Phosphoproteomics for Analysis of Signal Transduction Pathways

you I’d like to thank Rob again and mark for the invitation and it’s a very nice lineup of speakers and lots of interesting focus on this field of interact on biology and I’m going to talk about a slightly different angle on the team and talk mostly about foster proteomics and what we are doing in the context of cancer and also normal signal transduction and I’ll just recap all the you’ve had an introduction just recap some basic aspects of mass spectrometry as they apply to global analysis today and then talk was stabilized to top labeling with amino acids in cell culture or silac and how we are using that to look at to identify therapeutic targets in cancer and to and using it for pathway analysis how we have used it to the sex signaling downstream of a novel cytokine time extramural Limpopo in OT SLP and then talk about some of our by informatics efforts net path hpid and human protein pedia as community resources and we are very happy that Mike Washburn is here and you know it’s because you know in part because of his major efforts in the market technology that there is no running away from the fact that LC ms/ms is here to stay and multi-dimensional chromatography and coupled to high resolution mass spectrometry that’s here today I think we can talk about where we can go in into the future and this is what can be achieved in an average at least in research labs where you have very nicely tuned mass spectrometers and by informatics efforts this is what one can do today and this was definitely not doable as recently as about three to four years ago where today we are actually approaching transcriptomic and genomic types of analyses you can actually attempt to analyze the whole podium of itself cell line or tissue by fractionating it key to mass spectrometry today and to going deeper into the analysis is being able to fractionate or enrich and this team will come over and over again so this is just one way to do things so you’re one sample the entire proteome of a cell line or two shew in one tube rapidly becomes 30 to 60 or 100 fractions so that means that’s 100 different experiments that can be done and each fraction can then be analyzed by LC ms/ms and you may sample this is just some rough numbers so you can think about this millions of peptides in the experiment and until recently I had this number which was still an optimistic number to about 2000 to 3000 but now some labs are showing that essentially you can achieve such really in-depth numbers where you could be wondering well am i talking about transcriptomic of proteomics you can get quantitative measurements on almost 10,000 40s today depending on how detailed your other steps are and that’s why I’ve written some ranges sample prep and fractionation may take you a few days and the mass spectrometry can take you anywhere from a week to four weeks or so to be achieving these kinds of depth that means that mass spectrometer then will be dedicated for that one kind of analysis and time for the data analysis it can actually take you just like many people who got into the gene expression measurement early on they would take months to analyze proteomics data isn’t that it’s really the early days of proteomics and you can do a very very nice experiments and keep mining the data for months one month is just one guideline but really what is catching on is and you have heard this over and over again for many different speakers many of them have focused on label-free kind of methods which is where the research is going but I’ll talk about other kinds of quantitative mass spectrometry where you essentially introduce a mass difference either in vivo or in vitro I won’t talk about most of them but mainly focus on one method called psylocke that I developed when I was in Mathias Mons lab in Denmark and this really if you are expecting the changes to be subtle this is I would say one of the better methods because of many of the advantages and and basically what ends up happening is you encode your mass difference into every peptide of every protein without doing any of the hard work or the chemistry the cell is doing it for you and you essentially you grow the cells in heavy amino acid containing medium and you can use many different labels depending on which four years you’re going to use if you’re going to use trypsin you grow it in heavy arginine and lysine and I think Jeff talked about many of his experiments where they’ve

used Psylocke and you have a control situation you have some activated state you mix samples this is just one way of doing things where you can run a 1d gel you can do an in solution digest scx fractionation name it but the advantages from the ms/ms data you get the identification of the peptide map it back to the protein but then this is the finding the needle in a haystack where the ratio of the peptide signals that is in the MS can actually tells you what is it that you are looking for and in many of the scenarios in signaling or in biomarkers the kinds of situations that we are interested in is not the abundant background which is no difference between the two states but one case one out of a thousand out of a hundred out of a turn or out of a million depending on which biological context you’re operating in where you really want to zero in on differences at the MS level so it’s a very simple in vivo method it doesn’t require any extra processing steps it’s uniformly labeling of proteins and you can choose amino acids and we have developed a website Sarlacc dot org where you can go and we try to keep it updated on many different papers that have come yes do differently we take the c-13 containing one and and different people are still devising software I think what we are really excited about is and this is where the development has been the least is for the data acquisition software to ignore the signals from most peptides in real time which have equal abundance