 My name is Matthew Fulsack, I will be the editor-in-chief of the proceedings of this conference and I'm happy to introduce to you our special guest today, Kino Peter Germany. Peter is a training pilot chemist. He is from the Semmelweiss University in Hulobach. His major fields of study are molecular chapter rooms and networks. He is especially concerned with stress-related and aging-related networks effects. He also has 14 books and more than 220 scientific papers to carry out several honors and rewards. From 2008 to 2010 he was the member of the Vice-President Committee of the President of Hungary. He serves as an editor of cell stresses and chapter rooms of course one and of major scientific papers and he will be in May, coming May in residence in the Rockefeller Foundation in the larger center in the US. From his many publications, he specially commends to you his book, Weeklings, which is really a very interesting read and a very good read in the field of networks he serves. Yes, and this single to the end is crisis respondents and crisis management. Peter, thank you very much for your introduction. And as you can hear, I am coming from Hungary and approximately 100 years ago we were together with the Austro-Hungarian monarchy. But after the Austro-Hungarian monarchy we are in a constant crisis, at least we feel so. Not only in Hungary, I don't know, maybe you laugh, but certainly yes. I might actually be about this crisis and I would like to introduce you to the responses to crisis from the point of biological networks. But the key point, as Manfred has told already in his introduction, would be that this is fairly general behavior and you can actually see about yourself or understand yourself a little bit better perhaps, or at least I hope, by the end of the lecture. Now, the first slide is about how we think about networks in my group. My group, my network science group, is a pretty much multidisciplinary group consisting of not only physicians like in a medical university, but obviously from physicists, from mathematicians, from even economy people. So we try to understand networks as a general phenomenon. It is very useful, or it is possible, because there are many, many features of networks like the small learners, you might have heard about it quite often, this is the famous six degrees of separation as we usually call the very well-connected networks, the hubs, the existence of nodes which have much more connections than average, the nested hierarchy, and the leaflings, which have been already mentioned in the introduction, which seem to be all pretty general features of all networks, including social networks where the nodes are people, the connections are personal contacts, and biological networks where the nodes may be cells like in our brain, or proteins like in our cells. In my group, therefore, we are thinking always about this generality of network properties, and this makes us to utilize the billion years of experience, among others, in crisis of biological networks, and we use it for everyday life, and moreover, it is giving us a judgment of importance of the finding, because if the finding in network science was true only for a particular network, that may be an important finding, but if it was true for many types of networks, including social networks, biological networks, and others, that is a fairly general finding, which may have a more general application that may be even more important. Lastly, or not least, I would like to draw your attention that if you generalize problems with a scrutinized scientific methodology like network science, then you may end up with quite creative, quite novel solutions to your scientific problems. I'll just give you one example in the next slide, but the basics is that when you stop, when you don't have the solution to your problem in network science, in social networks, for example, then you may reconfigure, reformulate your original question to a biological network problem. Then you may talk about it in different words, and different solutions may come to your mind, or vice versa, you may reconfigure a biological problem to a social problem. Now, I'm just giving one example. This is a paper published in Nature by several random authors approximately two years ago, and the paper, as the title says, is about early warning signals for critical transitions. Here critical transitions are changes of complex systems which are unpredictable, which are large, and which are quite dangerous, especially if you are part of the complex system. The authors were comparing three completely different ecosystems, so of neighborhood in the environment, market, we all know that, and climate. And very surprisingly, but very importantly, they found at least three common responses of these various, very different systems when they are approaching a crisis, when they are approaching a critical transition. Now, these were the following, a slower recovery from perturbations, so if the system was changed, then just before the crisis, it could return to the original state much slower than otherwise. Second, which was related to this first one, an increased self-similarity of behavior, so the behavior was getting to kind of a routine, and it became more self-similar. You can measure it quite nicely by various measures. And last but not least, an increased variance of fluctuation patterns, which I may kind of transcribe that once the system has been changed, then it was changing to the extremes mostly or more frequently, then it was close to a critical transition than otherwise. Now, when I was reading these three kind of statements, it occurred to me that, wow, we are dating this ageing in my laboratory, we are in a medical university, and if you think about an old person, a really, really old person, your grandmother, grandfather, or even great-grandmother, great-grandfather when they were alive, then you may recall similar behavior patterns. When they are getting old, then you start to develop daily routines, and then you start to stick to these daily routines more and more, and if a disturbance is coming, then