 So now it's our pleasure to welcome Jose Manuel Garcia-Aznar, also known as Manu, which is shorter. And so Manu is full professor at the University of Saragossa, his principal investigator of multiscale mechanics and biology group at the Institute for Research in Engineering of Aragon. And so his talk will be on cell mechanics and cell modeling, so for migrations and other problems related to cell deformations and related to his ERC grant in silico cell. So once everything will be ready, so we'll have the pleasure to listen and watch to his great story. So I don't know why, but it doesn't work. So I have passed my presentation to other computers, so I'm sorry because there are some animations that didn't work, but most of them are working. I hope. Okay, so I'm going to present my work that I entitled Individual Cell Migration from Computational Modeling to Experimental Validation. So I would like to remark that what is very important is cell migration for many processes, many biological processes, like disgeneration, morphogenesis, and cancer invasion. So it's very relevant to try to model how cells are moving. And for understanding this, it's very important to understand the multiscale phenomena that our organizing is, how our organizing is organized. And in the case of the cell, it's very relevant to focus in the interaction what happened between the cell with the cellular matrix and what happened with the cell with its own components. So we are going to focus mainly in these three scales. And from the beginning, we have developed many different kinds of numerical models to simulate cell migration combining different approaches, like at TSU level using classical mechanics reaction diffusion analysis with finite element analysis. Later, we tried to go down into the scale, simulating individual cells using different kind of approach like box cell, finite element analysis. But in addition, we try to understand what happened inside the cell simulating the cytoskeleton. But first of all, we need at least a little bit to know how the cells normally move. So the migration cycle is divided normally in these three processes. First is polarization. Normally the cells are going to receive signals and depending on these signals, the cell is going to polarize. After that, normally cells create protrusions in the front part of the cell and later there is adhesion formation here, new due adhesion and in this part the adhesion starts to be quicker. And after that, this process, normally what happens is the cell is growing in the front part and in the back part or the rear part, normally contraction appears. And along when contraction appears, the cell is moving. This is sorry because this is the first simulation that... This is the first simulation that it doesn't work, sorry. But normally we can see that this phenomenon is quite good in 2D dimension. But in 3D dimension that we are going to see more simulation, more real experiment, and we can see that this phenomenon doesn't occur exactly like it's described in 2D. In 3D, the phenomenon is completely different. To advance I have to... Ah, okay. So normally there are different kinds of stimuli that are regulating cell migration. We have chemical cues or mechanics like stiffness or electric field or some kind of signal that is binding to the matrix. So our first idea was to try to simulate cell migration using classical continuing approach where we simulate the cellular matrix taking into account different kinds of flows and incorporating the conservation equations. And later we try to incorporate in our model solubile factor, rouse factor that are moving that is regulating chemical signal. And later we incorporate cells. But cells can exert forces, are elements that are passive but at the same time are active elements. So to take into account this phenomenon, we consider cells and cells can exert forces on the matrix. And these forces we simulate as a first simple approach the cells like a set of springs and later a system that is the contractile system that is very similar like it's following a classical yield law that is normally used in our model. So in this way we incorporate the forces due to the cell and we are able to simulate... Okay we incorporate this like incorporating also possibility of migration, proliferation of cells, differentiation and we are able to simulate, we incorporate the different factors oxygen, rouse factor, angiogenesis and fibroblast in classical reaction diffusion. Oh sorry. The question and after that we can see here for example the phenomena of how one wound is healed due to this phenomenon. So in this case it's a wound that is not very big and so the cells can repair the heal without problem because normally we cancel. We have analyzed different conditions like when the wound is larger and we have used stitches, this kind of thing. But here my main intention is you can see here how the cells are moving in a classical continuing approach. So our idea is to try to understand how one cell is moving. For that we started to work with other groups that do experiments in microfluidics. We started to collaborate with a group in MIT with Professor Royer Cam that developed this kind of chips and in this kind of chips we have here a gel that try to replicate to simulate our tissues and later here we simulate one cell by a very simple approach only like a set, a set of box cells and we solve the problem in a very easy way. So we applied a gradient and we also applied a chemical gradient and we can simulate for example how the fluid flow changes around the geometry. But our cell is going to... Okay, this is a validation in collaboration with this group and later what we did is a very simple model where we have the cell, it's a set of box cells we have the nucleus of the cell and later we are going to add an element in function of the mechanical or chemical stimuli that the cell is receiving. In this way this model is purely phenomenological. So in this way we have for example a section of the cell we have the boundary and