and then specifically fragment is this is something that you know we are yet to see we are working with certain vendors but they don’t seem to realize that that is really the next level that will really revolutionize and save time in a given mass spec run but that’s unfortunately not there although there is activity in that we have now you can compare five different states at the same time by multiplexing different masses of these amino acids so that’s the background you know we can use silac as a way to do quantitative proteomics but we want to chart interaction so you want to look at pathways in cells and how exactly can we start doing this well one of the ways and you heard about this is you can determine ordering of the components protein-protein interaction studies you can determine the interactions but this is something that and Claude talked about mapping of enzymes kindnesses and phosphatase in her case and also in the kinds of examples I’m going to give you in signaling is a key step and this is not an easy one because kinase substrate interactions are transient state cometary of phosphorylation is low and I tell this to people that you know it’s very very easy to do a global analysis it’s almost that when we end up publishing papers we are misleading many members of the community because they take that a little too seriously and apply that to their favorite protein Global proteomics is easy when you’re trying to find every site on your protein of interest suddenly things are 10 200 times more difficult so analytical methods to identify phosphorylation sites when even when you know there there isn’t they’re not easy you can do global analysis but deep and comprehensive that’s a different story and then finally establishing a protein as a direct substrate is actually not trivial and by definition you cannot do that in the context of a cell because in a cell you have other intermediary pathways and proteins so you have to choose the situation which is in a test tube where you mix a kinase with this potential substrate well you have proven this is a substrate but now this is an in vitro system and therefore it’s prone to errors it’s not the gold standard you go back to cell it’s a catch-22 of anything to do with these kinds of PTMs I’ll talk about how we are using these charting of pathways for personalized medicine and they’re a very small number of examples out there where it is possible to do some sort of a targeted therapy and it’s mainly known for certain cancers and Gleevec was one of the early ones where it is targeted against the kinase that that’s active in chronic myelogenous leukemia and one can wonder how is it that we are going to deliver on the promise of personalized medicine especially in the context of cancer how can we find the next glimmer that’s the kinase inhibitor against bisa against Abel and the key is not in developing more of those small molecules many small molecules against most kindnesses are already being developed by the pharmaceutical industry what we really need to find to deliver on the promise of this personalized medicine is we want

to find therapeutic targets in individuals with cancer that small molecule is being developed but we need to find which target is playing a role in the causation of cancer in which human being and I’ll illustrate this with our study on pancreatic cancer as a prototype we are taking a proteomic approach so for many of these small molecules that are being developed there again Student Union or tyrosine kinases and our goal is to identify activated tyrosine kinase in a subset of cancers by comparing the phosphor podium we are doing this as proof of principle studies in cell lines these are primary cell lines from patients with pancreatic cancer these patients along that but we have the cell lines where we can study and our idea is that although they have the same label to the pathologies they are all pancreatic adenocarcinoma but what’s happening inside the cell in terms of which pathway is active is different and we want to find those activated or hyper phosphorylated a key feature of kindnesses is that when they are activated they auto phosphorylate and our goal is to find these hyper phosphorylated kindnesses that are specific to cancer cells and then figure out what they are doing in those and we are using the Select approach where we take men carry cancer cell lines compared them to their normal pan Carrick epithelial counterparts and then look for this type of a signal where we see more phosphorylation in the pancreatic cancer cell and compared to a normal pancreatic epithelial cell and many people forget that cancers are heterogeneous once they get a label so we may think about breast cancer well we have the hereditary and sporadic but all the sporadic ones are not the same here this is just one way to profile the proteomes of these individuals with pancreatic cancer on the left h pn e and h PDR to normal pancreatic duct epithelial cell lines all the others are from different individuals and this is a Western blot IP and Western with anti phosphotyrosine so we are looking at a snapshot of what is happening in terms of tyrosine kinase activation the tyrosine kinase is and their substrates first thing is even a cursory look tells you they are different we didn’t need to sequence the genome to figure that out and then this is one of the normals even the two normals are somewhat different some of them have bands and that’s all we can call this is a low resolution method it’s a 1d gel they have some bands that could be one protein or many proteins migrating at that position that the hyper phosphorylated and then there are others were remember it’s a tyrosine vessel what there is no extra phosphotyrosine content in many of these cells and some of them looks somewhat like each other in general they are pretty different and what we did was zeroed in on this one cell on P 196 as a porta where we see robust activation of multiple