 Great. Okay. Well, thank you very much. I think it was very fortuitous that the organizers picked this afternoon to give the presentation because I get to give some good news. They did two hours ago. We got the final email that the gastric papers formally accepted at nature, so so that means Actually, had it been the opposite, it would have not been fun to give this talk, but anyway, um, yeah So let's see. How do I do this? Perfect. Okay So this is the this is the most important slide of the talk, which is to thank all the people involved in this project We had a wonderful group of people both from the TCJ side, but also the huge group of TCJ volunteers people who were involved in this project despite lack of the funding and so forth that really made this work and so Moving on. It's a little bit of background. So Stomach cancer is a bad problem. Stomach cancer is one of the top three causes of cancer death across the world killing over 600 or sorry over 700,000 people every year Now there's two main categories. There actually was more than two main categories But from the eye of the pathologist, there are typically two main groups that gastric cancer stored into the intestinal type tumors and the diffuse type tumors and This will be important for some of what we'll talk about later here But the intestinal type tumors are a more typical adenocarcinoma. They grow in glands and they have more of the, you know, the classic cancer associated cancer genome features the Diffuse type tumors in some ways resemble the lobular type breast cancers that we've heard about earlier. They have a CH1 alterations that grow very invasively and they lack co cohesion Except but compared to the lobular breast. They have a much worse in survival Now this is one way of thinking about gastric cancer But actually and within gastric cancer is a big problem that there's a lot of different ways to think about gastric cancer I mentioned that the two forms of histology and intestinal cancers and diffuse cancers people are also argue a lot about cancers in different part of the anatomy of the stomach are tumors of the gastroesophageal Junction different than those of the body or the Pylorus people argue are the gastric cancers in Asia different than those in the in the in the West are the You know, then you have the MSI and the, you know MSS the rb2 is you had all these different kind of ways to think about gastric cancer But importantly when it comes time to actually thinking about how to take care of patients and how to do clinical trials You know all this gets ignored and essentially if you were to listen to a clinical trial Or someone talking about how to treat gastric cancer They would say we did a trial of x-drug in patients with with gastric cancer They wouldn't talk about any of these kind of variables and so invariably then the next sentence would be we did a trial And it didn't work because of course you did a trial in a very heterogeneous group of patients and often without even thinking about a biomarker and the trial is sort of doomed to fail from the beginning and so with that as Prelude we wanted to think about this what our goals were with this project Now unlike many TCJ projects that were sort of drilling down on a particular subtype This project really embraced gastric cancer fully and with that our goals were to see if we could come up with a better way to classify these these These cancers not only have tried to classify them, but try to do it in a way that's Relatively clinically applicable and you know that doesn't involve, you know a million dollar work up to put someone in a Category and as we do that can we try to identify some of the key pathways? They're active in different classes of these tumors and then within those different classes. Can we find candidate? Targets and biomarkers that can be a foundation for for further therapy Oops this slide always gets messed up. So let's skip that So then As we aim to have this as our goal There's a couple ways to do that, you know one would be to say okay We have some prior ideas about gastric cancer We're gonna divide them by intestinal and diffuse for example and then we're gonna use that as our as how we're gonna You know organize the paper But we decided rather than doing that let's say let's follow the results of the data. We have all these data Let's see how the data self-organize and then try to you know from there figure out how we're gonna move forward with this analysis and so what we did then was use two different forms of the classification one was the Cluster of the clusters that we talked about or heard about earlier and the other was eye cluster So these are two ways of seeing how the data organize But then following doing that the goal was then to for us to try to identify some key features that mark the different classes of gastric cancer and then use those key features then to create a way to classify the cancers in a more simple Method and so basically so we would sort of use you know the total Data to find out how the data self-organize find key features that go with different classes of tumors and then use those simple features to