026 In that common ancestor, there were stem cells. They migrated. 00:54:09.246 --> 00:54:10.666 They built complex bodies. 00:54:10.926 --> 00:54:16.966 And so what more do you need than those mechanisms? So most of it's already going to be there. 00:54:17.226 --> 00:54:21.046 And you say you're not surprised. I'm a little bit surprised because I think 00:54:21.046 --> 00:54:24.846 that maybe in rodent brains, there's more that stem cells do. 00:54:25.106 --> 00:54:28.946 But for you, that's not a qualitative shift in what they do. 00:54:29.026 --> 00:54:31.406 It's just a bit more richness, perhaps. 00:54:31.526 --> 00:54:33.946 Yeah, it's more richness. I don't think it's more complex. 00:54:34.726 --> 00:54:38.666 So look, now that you gentlemen refuse to see the light, I'm shining on this 00:54:38.666 --> 00:54:40.206 issue. Maybe we should get to the finish line. 00:54:40.806 --> 00:54:43.846 And the question there is, so Yaki, as you also shared with us, 00:54:43.966 --> 00:54:48.086 it's actually interesting. saying the community of people doing your kind of 00:54:48.086 --> 00:54:52.886 work, like modeling these developmental, it's actually rather small. It's very small. Yeah. 00:54:53.226 --> 00:54:56.726 So you're sort of chipping away at this, making progress. 00:54:58.386 --> 00:55:02.566 So, but in that sense, given what you've seen and learned in, 00:55:02.586 --> 00:55:07.186 in this sort of as a solo agent in this, this field of developmental biology, 00:55:07.386 --> 00:55:11.006 what would be Jackie's law that we should adhere to in trying to understand 00:55:11.006 --> 00:55:14.866 the developing biological system? Well, keep it dynamic. 00:55:15.126 --> 00:55:18.426 Look at the dynamics. I think that is the most important. 00:55:18.766 --> 00:55:21.666 I think many people in this field fail. 00:55:22.631 --> 00:55:29.271 To look at, to think dynamic, to think that an event now has effect in the future of the animal. 00:55:29.851 --> 00:55:34.471 And if you look at that, you'll find an all new world of results. 00:55:34.731 --> 00:55:36.211 That would be my law. Okay. 00:55:36.671 --> 00:55:40.391 Now, Tony likes traveling, and so I ship him around the world. 00:55:40.431 --> 00:55:41.331 So when are you coming to Israel? 00:55:41.611 --> 00:55:44.511 Exactly. I'm going to tell you that. I can tell you this exactly. 00:55:44.711 --> 00:55:49.611 Four years from now, he's going to come to Israel, and he's going to meet up 00:55:49.611 --> 00:55:52.311 with you, and he's going to check a prediction he's going to make today. 00:55:52.311 --> 00:55:55.691 Tony is going to ask you, look, four years ago, you made a specific prediction 00:55:55.691 --> 00:55:59.331 that you would observe X in your simulation. 00:55:59.711 --> 00:56:02.711 And today, I want to know, four years in the future from now, 00:56:02.831 --> 00:56:04.871 whether this was validated. 00:56:05.071 --> 00:56:10.111 So, what's the one prediction you're going to make for us today that you will see tested by then? 00:56:10.371 --> 00:56:12.591 Yeah, well, I'll give a different kind of prediction. 00:56:12.891 --> 00:56:15.731 I'll predict that more people will do my kind of modeling. 00:56:17.391 --> 00:56:21.771 Come on, you can do better than that. It has to be about the mental biology. No, no. 00:56:22.831 --> 00:56:27.651 Developmental biology. Okay, so I predict that more people in development biology will do. 00:56:27.831 --> 00:56:34.511 No, I really think that what I'd like to see is not prediction in the lab that 00:56:34.511 --> 00:56:37.111 I've predicted and was accurate. 00:56:37.331 --> 00:56:41.171 I would love to see it comes out of the model, but I would love more to see 00:56:41.171 --> 00:56:47.451 people doing my kind of science. and maybe in the long run there will be a model 00:56:47.451 --> 00:56:51.751 of a human being that we can play before we go do anything. 00:56:52.351 --> 00:56:56.071 That can be something. This would be my vision. 00:56:56.311 --> 00:57:01.671 I'm not looking into a particular prediction that will be validated. 00:57:02.171 --> 00:57:05.471 All right. Giacchiosetti, thank you very much for this conversation. You're welcome. 00:57:07.151 --> 00:57:10.331 Antonio, you're a bastard. You're supposed to make it difficult for him. 00:57:10.551 --> 00:57:14.511 You're not supposed to agree with him. The CSN podcast was produced by the Convergent 00:57:14.511 --> 00:57:20.631 Science Network of Biometrics and Biohybrid Systems, a project funded by the 00:57:20.631 --> 00:57:23.231 European Sevens Research Framework Program. 00:57:24.811 --> 00:57:30.131 For more interviews, recorded lectures, or upcoming conferences in the field 00:57:30.131 --> 00:57:36.171 of biometrics and biohybrid systems, go to csnnetwork.com. 00:57:36.400 --> 00:57:44.560 Music.