you are getting really lost, more and more, once you are getting older and older, if someone is disturbing or something is disturbing you from this daily routine. So as a very short summary of these kind of thoughts or generalizations, you might say, or you might challenge yourself to say, or challenge the audience to say that ageing is an ageing early warning signal of a critical transition, and I see now you guessed what is a critical transition, and I think you were right, so indeed. From the natural point of view, death is a critical transition because before the network was connected, it had a giant component, we say, but after that your brain, your cells, all of your metrics in your body are disconnected quite abruptly, quite unexpectedly most of the time, so it is indeed a critical transition from the bona fide point of view. Now, I don't want to distract you with negative messages and very early warning, so I'm giving you the positive side as well that this will give you hints how you can prevent or how you can delay these critical transitions, and one of the major ideas is that prevention or delay is possible if you find the nodes, elements, actors in this complex system which have a more unpredictable behavior than others, why? Because in a critical transition, behavior is usually getting similar, so the nodes are getting to a similar type of behavior, they are synchronized in a way. If I would shout now that fire, everyone would go to this direction, maybe some of you would go to that direction, some of you might survive actually better because they would not kill each other on the single direction or single entry, they came to this room, so if there is a behavior which is unpredictable, that may be helpful in different systems to survive, like omnipotent operators in the ecosystems, market rules and stem cells, actually in our body, those are cells which are pretty much unpredictable, each type of cell actually in the body what is needed. Now, if you would like to generalize this or rationalize it in a silenced point of view, then you may really find three different behaviors in networks. The first one is a problem solver which is a specialized node which is solving a certain problem, a certain task in the network. The next one is a problem distributor, this is usually a hub. This is not solving the problems itself, just distributing it to other nodes which are in the solver. Now the most important from the current point of view, from the lecture's point of view is the serial which I called creative node. This creative node is changing its position in various network positions, not only in this tool, this was just an illustration, so it's very dynamic and actually samples the network, samples the properties of the network, samples the possible solutions of the network and therefore it is unpredictable and it can be creative from the everyday point of view because it can assemble different segments of another solution from various points of the network and therefore the system itself can enjoy this solution and creativity. I just would like to remind you that from this point of view the creative person has an unsatiable appetite for novel environments and for novel solutions and therefore this person might be called as an entrepreneur from the Schumpterian sense and actually this type of person or this type of node is a network entrepreneur because it is exercising this kind of property in the network. Now to exercise such a property you actually need a continuous refocusing of social categories if the node is a person and by this refocusing the creative person understands that there is no curve or values that connects isolated people and at the end what is most important exponentially enriches herself and himself with knowledge, with new knowledge absorbed from the network moreover this is actually a self-amplifying circle so it's quite beneficial. Since this creative node position is changing network position all the time this also means that it is vacant empty positions and letting others to occupy their position which he or she was occupying before. Now this is a very interesting feature of dynamic networks because this allows these dynamic networks to change the centrality all the time of the nodes and those nodes which are in the middle which were centered at this particular moment may be on the side and may be supporting those nodes which are now in the center, normal nodes. This is very similar to the function of our brain. As you are sitting here, as I'm standing here our neural cells are getting active but each second different neural cells are active and different neural cells are in the middle of other neural cells which are also active and this is changed so that neural cell which is now in the middle in the next moment may be on the side and may support the next neural cell which is at that time in the middle of this particular circle so this dynamics and this dynamical change of centrality is a very important feature of complex networks and I just would like to make the parallelism or the analogy for the society that a society, a school, a firm is working really well, really in a complex manner if it is dynamic like that so the boss is not always the boss in the traditional sense there might be a chance for other people to get to the middle to enjoy the support of the community and this is dynamic. Now this is a very complex idea I'm just showing you for a moment to show the bad and suffering life of a biochemist because biological data are messy and therefore you have to use many types of biological data if you would like to assess a network type phenomenon and you have to figure out which are those changes which are robust despite changing the biological source the data source, despite changing the method how you analyze your network and so on so we are trying to find this type of robust changes and in the next minutes I will tell you of some of these robust changes in crisis in these networks in crisis but before that I'm just giving you the method how we analyze this network this is a method to understand, to