different kindnesses and substrates we don’t know what they are but that’s where mass spectrometry is a great tool because it’s an unbiased way to figure out what is it that we are dealing with so we decided to compare this with this using silac this was gone in heavy medium this was gone in light medium and we carried out enrichment with anti phosphotyrosine antibody purify the proteins run them on a gel cut all of the bands out and did ms/ms and these are three different panels that are put next to each other showing that we saw this is the from the pancetta cancer cell line this is this normal counterpart we see increase of EGF receptor EGF receptor pathway substrate 8 and EGF receptor pathway substrate 15 this is heavy light heavy light so we have a control built-in that’s what I meant by needle-in-a-haystack because in any of these kinds of affinity enrichment methods and you have heard many of the speakers talk about this there is a certain background here the background is being used to compare to the activated state so these three Peaks are coming from the pancetta cancer cell line that we want to study these are the quantitation ratios for each one of those proteins and what is basically turning out and these are the ratios I’ve colored the proteins that we identified as being more abundant in anti phosphate are seen immunoprecipitate indepent carry cancer cell line versus control so you can see starting from the most upstream kinase here EGF receptor we see many of the substrates in this pathway lighting up so it seems that EGF receptor is probably the most upstream receptor that is active in these cells and we went back and we use more conventional methods until now we are using mass spectrometry now we would like to see what is happening to the protein levels and their phosphorylation status by ways that you know many biologists can interpret more intuitively and what you see is this is EPS 8 and different

proteins the top panel in each case shows phosphorylation in that cancer cell line and it’s more in every case but what is most surprising here is that the EGF receptor the protein level itself is not changed I’m saying surprising because often you find in the biomarker feel pathologists love to take antibodies and just throw it on their sections immunohistochemistry type experiments and say this is involved or overexpress what we are talking about is not over expression as in her2 gene amplification that happens in breast cancer that is something that people have taken maybe too dearly to heart and maybe it’s sending most of us along the wrong what we are looking for is activation of molecules activation of kindnesses we don’t need to invoke a change in protein levels you can also have situations we have protein abundance maybe change but if the kinase activity doesn’t change it’s not active if at all there is less EGF receptor in this cancer cell line than the control but you see this abundant hyper phosphorylation what we have not figured out in this case is why is that happening you can have many different mechanisms we have gone back to this cell line and sequence the coding exons and there is no mutation in EGF receptor which would be the simplest explanation so we don’t know the cause we may have increase in ligand we may have mutation in some other proteins that control EGF receptor activity or something upstream of that that means a phosphatase possibly that negatively regulates EGF receptor may be involved we don’t know the full answer but we know that EGF receptor is involved and when we go back to what we did was we took this cell line and many other cell lines from these patients and we grew them as xenografts in mice and initially we are very disappointed because we pulled out EGF receptor which has really in in some ways you can say beaten to death by many people proteomics and otherwise it’s really a well studied pathway so we’re disappointed but the good news here was they are inhibitors small molecule inhibitors like a lot in him that are available that we could use to test to see whether this cell line is actually going to respond to inhibition and this is showing staining with a different reagent now a forceful EGF receptor specific antibody that’s this brown color in this case shows positivity this is one other cell line where we did not have any mass spec data that also shows staining and these are two different cell lines do you must drive from those two cell lines where we don’t see any staining and when we treated these mice bearing these tumors with our lot unit what we can see is this is where I showed you the mass spec data we see that the tumor dresses in that and this other cell line where we see positivity but not in these two other ones that means now we have a higher throughput way of again now we are lucky with one other reagent we have a phosphor GF receptor specific antibody so we use mass spectrometer as a discovery tool we figured out that EGF receptor is active here and then we go on to use small molecules and other kinds of reagents to figure out whether they are going to respond or not oh well and whether we can find out if they have activation of EJ receptor in a high-throughput fashion we also had access to other xenografts that our collaborator had actually published they had data on this where they were testing many different drugs on a panel of pancreatic cancer cell lines and what they concluded from their studies was that EGF receptor inhibitors such as a lotta nib have no effect in pancreatic cancer we went back to the data that they had and we noticed this is basically their graph one of them seemed to be responding to a lot in it but in the context of the big picture it was left out and it was ignored we went back to sections that they had from each one of those tumors and the one that showed this response was the one that showed positivity that means now we have a predictor of potential response to a small molecule inhibitor which was hidden in the data and there are many studies that major medical centers have done in the context of pancreatic cancer and they continue to do that in the context of other cancers what is really unfortunate is that if you’re using a small molecule shouldn’t you be looking at its target and whether that target is active because if the denominator here is this big number only a subset of which that respond to the small molecule the whole stage is biased against you you are going to conclude that your drug