we decide if we add... if the cell is going to grow or not depending on the mechanical stimuli for example we have this part of the cell this is the part where we have the maximum stresses so there are different possibilities for example for this box cell and given that this box cell is closer to the stress to the maximum direction of stress this box cell has more probability to appear. The model is a simple idea and with this idea we of course we couple chemistry, a flow condition and mechanical condition and we establish the probability and we add elements and remove elements I'm sorry because these two simulations didn't work sorry but the next one is working and for example here we can see how one cell is moving inside one scaffold changing different kind of stimuli we are able to regulate how the cell is moving and it's a pity that I couldn't show you the previous presentation because the model replicates, reproduces very well the moment that the cell is presenting but this model is very simple it's not able to reproduce correctly the shape of the cell so in this case we have problems when for example we have a fluid flow that can move the cell for example there is a resin paper where there are experiments to measure the properties of the cell and where under a fluid flow are forced to pass through this channel and this you can see how the cell are deformed we try to simulate this phenomenon but it's not possible to reproduce accurately the geometry of the cell with the previous model so for that we define a new approach so we need to improve the representation of the shape of the cell and later we need a good fluid cell interaction for that we define an immersed finite element approach in which we are going to couple both phenomena so on the one hand we have the solid equation that we solve later we have the fluid equation and in the interface we couple the velocity and also the normal stresses and with this model we solve this in a coupling formulation later we use how we do this we have a fixed mesh for the fluid and later we have a mobile mesh that is regulating the movement of the cell so with this we use a level set function to define the geometry but here we have everything is fluid here everything is solid and these intermediate elements we have an interface where we distinguish between solid and fluid so what we are doing is we insert, immerse the geometry we solve fluid solid equation obtain the displacement, update the mesh and later we update the nodal variable so for that we need to map back to the less configuration to update other variables like stress so in this way we started to simulate for example how under a fluid flow condition how the cell is going to the form when it's going to move inside the channel and you can see that the prediction of the shape of the cell is quite similar quite similar how the experiments are determined but this is really a qualitative validation it's not really a quantitative validation but our idea is this cell okay we later did a variability an analysis of different parameter sensitivity analysis later one of our main questions but these cells have nucleus so it's interesting to try to model a cell with a nucleus so you can see here that we have a nucleus we consider the nucleus as solid later we have a fluid here and another fluid here in this case we have two interfaces and we can see how the fluid is going to the form the cell and in this case we are able to simulate not only the stress of the cell but also the shape of the cell but also the stresses and the strains of the cell so with this model we are able to predict better the shape of the cell but our interest is also to try to understand what happened in the cellular level what happened inside the cell so for that we found a very simple and interesting experiment this experiment a cell was set between two plates the up plate is flexible however the low plate is very stiff when you put the cell inside you can observe that if this stiffness is very low the cells are lower for force however if we increase the stiffness of the plate the force that the cell is doing is much higher so there are many researchers that consider that this is the system that the mechanism system normally the cell uses to sense the stiffness of the medium of the surrounding tissue normally the cells contract the problem is you can see here that initially you can observe that if you increase the stiffness more or less the force that the cell is doing is linear however there is one moment if the stiffness is higher of this value the force is saturated it is for a stiffness the cell is not able to sense the stiffness because the maximum force that the cell is able to do is achieved so the cell is not able to do more force it's like oh see we tried to for example bend a bar of steel depending on our muscles it's going to be but there is a value of the stiffness that we are not able to so it's very interesting this result because we can observe that for example when a tumor exists the tumor tissue normally presents this kind of stiffness properties so the cell one of the problems is the cell are not able to sense the mechanical stiffness because for them it's hard tissue so the sensitivity of the cell is in this region why it is happening? we try to understand what are the mechanisms that regulate this saturation for that we try to simulate the full cytoskeleton using a particle based approach where if we see here we are going to simulate the acting filament and later the cross linker of this filament and in red the myosin motors so in this way we use a Brownian dynamics particle based approach and later an acting network that is crosslink and with this we simulate a piece of the cytoskeleton of the cell later using the Langevin equation and defining extension, bending potential and repulsion potential between these particles and also we consider the possibility of unbending between the different crosslinker and we assume this law for the walking of the filament the filament, the motor sorry the myosin