find groups in network groups are densely connected segments regions of the network which are not so densely connected to their neighborhood to their neighboring this definition is pretty woke definition it is not a precise definition it is not such precise as this formula on this board and therefore it is very difficult to construct methods to find groups in the network clustering is a notoriously difficult problem in network science our solution for this program has some steps which are giving you or giving the people who are using this method a pretty detailed knowledge about the network structure the very first step is to construct a local influence island a local island of influence to each node or each edge of the network so actually I am having here three local islands this is actually a network of network scientists themselves how they collaborate so each node is a person, a network scientist and each link is a collaboration by a joint publication so these are actually three network scientists there are no network scientists on the field and this is the influence zone of those network scientists where they are in a very close and dense interaction with other network scientists now in the next step we summarize these local influence zones and actually determine or define a certain dimension, a vertical dimension which we call community centrality which is actually the sum of local nodes so when a node is having a high position in this vertical centrality it has a high community centrality this means that this particular node was a part of many local influence zones so that particular node is high because it is participating in a large number of these intimate contacts with the local neighborhood and dense contacts in the local region interesting, if you define this kind of community surface there you may understand or you may end up to the solution but eventually the hills on mountains, on this surface are the groups themselves are the communities the network communities themselves and eventually these communities will be very much overlapping if you would like to get more information that you can download from our website even as a scientist blogging which makes it easier to use and our website is labors.hu Now this understanding of community centrality so actually summarized influence on the network and from the network to that particular node makes you to understand or helps you to understand the impact of that particular node in the overall network configuration and overall network traffic interaction structure which is any very interesting example in crisis and from now on I am focusing on crisis events in yeast and in other organisms this is the protein-protein interaction network of yeast you know from beer or from bread or from various daily products that you consume here a node is a protein and an interaction between two nodes is a physical interaction between two proteins interestingly when the yeast cell is allowed to divide because there is plenty of food in the neighborhood so there is no crisis at all it is dividing exponentially then those proteins which are involved in the division in the expansion of the yeast cell so for example proteins involved in protein synthesis in multiplying the proteins themselves are the most important proteins of this network structure importantly and interestingly these proteins are losing their importance in crisis because in crisis the yeast cell is top divide so it is not growing any longer but importantly the proteins which are degrading the proteins so doing just the converse of the previous task are getting the importance because in crisis you have to make an order in crisis you have to get rid of those proteins which become erroneous which were having a bad structure or were oxidized or having another problem or some other clinical change during the crisis this amount of protein degradation is becoming important interestingly a lot of survival processes are getting really important as well understandably because the cell has to survive this particular crisis now this again is a little bit complicated complex figure focusing your attention to this white take home message in a stressed yeast network the major observation was that the groups which are overlapping groups are not overlapping any longer to the extent as they were before the crisis during the crisis so in a crisis what is actually happening that a group is getting more condensed is getting more closed from the environment and intergroup connections are getting less in the crisis than they were before so before the crisis the yeast cell was behaving like a complex organized entity everything was connected with everything it was really a large traffic going through the various groups and they were exchanging various information and physical reactions so for these things in the crisis there are some remaining connections obviously because this makes the network a network but the number of connections between the groups is small now if you remember or recall our everyday behavior when a crisis is coming in our family, in our firm in our community then this is how we react at the very first moment to the crisis that let's get in closer let's try to help each other only in the very closest community let's close the family and close the family lines and help try to survive together in this small group so the behavior is pretty similar and one reason why it is similar that noise and damage cannot propagate from a group to another group if the connections between the two groups are not so many as before in a crisis problems are propagating all around in the cell it is free radicals actually because in the cell in a crisis free radicals are abundant and they are damaging the proteins if there is a group of proteins and it is connected with another group of proteins very intimately then the chance that these free radicals problems are processing from one group to another this is how this joint then the chance that the problem is processing from one group to the other it is small I will give you just a very everyday example if there