doesn’t respond when reality 100% of those cases are actually responding and I think this is the kind of a paradigm shift that we need to have in the so many of the kindnesses and just to show you that this is not really peculiar to Pam Carrick and so there’s nothing strange about when Kerri cancer we are systematically studying many different cancers it’s a fatal cancer gastric cancer and breast cancer and we essentially see the same thing the top panel shows you again the same experiment now auntie phosphotyrosine Western body is just we have one key and we are trying to figure out how many different locks it can open and you can see in many of them you see no signal that means there is no blatant upregulation of tyrosine kinase activity but look at the others and when you this is a small set of triple negative breast cancers for which there is they they are really tough to treat and we have no handle on that and how do we define them it’s really amazing how in 2009 it is still by exclusion the reason it’s called triple negative is some of the early markers that were identified are absent on this not to say that when those markers are present we have made such wonderful breakthroughs that’s estrogen receptor progesterone receptor and her two three markers and that’s always we are talking about not the activation of those pathways but the amount of protein that is encoded by those three genes here they are negative for those and yet you can see that they are heterogeneous even at that level because these are all grouped together as one unit now and you can see molecularly and it’s possible that in some other ones it’s a serine threonine kinase mediated pathway that is active so we are basically now taking this large fan of cell lines and we are trying to find out what what exactly is going on inside I’ll change gears and talk about our studies on a ligand or cytokine mediated signaling scenario time extreme Olympic waiting is novel it’s amazing and when I tell you the signaling story you’ll realize we are calling it normal all day two has identified almost 10 years ago it’s because it’s still not a famous cytokine and not much is really known but what has really happened in the last 10 years since the cloning of the cytokine is that it is a central player in asthma and other kinds of allergy when they were preparing a patent application I was involved in cloning the receptor I was almost laughing because the lawyers that put in everything that this is going to do everything including Allergy Asthma they just named every immune disorder and I was laughing because basically these are templates you can put any cytokine any receptor and it’s the same cand text about ten years later I was to be proved wrong from my skepticism because this cytokine in contrast to many other cytokines is actually very important and a center pair in asthma people have shown that in many different mouse models and and also human studies now but not much is known about signaling if we know what is happening intracellularly we can use those signaling proteins as potential therapeutic targets so what is T SLP and how does it signal well we know about the receptor subunits I cloned this in 2000 time external reporting receptor t SLP receptor d SLP engages aisle 7 receptor alpha chain which is shared between aisle 7 which is a more famous cytokine that’s why we still call it relatively new although it’s 10 years old it uses this combination of subunits when aisle 7 uses this and basically this is what was known in 2000 and in 9 years the amount of progress that we have made as a community and it also talks about the sociology of science who studies these types of cytokines in those kinds of cellular immunology asses they’re generally not biochemists and also the tools are only now becoming very powerful to be able to do this in 9 years we have figured out that T SLP phosphorylate stat5 transcription factor and a KT that’s it and there are lots of big holes and what what is known for aisle 7 but many other members of the aisle 7 family of which interleukin 2 is a prototype is that Jack family these are central molecules that take the signal from the juxta membrane domain go through stats and other molecules all of those were black holes here after nine years of research and in fact what is amazing is in this particular case because people are still keen on using the tools that they had jack family kindnesses were shown by

many groups stated in print that they are not activated that data was based on western blotting with anti phosphatase an antibody where you enrich those jak kinase is it’s a standard as every by chemists uses it so you enrich for those proteins and you Western law with anti phosphotyrosine and you don’t see any difference you say it’s not activated and then but the scientists now say it’s not the jack family kinase that is responsible for transmitting the signal we should try to find those kindnesses and we’re also very active we thought wow this is a great chance at trying to do something major for biology so what happens to the other siblings of this TS LP aisle seven family this is the the most well studied protein in that family interleukin 2 signaling and it shows this is what we know jak1 and jak3 kinase is transmit the signal through stat5 stat3 and all these other kindness this is what is known after many decades of research on this really a prototypical cytokine il-2 this is what we see when