make a simple Way to classify the tumors moving forward and so the goal then is would use a classification that can be more easily applied towards people in real life And so this is the first way that we went about doing this the cluster of the cluster assignments I think this is should be a pointer. Yep. And so when we did this I mean the key result here I don't want to go into all of the Details but the key result that as you started doing this you found four clear groups, you know One group was dominated by tumors that were MSI positive one had mostly diffused type tumors One was dominated by the tumors that were positive for the EBV Epstein-Barr virus and one had tumors with really wild and wacky copy number profile Profiles and again when we use the eye cluster approach here. We had five groups, but it largely capitulated what we saw previously I'll just zoom in a little bit here So this first group here were dominated by the diffused type tumors This is Lauren so Lauren diffuse and black again We had a group of tumors that were mostly EBV positive We had these two groups here that were very similar that had a lot of aneuploidy and then again over here We had a group of tumors where the MSI positive tumors Sort of form their own In the cluster and so seeing as that these were some of the features that really mark these different classes of gas or cancer We decided to use this as a way to form our Are the categories moving forward so after all of this we then reclassified the tumors by this Simple metric so first we had our 295 tumors and we pulled off the EBV positive tumors Then what was left we pulled out all the MSI positive tumors and then with the 205 left We separated those into the aneuploid or the chromosomally unstable and this last group here, which was Genomically stable tumors Basically meeting ones that lacked EBV lacked MSI and lacked aneuploidy So what we called that more had the more more flat Profiles now as we did this this was not independent of how we thought about gas or cancer previously so just looking at his histology first the most striking feature was that the Diffuse type tumors were overwhelmingly present in this group here these Genomically stable tumors, but it wasn't absolute So there were tumors that of the diffuse histology and EBV MSI and in the aneuploid group and not only that There also was clear differences in the spectrum across Anatomy so if you go in the G junction you get many more aneuploid tumors as you get towards the pie Loris you get more MSI and Diffuse or stable tumors in the middle of the stomach you get more EBV Okay, so Does this help doesn't matter? Let's give an example here in the EBV positive tumors So what made EBV positive tumors their own group was mostly driven by the strong methylation differences, this is a sort of agnostic methylation and the clustering led by Toshi and Hui and when they and when this was done, what's seen is that the EBV positive tumors form their own Separate group which are distinct from the hyper methylation seen in the classic simp or MSI tumors and these MSI tumors largely have silencing of MLH one, but none of the EBV simps have that And so this helps helps us define this as a distinct group And so do we learn things from that information so the answer is yes so for example for some time it's been known that a certain percent ten fifteen percent of gastric cancers have activating mutations of PI three kinase but by by creating this way of Categorizing these tumors we see things we hadn't seen before such as the fact that the EBV the EBV tumors have an overwhelming percent of mutations of the PI three Kinase actually overall 80% of them have PI three cas mutations with a roughly 70% being at sites that are cosmic sites or Recurrent sites whereas you go along to these sin tumors. It's down to three percent That's something that hadn't been previously noted Now also we could try to find new events not just Or in war about the events we already knew about like PI three kinase So within our copy number analysis for example, we saw an intriguing novel peak that was on chromosome 9 Which we initially attributed to the gene Jack 2 these Jack 2 is a tyrosine kinase and so you know It's an amplified kinase that makes a lot of sense Well it's quite Intriguing those when we looked at the different subgroups that this novel Jack 2 event was very enriched in the EBV positive tumors present in 15% of that group stronger than even the Annuploid tumors where the other events were much more enriched Now we drilled down a bit more into the genes here And so on as you zoom into this little piece of chromosome 9 again You see Jack 2 but there are also other genes in this event a couple of them have rather Unassuming names such as CD 274 and PDCL 1LG 2 the kind of names that you usually Ignore in a list because the names are so boring But when you look these actually have quite exciting names not for the gene but for the protein because these genes encode for PDL 1 and PDL 2 so these are very well-known proteins now which are Immunomodulatory proteins expressed by a number of cancer cells as a mean of a means of down