is an an infection there is an epidemics going on then you impose a quarantine and you don't let your children go to school this is the very same behavior because you disjoint the group of the school from the group of your family and therefore you are not allowing the procession of the problem which is in this case an epidemic infection so these systems are fairly similar in their responses when there is a crisis going on but I would like to draw your attention to the second phase because we are very good to learn the first phase how we can close our communities when there is a crisis this is the first phase but there is a second phase so the second phase is to reconfigure to allow the system to have links between communities which were very distant from each other in the original network now this is a completely reconfiguration of the network and this allows the whole system to develop a new solution in the crisis now this is a point where we are not really good in society especially in Hungary, I don't know Austria but maybe you are much better in other countries but in Hungary we are certainly not very good in that so we may have a lesson to learn oh, I missed a few slides I just would like to show that this phenomenon is pretty general even from the scientific point of view Uriel and his co-workers were summarizing the metabolic networks of 117 bacteria which were living in different embryo the difference in the embryo met was how the embryo met was changing several bacteria which are here were living in another bacterium or in another organism so they were symbiotes they had a very constant embryo met because they were in the middle of another organism other bacteria were living in a very variable and they are not what was the take home message and what was the essence of this finding that the modularity the more separate modules was increasing quite significantly 0.25 to 0.45 and this was how large it significantly increased when the bacteria were transferred or when they were living in a constant embryo met versus very changing embryo you may recall this kind of yeast situation when the yeast had a kind of happy life everything was okay, there were resources that was not so much changed and the yeast had a low modularity so they were not so separated modules in the yeast cell however the situation was changed, there was a stress all of a sudden the yeast modules were becoming separated but I would like to recall that that was a protein protein interaction network and this is a metabolic network so these are two different network behaving the same way these are all data about the metabolic network showing the very same feature and this is a telephone call network obtained from other last time I was watching and these are telephone calls in an originally black one black line on the bottom the red line on the top is describing telephone calls at a bombing event at a terrorist attack what turned out that in an everyday life the network, your telephone connection network is having communities which are pretty much overlapping so you are calling distant people, distant friends and so on the contrary when you are in a trouble when there was a terrorist attack then you are not calling distant friends then you are calling your mother your spouse, your daughter, your son the very close people they are the closest to you and therefore the network community which can be observed during this is a very close community is a very dense community and actually these communities are getting very distanced, very discriminated from each other during this stressful event so actually social stress and quite abrupt one this terrorist type of event is producing a similar type of network change as the others I am just listing you here other networks like the brain, ecosystem other social networks which are changing similarly when the other resources are changing so the general take home message from all this segment of my lecture that if you have plenty of resources and you have no stress then the network is very close to a random network if there are communities well in the random network there are no communities really but if it's not completely random if there are communities they are overflowing each other they are indiscriminatable from each other the network structure is not really well defined if stress is imposed if resources are less then the network is developing communities after a while the communities they are getting distracted from each other and at the very end eventually these communities will break and so these communities will form sub-graphs or sub-networks and the whole connection structure of the network is broken this is an extreme stress or extremely low resources now this behavior can be observed at the phenotype as well so how the network behaves how the complex system behaves and very interestingly in many systems I will show you only two different phenotypes can be observed if there is a lot of resources as compared if there are very few very little resources the first network or first complex system I am showing you is ourselves human metabolism it has been defined or understood quite a while ago actually 10 years ago approximately that human metabolism can have two basic forms one form is when we have a lot of resources when there is a lot of food available this form we call a kind of large phenotype which is an overspending phenotype in this case human metabolism is not really taking care of the efficiency I mean digesting food only half because if we have food tomorrow it will have food in the evening it will have food every time so why to bother to digest it completely very efficiently so this is what we call the large phenotype on the contrary if there was family if there was a scarcity of food not if there was no resources for a longer time than a different phenotype develops which we call small phenotype this is a very swift phenotype that metabolism is digesting each bit of food because it is not quite sure that I will get another chance to eat something in the evening or in the next morning so whatever I have I have to