we started to do the biochemistry and I think in the context of protein interaction you have already seen at least some pictures like this where if we activate cells with TS LP and use the tools that we have anti phosphotyrosine antibody we basically see no difference now does that mean there is no difference well we don’t know but when you do another kind of an assay once you know what molecule you are looking for look at this here this is enriched that means immunoprecipitate stat5 and you can detect it now you ask the question is that five tyrosine phosphorylated the answer is yes but it’s just hidden behind the bands he said less abundant molecule that is getting phosphorylated but if you had to just go by the see of signal you’re not going to see so in that background what we can now do is we can use psylocke that’s the power of this quantitative proteomics when apparently you don’t see any signal you can go in and try to figure out in one step what are the kindnesses and the pathways that are being activated by this and here basically we take a cell line that is a reporter cell line that responds to TS LP and we split them into two one we start going in heavy amino acid containing medium all the proteins are labeled then we mix them because we always know that the heavy peptides are coming from the TS LP treated cells that the same cell and now we can take this very complex peptide mixture the millions of pepti that we are talking about and we can do many different things and and now the previous experiments I showed you we were doing enrichment at the level of proteins now technology this is you know one or two years into the future now I’m going to show you data where we are doing the capture of phospholipids we have already done the digest here all those peptides of which most of them are non phosphorylated are there and we are going to capture anti phosphotyrosine peptides using an antibody and on the other side we are going to use in the scx chromatography alone or in combination with titanium dioxide and i’ll show you just some represented data all this is and this is all unpublished and it’s very recent so we’re going to capture the phosphotyrosine peptides from an antibody IP and this is other kinds of enrichment principle and our goal our objective was to find a candidate kinase we wanted you know we knew it was not going to be jak 2 and this is now partially so we we found hundreds of peptides phosphorylated peptides tyrosine phosphorylated and I’m showing you the tyrosine phosphorylated peptides here and the ratios and most of them don’t change because just for the cell when the not activated by tea SLP these are still getting phosphorylated for the normal growth but we are we don’t we don’t care about those we want to figure out what are the kindnesses and their substrate that TS LP is going to induce and this is what we find see if we just look at the top two candidates at 5a and stat5 B you see that wow we are on the right track that already tells it if you look at the bottom this is just a small collection we have two but most peptides are in that group CD k3 and GSK 3 alpha these are this is a doubly phosphorylated the singly phosphorylated this one-to-one ratio and we have found many different adaptor proteins and tyrosine kinase –is that are getting induced when you add to your selfie and until now a lot of the data that you may see from biological experiments is usually dependent on Western blot with anti phosphotyrosine which is not

strictly linear so these numbers may not make intuitive sense to you but these which are amplified signals on the substrates of kindnesses stat5 a molecule is between 3 and 10 so and let’s look at this jak 2 we found 1 point 8 fold difference on this side and 1.4 there and look at Lin which is another tyrosine kinase this one peptide is actually down regulated and a different site is up regulated so we have found biochemical evidence by quantitative mass spectrometry that jak 2 is forcefully we went back and redid the Western blots one more time you cannot catch this difference that one point eight to two fold you cannot catch it and we have repeated the experiment we find the same thing we have gone on to do functional studies now we have used jak 2 kinase inhibitors and we are now doing SI RNA mediated experiments to show that what is driving proliferation in these cells is jak 2 and this is 10 nano molar which is very good for this inhibitor that we use in these studies Lin I pointed out that we captured two different phosphorylation sites one of them was at the c-terminus and this is the picture that we see this is unendorsed and this is TS LP induce this is reminiscent of what is seen in sock kinase which is a prototypical member of this family where the c-terminal tyrosine is an inhibitory tyrosine for that kind is to get active the c-terminal phosphorylation is suppressed or is d phosphorylated so we CD phosphorylation of this and we see hyper phosphorylation of that this peptide comes from that this peptide is here so Lin is another tyrosine kinase that’s getting active and BTK is another tyrosine kinase that we’ve captured in that one experiment single experiment we have also these our initial data again and these are some of the data that we have from looking at Syrian threonine phosphorylation triggered by TS LP and now we are talking about a different class of proteins and different magnitude of change tyrosine phosphorylation was more subtle here we are seeing differences at the level of that foster peptide the ones in red are the ones that we found by ms/ms to be phosphorylated tenfold four fold and and again hundreds of them were not changed so the majority of them are unchanged and we are trying to put all of this data together in some sort of a pathway picture but most of these cytokines the major action is to Tarson kinase a and we believe that we have already found them this was the picture of TS LP signaling when we began