motors are going to walk along this acting filament following this law and normally they tend to move where the bar ends and with this model we put all together and we try to understand which is the energy and behavior when we change the properties of the surrounding medium so we change the elasticity of the surrounding medium and we can see here that we simulate this and we can see how when the stiffness is very low in a big contraction however when the stiffness is higher the cells, the myosin molecules try to move but they are not able to deform the external geometry so we observe here that also the topology, the geometry of the connectivity the morphology of the network is different here it is more connected however here we have holes in the mesh but what is more interesting is we are able to predict the saturation phenomena so when we increase the the stiffness we increase the forces that the cell are able to do but when the stiff is very high we can see here how the force of the curve is always here so this model is able to qualitatively to predict how the contraction is regulated by this myosin acting system so the problem is when we finish this part is we consider that we have models that work quite well but they don't we don't know if they are good predictions or not so we try to determine to develop more quantitative analysis in order to compare with real data so in this sense what was mainly numerical started to create experiments in our lab so the idea is we are going to try to understand how the extracellular matrix properties regulate the migration of fibroblasts in 3D so for that we compare two different kinds of materials collagen, that is normal tissue that we have in healthy tissues and fibroin tissue that is a tissue that normally occurs when we have a good and we incorporate different chemical factors to understand how these chemical factors regulate the behavior of the cells so this is our strategy that we try to do is first of all we characterize the hydrogel to fit the experiments so the numerical model and we try to develop our experiments where we measure what we need to create our models and later with these models in fact now we have first results we are trying to improve our experiments and to establish this loop in which we hope to advance quicker than when we use results from other groups or even from the literate so I'm going to show here some preliminary results so what we have for SHERP normally is in this paper there is an interesting review of how cells behave in 3D so our experience with the previous simulations that we developed with the simulation with experiments of microfluidics in cancer cells were epithelial and these epithelial cells normally move using an amyvoid system that is completely different what we observe, like we are going to show now in the case of the fibroblast healthy fibroblast or mesenchymal cells normally migrating a different way how of all in 3D there is a competition between protusion that the cell are creating and there is a competition between this that in between this protusion it depends on how is the geometry and the mechanical properties of the surrounding matter so our first knowledge was completely different and all the work that we have developed it wasn't valid for simulation fibroblast in 3D so we found this paper and it corroborates with the experiment that we made in 3D so the first idea was ok we are going to characterize the gels to understand the difference in the geometry and we can see that the collagen is this one the fibroblast is this one and the geometry is not exactly the same so you observe the collagen is mainly a fibroblast tissue but it's mainly the fibroblast I've been however the fibroblast is more is the fibroblast so the pore side is bigger is higher in collagen in comparison with fibroblast the fibroblast has present a similar radius and it's more bundled in the case of collagen than in fibroblast in addition we measure the permeability where we find also differences in the permeability how the fluid is moving inside we observe that in collagen is more permeable than fibroblast and of course we quantify rheological properties G' in collagen and fibroblast and we observe that both of them have a non-linear behavior in function of one level of strain that if your serve is lower in the case of fibroblast fibroblast and the properties are much higher in the case of fibroblast in comparison with collagen so we have to keep in mind this and we are going to observe a very different behavior in the cell during immigration due to this reason so the first idea is we are going to put the cells in our microfibre chamber but the problem is we are going to use a chemical attractor to in some way regulate the cell movement in one specific direction but what we observe that we need to understand how the chemical gradient is formed inside the material because we didn't know how the diffusion is produced so this is the kind of chip that we use that is described in this paper and normally we have one central channel where we have the gel is here and here we have the cells and normally we add a graph factor in this case is the platelet derived graph factor that is the graph factor that occurs normally when we have wounds and when we include this factor what's happening this factor is moving through the gel creating a gradient and normally what we observe is there is a part of this factor that is degraded so in contact with oxygen is degraded we have to quantify this that is not easy later is the diffusion but this diffusion is indicated in grey color and later you see that there is yellow part of the factor because it's bound to the matrix so it's very difficult to validate this and to understand how the gradient is formed and which is the time to achieve a stable gradient our idea was to use a simple computational simulation but we need to validate so to validate we use ELISA, ELISA is an experiment that we are able to evaluate which is the concentration of the factor here at the beginning okay and here and