regulating the anti-tumor immune response and these are Targets of a number new emerging inhibitors And when we go into the expression data we could see that actually the EBV positive tumors have higher Expression of both PDL 1 and PDL 2 with a subset having true outlier Profile is being the cases that are that are the ones that have the 9p Amplicons and also within our gene expression. We could see that the actually the EBV tumors have a strong immune present Signature which goes along with our knowledge that EBV positive tumors have a strong immune cell infiltrate And so now if you could see from the sort of forming this idea that EBV positive tumors are their own group We've now found two very strong Therapeutic targets to take forward in this in its tumors both inhibiting the PI 3 kinds pathway as well as now bringing these novel PD1 and PDO 1 PDL 2 and inhibitors into this class of tumors Okay, so those are the EBV positive tumors But that's in 10% of gastric cancer No, 10% of gastric cancer still the 70,000 people dying every year, but what else can we say about the other groups? So just quickly if we go into the aneuploid type groups that the the the the sin tumors What's very striking is that all of the amplified targets that are present, you know recurrent amplifications at EGFR Irby 2 Irby 3, you know FGFR 2 met and so forth a number of targetable alterations and of these only right now Irby 2 is currently being Used in clinical practice. So in a number of new opportunities there to bring forward other biomarker driven therapies Now the MSI tumors as you would expect have a lot more mutations less copy number aberrations And here we scored these mutations by gene with the key point being the dark green is where you have Mutations present that are cosmic sites and light green are non-cosmic sites because you could also have a lot of passengers these hyper mutated tumors and A lot of these mutations are actually hotspot events for example recurrent mutations of Irby 2, you know, we have a number that are as 310 f that have been shown as being activated by Matthews group also recurrent hotspot events such as Irby 3 v 104 m These are recurrent hotspot activating mucent activating mutations shown to be Drug sensitive that are present here that aren't being acted upon right now Okay, but last we have this one group here the Genomically stable group or largely diffuse type cancers So these are really nasty cancers really bad actors And as you see from this view here looking at the PI3 kinase and RTK pathways, it seems a bit quiet There's not as much else there compared to the the other classes And so here's where we turn to trying to find Novel events looking at our analysis from mutsig from this cohort This is just looking at looking at the non hyper mutated tumors here separate out by the EBV's Stable and the sin tumors seeing things you'd expect like p53 heavily in the sin CDH1 heavily and diffuse PI3 kinase heavily and EBV as I mentioned as well as arid 1a But one novel event that we honed in on where these new mutations of row a that were enriched in the diffuse type group So if you look at these events, they have the pattern that look exciting. They're highly in the clustered events at row a Gtpa signaling protein including five mutations at codon tyrosine 42 So these two domains here are two areas where you have frequent mutations are two Parts the protein that will lie right next to each other in the effector area So row is a gtp bound protein when it's bound to Gtp like a rash protein is active and this is the area here where you signal to downstream effectors such as rock one here and Again, so these were heavily enriched in the diffuse type tumors or the geometrically stable tumors here So 15% of this cohort So row a is that a gene that's involved in it's it's signaling is important for Organizing this the actin the cytoskeleton enhancing invasion and migration. So again, these were present in Diffuse type gastric cancer. This is a class of tumor whose hallmark feature is that it grows in a way That's not cohesive and the cells are highly invasive. So this makes a lot of sense teleologically Now this pathway also came up from our analysis of candidate fusion genes And so both from overlapping the the low-pass hole genomes and the RNA seek was what came up very intriguingly where a number of alterations linking these genes together to cloud in 18 on chromosome 3 and Our gap 26 on chromosome 5 and these all of these fusions involved bringing the the the our gap protein onto the UTR of cloud in 18 This is a really small target like a little to Kb target now You have a situation here where you have a fusion gene that hits the the UTR. So right here You have a stop codon. So when we saw this we said, okay Well, actually you're not going to get a mature fusion transcript because you have a stop codon here, but the work led by Andy mungle and re and bulby With the transcripts figured out something quite cool And that was that you actually have an ability here to activate a cryptic splice site where the splice site from exon 12 of The second gene goes into exon 5 of the cloud in 18 finds a