use 100% efficiency or close to 100% efficiency interestingly these two phenotypes are changing very slowly you need sometimes as much as three generations to change from one phenotype to another but this might be a reason why in developing countries in the sense like India and China and major powers of the world by now when food was not available plentiful like 100 years ago but now it is not for many well not for everyone but for quite many people it is available diabetes and obesity is so prevalent because the metabolism is still strictly but the input is a large input or became a large input or larger than it was now my last second and last example is of Janos Kornai the famous Hungarian economist and Mr Kornai is famous because of many things but one thing certain that he defined socialism I mean the Hungarian system 20 years ago or more than 20 years ago as a shortage economy now in his recent work which was published now in Hungary and it is coming out in English so I am kind of warning you if you would like to read it then in a year and find it in the English market which is a source about capitalism Professor Kornai defines capitalism as a surplus society so it is very interesting that even in the society level you have this dualistic function a dualistic feature that if the resources are low then you have a society which is kind of swiftly which is kind of shrinking which is kind of having a very very short structure that is socialism and if the resources are not so they are pretty good then you have a society which is a surplus society which is capitalism so obviously these are very very distant kind of analogies I don't want to prove that scientifically I just would like to raise your attention that such interesting parallelisms may be found by thinking about the behavior of systems under different conditions but here you have to concentrate on the bottom of the figure because this is an important take home message I am here comparing two systems which are the kind of small phenotype system and the large phenotype system which are on the two sides of this bottom panel the small phenotype system is usually a rigid system which means the structure is very dense as it was in the dense communities structures in crisis and therefore it is unable to change so if you compare its possibilities to learn something to adapt for something and to remember to the adaptation so the effect of the adaptation which means memory type of women then if the system is very rigid then it is unable to change because it is very rigid you are not able to change I am really having difficulty if I would like to change this table this is pretty rigid but once it's changed already changed then it remembers then it has memory because it is rigid it can keep the changes so it's like another fact it is learning something to this large phenotype at an extreme form that is a very flexible system if you have a structure less system it has no modules if it has modules and the modules are overconfident you cannot discriminate them and so on it goes to a random network if a network is like that very flexible then to adapt to a situation that's no problem this growth can adapt to any situation this is flexible but to keep that situation to have a memory to really remember to their particular change that is really impossible I mean a growth like that it is not doing it it doesn't have a memory it is not rigid to the extent to have a memory now what is really useful what is really complex what is really helpful for the evolution that is what is in the middle that is a balanced system that system does have parts which are rigid but at the same time it does have parts which are flexible so therefore at the same time it can adapt, it can change but it can remember, it can stay it can memorize or keep that change for a while so the adaptation efficiency is pretty high if you would like to get a scientific treatment of this change is not in the sense of flexibility or rigidity but in the sense of robustness which has quite many connections with these ideas that I would like to draw your attention of this nature when I am talking when they actually proved in the simulation system that the systems with immediate robustness are adapting faster than other systems with extremely robustness extremely small or extremely large so going back to the previous example when there was a stress then actually the system was shifting its composition, structure and behavior from a very flexible system actually too flexible system to a complex system which is flexible and rigid at the same time and this is what we actually observe when the yeast was having a crisis so getting to the end of my talk I would like to just give you two take home messages and the very first take home message is that biological networks actually offer the experience of billion years of previous survival so it is worse to look at them because sometimes our own behavior in the society is not that kind of perfect as I would say as biological systems behavior is surprising because in the society we have a generation time of a few hundreds or at best thousands but biological systems have a generation times of millions if not billions themselves which is having a greater ground for experience now the second take home message is that community rearrangements group rearrangements network community rearrangements maybe a general mechanism for a system level adaptation because it is cheap if you have if you have a plan to change your system in a way that I will change complete parts of the system I will renew complete parts of the system that is expensive because then you have to redrop that part from scratch completely new if you have an idea that I do have the parts of the system they will pretty much be unchanged but I just reconfigure the connection structure of these parts that can give a number of new responses at the system level but it is cheap because all the parts remain more as the same this is what is actually used by the yeast cell when the yeast cell is reconfiguring the protein-protein interaction network major