after a number of hours we obtain the sample from here and we know the concentration of the factor in each time, in each period of time so with this idea we establish a classical reaction diffusion equation where we have the coefficient diffusion follow this equation that the typical it's down like a question where we have several parameters, this is the radius of the molecular, the molecule this is the the radius of the fiber of the matrix and this is the matrix void radius and incorporated this we are able to estimate the diffusion coefficient and later we assume that the gradation is linear with the concentration of the factor and the binding we assume that is linear with the factor very simple approach and in this way we evaluate and we can see here this is the result after 24 hours and we can see here this is the initial conditions in the addition channel that is the green one and after 24 hours this is in collying, this is in fibrin we can observe here this is the in the opposite channel what happened after 24 hours we see that this is the experiment and this is the numerical prediction and we can see that the model is able to predict quite well how is the distribution of the factor so after that this is the chemical factor that we have after 24 hours in the collying and in the fibrin so in collying we have diffusion and binding however in fibrin we have mainly diffusion one that we now did we evaluate how the cells are moving in 3D and we analyzed the first thing is how is the vinculin in the acting distributed in the cell so we can observe that in both cases we have branches protusion we can see for example with the first simulation that I showed you real cells in which are more amyvoid but here you can see that we have a lot of branches and there are small differences in the shape and we analyzed the following cases control, no factor this is the first case we have cells inside later we incorporate the platelet that is the factor that normally we incorporate the healing of wounds and later we incorporate plebistatin plebistatin is a drug that removes the acion or partially removes the acion of myosin so myosin if you remember is the system that the cells use to contract so what happens if we remove contractions these cases and here we can observe first results I am going to remark simulation should work sorry ok we are going to observe this one that is the other is the global behavior but here we are going to see the local behavior that is very similar if we observe in the case of control we observe that normally the cell branches and later is moving in colline however the cell are launching creating branches protusion however the cell is not moving because we consider what happening is the matrices show a stiff and the pores are smaller that the cell are not able to move if we incorporate the chemical factor the movement in comparison with controlling colline is incredible I am sorry but we cannot see in the other simulation but here even you can see that the cells are moving much quicker than here in the other simulation in the other animation is very clear the advance is tremendously higher when we have the factor so it is clear that the factor is regulating the movement or the place toward the site of the wound for example but what happen with the matrix if we observe here the fibrolet is not able to move the protusion are longer than in this case however the cell is confined is not able to progress so it is clearly that even although the conditions of the ground factor the signaling is working the condition of course is clearly limiting, regulating the movement of the cell but what happen with we incorporate Blevistatin when we incorporate Blevistatin the branches are even longer the movement is more or less achieved we can observe but normally if you observe here or here normally the cells create branches and normally retract branches however when we have Blevistatin this phenomenon is occured less so later the most important is not only to observe this is to quantify this so when we incorporate BDGA protusion and motility is increasing and Blevistatin the protusions are rolling so we observe here this is the result when we incorporate the ground factor and we can observe differences between the cells the cell 1 is the one that is close to the ground factor the cell 3 is the one that is farther from the ground factor so the measurements of cell 3 are very similar to what happen in control experiment however in cell 1 you can see here this is the red trajectory the speed is much higher here we can see a comparison of control PDGF PDGF plus Blevistatin and we can see here that and this is collagen and this is fibrin and here we can clearly observe the difference in with PDGF the speed is much higher in comparison with control and much higher in comparison with fibrin and also it's very much higher that we compare with Blevistatin ok this is later we have more data that I going to advance but it's a comparison of all the phenomena more quantitative so we have a clear study of this phenomenon so the next step after we develop this result this experimental result we create a numerical approach, a discrete approach where it's a very simple moment our first idea is for us a cell is a nucleus and later we have a number of protrusions that define the phenomena try to describe the geometry of the cell and we are going to incorporate three main mechanisms we assume that chemical factor is regulating chemo sensing, so we are going to try to simulate how the cell are sensing chemical factor, later we have protrusions dynamics, how the cell create protrusions, retrace protrusions and later how the extracellular matrix that these properties constrain this protrusions so the cells are embedded in a extracellular matrix and we are going to simulate how it is intracellular for that we are going to use a reaction equation and initially we assume that the membrane receptors the cell membrane have receptors and all of them are distributed homogeneously over the cell surface of course if