and this is after many decades of studies these are the critical molecules in il-2 signaling this is summarizing the data generated from many different platforms many different labs over two decades today this is where we are it’s not too bad for TS LP where we still I yet to connect them that’s where we have to go back to the drawing board and to be able to connect them and to find other molecules is going to take a lot more time but we have already found many of these molecules and we can now carry out functional studies using never that again Jack – and even Lynn and BTK and tech and other kindnesses for which we did have biochemical evidence and the lesson that we learnt and I think maybe early on experimentalist er taught this that when you don’t see something you just say it was not detected as opposed to it’s not happening and that’s really the mistake we have made because every analytical platform has its own weaknesses we have to be careful in negative data if people had said not detectable which is really what we can say because somebody can come with another condition under which you will see activation so here that’s one of the reasons why maybe the field got to be misled and it ties in with other kinds of asses that others have talked about you can have multiple reaction monitoring now where you can just analyze only that one side you don’t people should need to go through this entire program experiment to monitor activation of jak kinase you can really monitor what full activation you are seeing at the level of phosphorylation site abundance and it’s possible to take unbiased discovery approaches like this to detect signaling pathways in a global fashion this is I’m going to show you a few slides it’s a and adapter molecule that I found in a proteomic screen when I was in Odense I called it Odin as an adapter molecule that that was phosphorylated by addition of EGF to heal ourselves and we have data from knockout Mouse that this is a negative regulator of proliferation but we really

don’t know how it works and we did we flag tagged it express it into nine three cells and these are the inter actors that we pulled out where we basically saw no signal from the control cells versus Odin transfected cells it’s a long list and I’ll just show you what was known or what is known about this is also unpublished the interaction network of Odin and we are hoping that this will shed light and we can figure what is it doing how is it negatively regulating EGF receptor signaling but this is what was known it binds it’s been shown in another higher throughput screen to bind to one of the members of a 14-3 three proteins and I’m not going to convert that long table that I showed you into this there are many different proteins some of them are adapter proteins this is a gdp as activating protein there’s a rust like protein people don’t know what that protein does some cytoskeletal proteins and enzymes in that one single silac experiment we have filled in all of the blanks but this only sets the stage for doing many other experiments and these proteins were already connected and in this Odin IP we basically identified all of those three and you have seen this kind of a team occur repeatedly too many of the talks today so we are now working on we have validated many of these types of interactions and are moving on so this is great we have no doubt that we can publish all of this data but the problem is although we were careful in going back orden is goes by many different names many of them are call it a kiya protein 0 to 2 6 it’s just an accession number now how do we figure out what is the published literature on Odin it’s nice that we can go into supplementary tables and figure but that’s really not easy for most people to do and we would like to share our data but also make it easy for people to be able to figure out when their findings are novel and what exactly is new and you asked the question about this literature search and I think it’s a big one because many of the experimentalists they are dying to conclude the day the results come out whether what they found is novel or not the problem is some some time ago maybe people were more cautious and they would say things like to the best of our knowledge this is novel but now because we have to be more and more aggressive to sell our stories so much we would like to just say no if this is normal but we don’t know there’s so much of literature out there most of it is hidden in the supplementary pages we would like to make sure that other people don’t spend their valuable dollars rediscovering these things or they take this to the next step because we cannot pursue many of these kinds of things in the context of TS LP what I showed you in the to context of pancreatic cancer many of these pathways it is experimental data and we would like people to be able to benefit so I’ll tell you some of the mini stories about some of the tankless things that we are doing it’s really they are definitely underappreciated at this point of time and maybe in this room we have a critical concentration of people who might appreciate these kinds of resources more than average but still I think there is a lack of appreciation for what we need to make systems biology happen and this is one of the examples that I’m going to give you net path which is the public resource of signaling pathways that we have started to develop because very quickly we have come from single genes to pathway analysis and we would like to be able to say which pathway is activated and this is our modest attempt and initial goal is to only work on ligand receptor type of events that can be clearly demarcated as a pathway and to have a detailed collection of all published reports or pathway reactions to and initially we have annotated them the pathways that are that have more relevance to cancer in immunology and to involve the community and TNF pathway many of you I think there was some biomarker studies about TNF that were mentioned but really when you go back to all that has been published on such a well studied cytokine it’s something like this it’s almost unmanageable even when you zoom in there are so many molecules that have been shown to be in this pathway that have talks and cross talks in so many ways that to think that TNF receptor cygnus is all the data that is published up to now there are other studies that I’m aware of where a thousand more proteins are going to be added to this pathway 1,000 we should not simplify biology any simpler than it is it is complex and we need to try to understand that biology and this is how detail some of the pathways are all derived from experimental data things are annotated