cryptic splice site splices out the last 30 nucleotides of this protein including the stop and creates an in-frame fusion. And so overall we had I think 13 of this events across our Our group of samples and the resulting protein and I'll get into what this would be doing in a second links together this claudon Membrane protein and at the very end of it hitches on the c-terminal portion of this row gap protein arh gap 26 And what was quite intriguing was just like the row a fusions these novel fusion proteins Were highly enriched in the same group of tumors the diffuse type class of cancers So both of them were present in 15 percent and it was a non overlapping group of tumors. So these new events were basically Present in a one-third of this class of very deadly and common cancers So, um, what are some of these genes? What could they be? So claudon 18 is a part of the tight junction This is a cellular adhesion complex. So tight junctions are actually what they would sound like actually Basically in an epithelial lining helped to sort of the rivet cells to gather Now our gap 26 is a row gap protein a row gap means a gtps Activating protein. This was the kind of protein that would act to reduce row a activity Which of course poses a question here as we imagine that these alterations would be activating this pathway But then again, so you do have this this novel protein here where you're Lopping on a part of this row gap domain to this adhesion protein so you're you know, you the Potential for this to be just the regulating this pathway both by modulating the location of the gap as well as its Function is is is quite large But to summarize what we've gone from is a situation where you have gastric cancer Without a good way to think about it in terms of uh to how to classify it or how to Take care of it to its new system now where we have four rationally Defined and robust classes of cancer where we have some clear salient features for each one We have the aneuploid type tumors You know with a lot of p3 mutation a lot of wrasse and rdk alterations We now clearly define ebv as its own subtype with actionable alterations such as pi3 kinase mutations And the overexpression of pdl1 pdl2 um We have the msi tumors that are highly mutated and have simp and we found a number of you know actionable mutations in this group And in the genomically stable tumors, we don't have as many simple obvious targets But we have a number of a new Insights including this very intriguing novel mutations of row a and this new um fusion Really helping us think about a novel pathway in this class of cancer And so with that I want to once again thank all the people who were involved in this project and thank all of you for your time Yeah, sorry wonderful talk Very very nice. So the question is uh, do we know anything about h pylori? Yeah in these cases And what's the any clinical outcome association with those four groups? And for example where they're each some group is more sensitive to chemotherapy um So we don't uh to answer your uh question. So in terms of h pylori We don't have very good annotation for that because unlike ebv the h pylori doesn't really doesn't get into the cancer cells And it's more on the Illuminal surface. So we we we weren't able to quantify that with Genomic data to the same extent In terms of clinical outcome the survival data are still coming in at the little data we have so far There's not big survival differences, but that's not surprising because you know survival and cancer cancer is just sort of like Bad and worse, you know, um, so there's um, and we we don't know yet about we don't have Data on sort of chemotherapy responses because there's so many different source sites and so much variability In the chemotherapy given we don't we don't have that level of the msi You would think that may be more sensitive to Possibly possibly you know in a msi colon is thought to be less responsive to adjuvant chemotherapy. So it's not you know Congratulations adam beautiful In in light of these g e junction tumors where pathologists struggle Are there any insights you think you can be gained from sort of merging? Some of your stomach data with some of the esophagus data that I know you're working on and Sort of analyzing those in the broader context. Someone would think you were a plant in the audience. Um, so, um, So, um, so i'll say that you know It's you know, it is a little bit arbitrary With these g e junctions whether people get called stomach or esophageal and because of history With the stomach project starting first all of the g e junctions were included in that bucket But we're actually merging for our next phase the esophagus and stomach working groups To try to look at these jointly. So as we move to the esophagus, we're going to really try to ask the question You know real tubular esophagus versus g e junction and so forth. Are they different? And so I think that's a key question moving forward as we move to the next phase here Okay, let's move on to our second speaker. It's arie hakemi Who will uh Talk about the integrated analysis of metastatic disease and clear cell renal cell carcinoma