the cell for example we have this chemical radian the cell is here, in this point we are going to have more chemical stimuli that in the rear part, so we here here the distribution in one map of the factor distribution over the cell and in this cell we use this kind of model so we have a factor here so we simulate the factor we incorporate that this factor come binding to a receptor and inside the cell we have what is called PI3K that is a chemical factor when this factor is bound to the receptor activate this factor and when this factor is activated the protrusion start to grow so for that this is a stochastic phenomenon that is the most the part that there are most highest computational codes and we have, we use a poison Poisson distribution we have to name it as stochastic and later we have a time resolution using the Gillespie algorithm that is quite usual in chemical reactions and later so we use these equations and this equation is activated in each point we discretize the cell surface and each point of the cell surface stochasticity and for that we use techniques in this space multivariate non-homogeneous Poisson distribution published by Sathman so with this idea we in the place where are normally the peaks of PI3K where the location of the protrusion is going to appear and the size the size of the protrusion the protrusion is going to grow depending on the stiffness of the matrix and also it's going to retract and it's going to create this dynamics and in function of this dynamic the cell is moving so for that we assume that an analogy with a surrounding matrix is the SLB theory so when the protrusion is growing you can see here that forces are going to appear there regulating and constraining the growth of the protrusion so for that we use the SLB theory where the strain is regulated by the value of the PI3K now later we have the forces that each protrusion is reacting and we assume that here in the nucleus we have the reaction and in function of these forces the cell is going to lead the speed of the cell, of the cell nucleus depending on the forces that all the protrusion are doing so here today I'm going to buy a ticket for lottery because it's so bad luck so we can see here how we are able to simulate the migration of the cell in 3D so these are preliminary results and we are working to improve the model but the model you see is very simple, it's not very complex the most difficult is the chemical reaction that defines how the intracellular cell will finally regulate the crowd and the current and the crowd of the protrusions but we have started to quantify and to compare our measurements with our simulation so it's clear with the model we can obtain many results and we can observe for example how is the how is the influence of different chemical factors depending on where the cell is we obtain the different speed in collaging and in fibrin if we compare with the experimental result we can observe that in vitro we have in average our results are quite good if we compare the size of the solution of the philopodium the prediction is not so good and we are working to try to improve this part but more the definition of the movement is more less good but where we are having problem is in the variability the problem is our model is stochastic so our numerical results present a variability in fact if we see here this is a box plot where we show the median of our simulation and you see that our chemical model incorporates the variability of the numerical results but the problem is if we see here in the gray area we are going to show the variability of the experiment so we can observe that the variability of the experiment is much higher that the variability that we have incorporated incorporated in our model in our opinion it is happening because we have incorporated the probability of stochasticity in the chemical reaction but not in the matrix we have assumed that the matrix we have quantified the properties but the matrix we have assumed that is homogeneous but it is not true is it heterogeneous that we show the fibers mesh that is clearly heterogeneous so we need to incorporate this to obtain at least this variability because if not we are not able to later the other problem that we have is to quantify measurements in vitro it is not easy of a fault in fibrin because the measurement is very very small in 3D when you do experiments in 2D the cells move very easily they are quite higher however in 3D the movement of cells is very slow it is very slow and in fibrin is even much much more slow so our challenge now is to try to predict also not only the average value but also the variability of these results so I would like to finish with some general conclusions so I have tried to show some numerical examples that we have used to investigate how different mechanisms can regulate migration and in our opinion numerical simulation is very good tool but the problem is we require validation because without validation the impact is not enough we require validation and for validation we need to combine technologies we need to combine experiments in vitro or in vivo but we need to incorporate image processing techniques to quantify to measure data and later with this information we can validate and test our model it is fundamental normally mechanobiology is relevant we have to investigate more how is this relationship between mechanical and chemical factor it is very important and also although computational analysis is very interesting and we have observed the previous presentation that we can develop a lot of realistic simulation in my opinion still not sufficiently powerful I think we need to develop novel numerical strategies that allow to link scales and reduce the computational cost I think it is fundamental so this work has been achieved thanks to the institution that give money for doing research and also people these results are mainly in the PhD work of two persons, Ayana and Federico and also the collaboration with other groups in Spain and around Europe and USA and thank