parts are preserved but the configuration the connection structure of these parts is completely renewed and this is what is giving another possibility for the yeast cell so this what we may also learn from such systems and eventually I mean this very last kind of pedal that I showed you before it is also part of the take home message that you may consider a nice and delicate balance between flexibility and flexibility in complex systems because most extremes are not really useful or not really helpful for changes or for remembering the changes in the longer term now in my very last slide I would like to just to acknowledge the contribution of the people of my room these are only the people who are taking part in this particular study including Gondrash, one of whom is in the audience and we have actually in symposium all this is the part of the of the room in this morning and obviously this group is multidisciplinary as I told you so it is open for collaboration so if anyone in the audience is that he or she wants to collaborate with us we are very happy these are the potential collaborators as Manfred told my book about that is really available so you don't have to pay for it anything which is in crisis a positive sign thank you very much present these networks in a very rich language which does not give us any information about how these networks become become encoded so what is the nature of the community system that generates this network actually I would formulate the answer or the problem in a different way we observe facts in this biological system as we observe connections physical connections of these proteins in experiments like these two hybrid systems or community precipitation systems so we do have a lot of data on these interactions and the coding is posed by ourselves from my understanding that we understand this complexity of data from the natural point of view which means that we define nodes in the networks we define the category which we consider to be a node which may be a protein maybe a person, maybe a cell in this case the definition might seem to be easy because ok I know what is a protein I know what is a person I know what is a cell if I know but many times this definition is getting very difficult if we got to the point that we try to define the interaction itself the edge in the graph that is getting even more difficult if we try to define the weight of this edge that is getting even more difficult I am just giving you one example about social networks they are working with many social networks but one of the social networks is the at house networks from the United States that is a social network about school children now the school children they are monitored for a longer time and sorry please yes I understand that but in essence how do we get meaning to the communication how do we code from your network oh I see thank you for the additional question or for the explanation of the original in a social network or in a neural network when neural cells or people are communicating we have everyday kind of concepts on the communication we have everyday kind of concepts of the information chain it is indeed more difficult to conceptualize the communication the protein interaction network is communication communication at the protein-protein interaction level is a physical type of communication these proteins are changing the shape all the time and these shape changes are not independent from each other these proteins are bound to each other in the cell either transiently or constantly so if one of them is changing its shape then the other one besides it has to change its shape as well this is a communication so actually this type of physical more less physical type of communication patterns whether I kick the neighboring protein or not this can be understood at this particular as a communication pattern without a code in the sense of Shannon or genetic code or these sorts of codes well certainly in from this sense of a view not such a well developed theory of coding at this level partly maybe because these systems are messy so not all very well developed and clear theories can be applied to these systems but in the sense of the connection structure of the connection data you have your real data and you are starting from your real data so this is the background I think first the one time that you were having all the feedback yes it is connected to self-organized criticality I can't really tell you at the moment how it is exactly connected to self-organized criticality because these changes in this systems I mean our current treatment or people's current treatment for these changes in these network systems are restricted from one step there was this normal state there is this stressed state now in the self-organized criticality you have not one step you have thousands or millions a number of steps so it's a whole continuous process once we develop data enough and we can analyze data enough for a whole continuous process I think then it will be a very relevant question what you posed by the self-organized criticality conceptualized together or not I think yes but this is kind of their feeling this is not quite effective I have a somewhat unfair question so of course natural technology is a wonderful approach and we have an excellent presentation we hope to use it for higher nature than a social organization but this is a conference on cyber medical system research which has a 40 years history and the history tells us cyber medical theory of living structures don't have to review thermodynamic approach open systems living systems as catastrophes and bifurcations so now we have natural technology so how do you what is the best way to analyze and follow these theories many times we live in one person's first school so how do you see the future of their exercise do you remain with us for the next 200 years or do you think it is just some kind of presentation of course patients come your school they help us and then something else will come thank you very much for this question I think this is not an unfair question this is a