in this map only when they change when cells that Express TNF receptors are given TNF that’s the context yeast two-hybrid kind of experiments are not here TNF has to perturb the system this is how detailed it can be so this is a snapshot of the home homepage of net path where we have done some cancer and immune signaling pathways and the idea here is not just the physical interactions and the kinase and other enzymes and their downstream reactions but also to compile a list of genes downstream genes that are differentially modulated by this extracellular stimuli and also a protein translocation events I’ll just walk you through some of these pictures so this is EGF receptor this is how many molecules 177 molecules as of the date of this curation were already described to be downstream of EGF receptor so we think it’s only 20 molecule that’s a fallacy that we have it’s not what biology is trying to tell us and then we have experts this is physical interaction category catalyst we have detailed comments written by curators and transport where when you’re idg of these proteins they transfer gate from one compartment of the cell to another and these are transcriptionally up regulated genes so that’s pathways and that we have built on top of this research that we began almost seven years ago as a non redundant database for human protein sets that’s the focused human proteins and our goal was to have a repository which would be a comprehensive repository of protein interactions and other special features of proteins like post translational modifications substrate expression tissue expression and enzyme substrate relationship and when we began this and even today what is very interesting is for many protein protein interaction repositories kinase substrate and these kinds of enzymatic relationships are not considered or classified under protein interactions although they are definitely not the most exciting interaction that you can have in a Cell this is the homepage this is a molecule page in HP Rd if you are already not familiar with this these circles are phosphorylation sites these are protease cleavage events and what we realize after many years of curation is that we cannot keep up with all the literature that we must involve the community directly and we started an initiative called human protein pedia where experimentalist could contribute their data directly either through web interfaces or by interacting with us and here you can see that this is the kind of metadata that can be captured from a mass spectrometry experiment and then it is also available if people have contributed their data that goes along with that experiment you can even inspect the ms/ms spectrum in this case and for this I have to thank the pro uux community that really came along to help us and Rob Rob actually worked with us he was one of the very enthusiastic parties to contribute data in this and the net path that we have that I described was done in collaboration with Gary Bader who is also here and these are the people in my lab who contributed to the experimental things and this is the nonprofit Institute Institute of Informatics that I founded seven years ago in Bangalore and that name itself is a misnomer really because it is a systems biology Institute institutional biology was just established in 2000 and this was founded in 2002 so that name it was a very catchy name was already taken and what we said was initially we are going to develop databases and we are going to focus on by informatics with a clear vision that we want to do everything and now we have a genomics lab so they do a lot of microwave type experiments and they are doing a lot of mass spectrometry now and we are very excited that they’re joining hands with not just our lab but many other groups in different places to do many of the thankless things that I think are really needed but I think the community has still not come in ways that it should we all focus on data generation and not so much on data sharing but I think that’s really a weakness that will be fixed hopefully in the years to come and thank you the question I had was what kind of heat is used for enhancing yes so so if you take some of the early fractions that come out of scx columns they are enriched in serine threonine basically all phospho peptide species and because sintering in first polish is abundant

those fractions are automatically enriched and then we use titanium dioxide to enrich and you’re right I mean the game is in Richmond if you don’t enrich you are going to miss many many of those events and we are still looking at the tip of the iceberg so you can take those Essiac fractions then go do titanium dioxide on each fraction so it’s a lot of experiments but that’s the only way to go deeper we are doing about 50-60 fractions in all of those cases the initial once we analyze without in event titanium dioxide in the first pass which is what has been done so far so the titanium dioxide part has not even been done for those and so what I’m showing you is some things that you’re already finding to be regulated so we have probably identified about 500 or so sites and those some of them are very nice adapter type molecules that we can find in that fraction