you very much and sorry very much for all the problems that I have many thanks very much it is wonderful of your work any question? Yes Mariano wonderful presentation I have a question the matrix is continuum mechanics and which is the coupling between is the force we didn't solve mechanics we use an analytical solution that the SLB solution so we consider that a protrusion is like a solution in one matrix and we use this theory so this is a simplification that we have made but it is clearly non-realistic it could be done in this way let's say you can solve mechanics mechanical simulation in the matrix and then couple it in fact now we are working on that the first step is we are coupling with finite element but the problem is that simulate the mechanical behavior of the gel this is not a problem but we need to incorporate a discrete model that simulate the fibers and incorporate the stochasticity of the geometry of the matrix and this is our next step maybe partially related to that also is because in principle even in your microfluid although it's relatively simple you still have like maybe the not perfect geometry, you have the oxygen you have the nutrition going there don't you think that a possibility would be like a full multi-physics system and then try to solve it on the supercomputer or do you say like no it's maybe a better approach like you do is like or kind of simplify all the different steps and then solve them separately and then try to integrate the information with all kind of simulation of everything I don't know what your idea is about Yes, our idea is normally a physiological matrix have a high variability in properties even these collagen gels depending only on the temperature on the day how you take the piece of collagen that you obtain a different result so our next idea is to work with matrices that perhaps although they are physiological we can control things we can control things like properties and we obtain a repetition of the properties of the matrix so the next step is ok we are going to lose to lose in some way physiological properties of the matrix but we are going to incorporate matrices in which we are able to control the size of the pore we are going to be able to control the degradation of the cell because here the cell are able to degrade so we are going to incorporate matrix where the cell are not able to degrade and we hope in this case to reduce the variability so the problem is even in this case that is a bitter experiment that is ok there is a high control of many things even in this case we have a high variability so we are going to try to advance in this direction to reduce this variability in this moment we find a new solution perhaps the solution could be to incorporate a model we incorporate the geometry but in any case it is going to be difficult to characterise a certain matrix but in any case I think we need to reduce the variability in order to improve our model and to improve our knowledge of the process any more questions? no please use the microphone thank you for the presentation I have a question I want to understand your model this model of random protrusion can are you not considering explicitly polarisation or you want to include it after or it is not important for this or it can polarise like as an emergent behaviour in our opinion there is no polarisation in this kind of moment we did not observe polarisation you see the first picture there is no polarisation so you plan the model because of this fibroblast migration and there is no polarisation because at the beginning of your talk you addressed the polarisation problem so I was wondering if you are planning to no because initially when you check when I started to work how the cells move the first idea is this scheme that I show polarisation, pla protrusion and contrast and when I started to study 3D it never happened and this phenomenon of polarisation normally occurs in amoeboid migration even in 3D in this case yes you have polarisation but when you have mesenchymal migration that is the migration that is mainly regulated by protusion and in this other review paper that is later they mention that they never found polarisation it is a phenomenon of competition between protusion so you don't need it another question in your gradient validation in the case of collagen you have a non-linear gradient basically because of binding in the fibre you have a linear gradient no, it is non-linear but in this case the non-linear is more view both cases we have degradation the degradation is similar in both because it is not depending on this so we assume the same degradation for both systems and the degradation gives you a bit non-linear but it is more non-linear in the case of collagen but then you compare the one concentration and the other you don't have a way to validate actually to visually see in the experiment if it is the non-linearity can you validate that it is very important to do it in fact we did it with Destran with Destran we obtained this non-linear behaviour and now we are trying to do it with the factor because the problem with Destran is much bigger and our idea is to try to do it with the factor but we are having problems to put a colour in the factor and to visualise we are having problems but we are trying it I am numerical so it is difficult for me these things thanks maybe a little bit more philosophical question what is at this moment your research question because in the sense you go you say I want to simulate things at some point you come from wound healing then you go to microfluid is it now that you want to simulate microfluid or do you want to understand the problem what is your research philosophy our main work is going to focus to understand which are the mechanism that cells are using for moving so the idea is to combine technology in order to understand this better our main focus is to understand how they move in 3D so with two main ideas the first one is for cancer we would like to understand which are the mechanism that cells are using for moving