very good question indeed there were kind of scientific fashions regarding many many types of the theories what you mentioned or approaches what you mentioned and network science was kind of rising very fast in the last 10 years I think this rising in popularity is partly because we do have now data and we do have now capacity to store this data and capacity to retrieve this data so formally if you had this data well you had difficult before the computer science age which is kind of 20 years or slightly more actually personal computers are me everyone's success and biology especially has no data like 10 years before to the extent which would be understood or would be analyzed at the systems level including network I don't know whether I will be a network scientist 10 years from now so I was not a network scientist 10 years ago I don't know whether I will be 10 years from now currently I feel that there are many ideas, concepts and behaviors which you can catch by networks and which are robust but that's why I am very much interested in networks at the moment I don't know for how much more time you will be able to discover this very kind of robust general and interesting changes with the network kind of description at the moment the field is very rich I don't see any polarization of the field so the field is getting at the moment richer and richer more and more interesting for me or for the people who are in the network size obviously this will have a peak obviously this will have a kind of downward trade when, I don't know, maybe 10 years maybe 10 years I think you will be interested I'm sorry how do you determine the border of a network well that is a very relevant question I mean if you are talking about a cell then more or less you have an understanding about the border of the network which is this in the cell that is belonging to that particular network protein-protein interaction network or whatever which is outside but even in this case where the cell really ends I mean the receptors on the outer surface that is certainly inside the cell I mean it is belonging to the cell the proteins which are bound to the receptors are they part of the cell or are they part of the outer territory the excess of the matrix so I mean this is a very relevant question it's very difficult when you say the cell so you start from a nominal system yeah, yeah and what that will be interesting is when you see the network as you study it that you see that it crosses your initial boundary well I mean from the network point of view if you like so you don't have any general knowledge about the system you don't have a conceptualization of the system then you have to rely on the prototype structure then you have to rely to find groups in the system and if you found very well defined groups then you may identify them as subsystems or that you may say that this is the boundary of the group but in fact these groups are overlapping all the time so that is one of the most difficult questions of the network science at the moment where to find the boundary of the system only coming from the network information and not to take into account any other information ok, sorry in the introduction as mentioned you said a committee in Hungary would you like to comment on how this research is purposeful to inform public policy if it can be or if it can be the network science as such yeah I think there are many many social problems which can be conceptualized as networks like at the moment in Hungary and I think in many other countries we have big problems with health care health care is actually a network of different hospitals of different institutions patients themselves so it can be conceptualized by this point of view so actually it can help on a policy from understanding the complexity of the system they work with this is one thing the other thing is this kind of more I mean how should I say, difficult thing to make analogies and what I try to make in my lecture and try to understand the examples or biologically what the systems gives us in various situations I think this may also help the public kind of understanding the decision making if there are people who have a sense of that sitting at the decision making so this might be more difficult yes, please true it seems my watching often takes place they have a network they start from scratch they sort of remember it probably from the head state the when made the networks which are often critical of the day communication of the system and when research is being made they go from the standpoint that the the structure is isolated or random not yours and there's not any memory in the system there's also in the EU there's a class in research they show that they have a certain kind of memory a certain pattern of the disruptors they come from some parts of the disruptor now there is to expect a different kind of disruption they also seem to have a certain memory have you can you share some memory in the network we are currently making this research so unfortunately I can't give you data or real results just an idea I would like to enrich your comment which is a very key point I think that formerly 100 years ago engineering was more or less 800 years ago had only a few parts had a very defined interaction pattern which wasn't really changing if it was changing the system was broken at the moment we do have engineered systems like the electricity systems which are not of this type which again shifted from this very kind of rigid structure towards the middle of the structure when they do have those flexible parts flexible means here more expensive parts so now we are sacrificing a little bit more money not to get the most utility system which is possible the cheapest system it is not really reliable it is not having this memory it is breaking easily but we sacrifice more money and therefore we are getting a bit closer to the systems which have been involved in biology so I see now the two sciences biology and engineering are more in common to tell each other because of this kind of behavior and change the question is this network analysis