and to metastase from one tissue from one organ to other so our idea is to understand how they move in this condition and the other one is to understand how healthy physiologic cancer like fibroblast move in normal conditions in order to improve our generation capacity so our body is able to regenerate the problem is a question of size the size is very small the body is able to regenerate but if the size is very big the body is not able because cells are not able to migrate so far so we need to try to investigate how we can improve this and for example you know that the salamander is able to regenerate all the limbs so our idea is to try to understand how has to be the condition in vitro to translate and to try to to tell mechanisms to improve regeneration so we have two main questions to try to stop this tumor cell and to try to improve the movement of physiological cells I have a couple of questions well three questions, two groups the first question is a little bit about the tool particularly for people who would like to start to do this kind of advanced research in modeling sometimes that's important that you're not limited by the tool or you don't see too difficult to the tool so to which measure are you developing your own your own computational solving tools and visualization tools and to which measure are you using already built tools in which you can friendly enter only your scientific calculation equations okay it depends okay I don't know if I understand the question but if I understood your question is normally it depends on the student in my case I normally give freedom to the student so normally when one student comes to my lab I said we are going to investigate this and normally we have different, for example we have our own finite element code we have a commercial finite element code but normally I work together with the student and we define which is the best approach to solve a one specific problem so my idea now is all the students I try that all the students try to do some experiment and some simulation which is not easy because normally numerical people only like to do simulations people that do experiment don't like to do simulation so what I think is very important to do all together and to understand everything because if you know what you obtained from the model you are able to try to investigate how you can achieve some specific measurement in your experiment so in this way we can achieve a good feedback between both I don't know if it's your question or not but ok but anyway and then there's other questions I have a little bit more scientific so first question is when you're simulating the deformation of the cell in 3D so I'm wondering so whether you're respecting so basic continuum mechanics low in elasticity like reversibility deformation tensor gradients etc because in the end what happens in the cell is very different from what is happening inside the cell you mean yes so when you're simulating the deformation for example the interaction of the cell with the wall of the pipette so what kind of what kind of solver what kind of theoretical approach do you use so is it elastoplasticity is it something an hyperelastic approach for the solid we use an hyperelastic approach and in this moment what we do is when we have the new geometry we extrapolate the variable that we have before to estimate the stresses so what we are doing is to map back obtain the data there and we update to verify the stress distribution so normally we estimate the boundary is moving and when we have the new position because we have a fixed mesh and later a movable mesh the boundary is moving when we move there we have to locate where was this point before so when we know where was this point before we update the variable there to the new location in the new position of the mesh you see my point this is what we do don't you expect don't you expect that in the real system you have some kind of fluidization in the case we only simulate the cell like the nucleus of the cell is a solid the sito planche and the sito skeleton in this case we simulate like a fluid because in this kind of tube normally the cell is more similar to a fluid than to a solid the cell is more similar to a solid when we have mesenchymal migration when we have amyvoid migration of this kind of migration normally the sito skeleton is uncoupled and the cell is mainly the sito plasma is mainly a fluid it's not a solid so in this kind of simulation only the nucleus we simulate it like a solid and then the last question I had from a scientific point of view is what is really rational to use fibrin because it's kind of tricky material when it's subjected to local forces it's able to locally partly deassemble then reassemble when the force is gone it's a material that you've shown comparison between collagen and fibrin but indeed fibrin has an infinity of states depending on yes you are right ok it was ok in my opinion it was our love experience in this kind of experiment so initially we have no idea we started with two material that are in a wound collagen and fibrin we started with that now we are mainly don't consider fibrin because it's too steep experiments are and we are mainly concentrated on collagen and now we are working for example changing the concentration of collagen incorporating different kind of so we are working with that and later we are working with other kind of material but fibrin was a fresh approach that we use but yes I fully agree with you it's not it wasn't very useful ok for us the idea that the cell is ok the constraints of the matrix is very relevant to determine how the cell is moving is it worth for that only the question it's lunchtime no other question no so lunchtime in the practical sessions again the same places 230 keep in mind that today is the last day and that presentations for tomorrow and the report the record has to be written but Jerome will explain it later no no the question I had for example of course you use a for example when you calculate the projections of the laminopods