is often done with a view of a big network we went on actions of global knowledge what are the possibilities of global assessment if you go into the network and the global assessment of the incubation of things yes this is a very common problem if you are studying large really large networks because you don't have the information we have about biological cells biological cells are pretty small from this point of view the yeast has only 6500 proteins and maybe 200,000 links this is nothing compared to the internet compared to the worldwide network or the whole social network of mankind or your brain now in these later cases indeed sampling of networks especially local sampling is a crucial point I don't simply are at the end of the story and that's why your question is so relevant how to find a good sampling system for the locality of the network to feel the complex structure of the network most of the networks many of the networks have a kind of more or less fractal hierarchical structure so you may have a extrapolation from the local structure to the global one but whether your extrapolation is really true well I mean I would be safer if you may extrapolate for three or more scales of hierarchy and then you extrapolate one more so the fourth scale or the fifth scale that is safer at the moment we are having one scale only this is the local kind of understanding of the network and we are trying to extrapolate to the fifth scale now this is not working so we have to get more advanced sampling and extrapolating structures I think you first please so the name is what happens if you really want to reject the quality is the technology silent how does that work it's a very good question actually I don't really have a good answer to this question my first answer is no that we don't really assess quality type of changes or interactions from the quality point of view in a different sense but I think that's my feeling that if we will advance in the dynamics in the network dynamics then we may have in the future some information from the dynamics from the refined dynamics on the quality I am just giving one example which is a very simple example and it's not worth well we are working with when you can define various social games like the prisoner's dilemma game the ultimate game and other games and we are using it for the interactions for the cooperativity of protein protein interaction networks now I start to have the notion that it is it is educative what type of social dilemma game you are using into the particular protein protein interaction network to the quality of that interaction so from such initial points you I mean one may end up not now later on to get closer to your question but at the moment I think we are far from it I am wondering with regards to combining with the question about the boundaries and the question about the memories looking at the epigenetic landscape how the system processes change taking this from what we are getting from the areas of profit physics looking at non-local correlations state correlations are you bringing this into there in terms of how the system actually aligns itself and develops its own the question of boundaries is there a role for the interaction? I think we are again my answers will be quite the same all the time we are getting closer to understanding from the network point of view as well epigenetics haven't been covered that well with networks at this point as far as I know because of the complexity of the changes and because of the complexity of the types of the changes which are this is the question which we discussed it is very difficult to conceptualize something in the network point of view when changes are different so we have this DNA methylation one thing histone phosphorylation another thing should I consider it the same type of link? this is when things are getting difficult from the network point of view obviously you have colored graphs and make up a network like that but the colored graph theory is not really at the moment of biology so getting back to your point this long distance interaction and kind of system level interaction from this point of view I think very well tackled by networks at the moment I would like just to refer to a long method that I showed when you have a local community you have a single node but that local community many times expands to the whole network not to the same intensity obviously it's more intense locally and it's less intense at a distance but there is another node it does have again a local community and these two local communities are overlapping each other and actually all local communities are overlapping each other so this is getting you a complexity of kind of long distance changes which you can formulate or rationalize from the network point of view at the moment so I think this is a very encouraging field what you mentioned from this point of view from the long distance complex interaction point of view because otherwise it's difficult to understand it but if you get a bit by bit for example by help of a network or other realizations then it is more helpful because your mind is too limited to understand it as such so we need help the network is provided would you rate your mind is the holographic information nestedness does that factor mean to the network of course is it playing with bombs and many other frames what do you mean by that sort of orders is that... if you start to think about network dynamics the only way now you can think about it so it's getting more and more into network science not each bit of that and not especially the papers which were 5 years ago or more but currently yes I think now last question I think because we have time for the referring to the the network that's had a dynamic control you know it's not always static way on the score is that coming from the work more on the current way how do you see it this type of redundant and degenerate kind of regulation is playing great role with major role in the biological system so yeah it is and others which are touching this issue I see thank you I think the way and I think