 The cancer genome, the cancer genome atlas, as the name suggests, has been focusing on genomic, identifying genomic events and did not include a large-scale proteomics profiling. Gordon Miller and Edin Lue from MD Anderson, they just started to measure protein level and possible protein levels in TCG sample. And we kind of get this data and we are playing around with it and we tried to correlate those proteomics data with those genomic alterations. And I should point out that this presentation is more about working in progress and we just started to scratch the surface of this interesting data. So RPPA stands for Reverse Phase Protein Array and it's a high-throughput proteomics technique to measure protein level and the possible protein level in relatively expensive way. Each site is developed with a single antibody and could contain up to 1,000 samples with controls and there are currently about 300 validated antibodies available. The Edin Lue and Gordon Miller from MD Anderson, they have generated those RPPA data so far. In six cancer types for more than 2,000 samples and including like about 200 antibodies, among which 50 antibodies are first for phosphorylation antibody. Here I list a few. For protein antibodies, including like P10, TP53, ER, AR, AKT, et cetera, for phosphorylation antibodies, including AKT, EGFR, B2, MAPK, et cetera. And we should note that several of the genes have both protein level and phosphorylation level. For example, the ERB2, ERB3, AKT, and MPK. And this is a pathway overview of the antibodies. The pink and orange dot indicate that the particular gene have first for antibody, first for antibody and protein antibody respectively. And from this field it's clear that the antibodies actually covers a lot of members in those interesting important signaling pathways in cancer. And one might doubt that if this technique, RPPA, if it is working for TCGA samples. And in the next few slides I will just briefly give some examples that can confirm our expectation from the data. And here is the ERB2, here's the correlation between ERB2, MRA, protein level and phosphorylation level. And from the slides we can see that the ERB2, phosphorylation level is highly correlated with ERB2, MRA expression. And ERB2, phosphorylation level is also correlated with the ERB2 protein level. And we also study, compare the ER and GATA3 protein level in breast cancer subtypes. And as expected the protein level of basolite and ERB2 is significantly lower than the lumina samples. And the same with the GATA3 protein level. And as Giovanni have mentioned, and the PTEN deletion and under expression are highly correlated with ATK phosphorylation. Actually it's across multiple cancers. And so I hope that at this point we have more knowledge about this RPPA. We have some confidence in the RPPA technology and we want to dig into this AKT phosphorylation in breast cancer. So the AKT activation and phosphorylation has been found in many cancer types. And phospho-AKT has diverse targets regulating a lot of important processes in cancer. And phospho-AKT contributes to breast cancer progression and confers resistance to conventional therapies. Inhibiting phospho-AKT could be beneficial to some patients. Okay, this is a box plot and histogram for two important phosphorylation site level in AKT. And clearly we can see that there is a long tail for both of the phosphorylation sites that has a higher phosphorylation level. The question is what genomic even causes AKT activation in breast tumors? From our pathway overview, we have some usual suspects including like PTEN, KRAS, B2, and PGFR1, and PIC3CA. The PTEN deletion and loss of expression is highly associated with phospho-AKT but not PTEN mutation. And for those KRAS, B2, FGFR1, and PIC3CA, actually they are not highly associated with high AKT phosphorylation level. So we wanted to know what other genomics even can explain AKT phosphorylation in breast cancer. To approach that, we separate the samples into AKT high phosphorylation and AKT low phosphorylation, and perform enrichment tests of all gistic region of interest and frequently mutated genes. And we found a couple of interesting carbon number changes, this too, and a few mutations that are significantly altered in AKT high, with samples with higher phosphorylation of AKT. And we actually test all of those genomics events and find that one of the application and AKT E17K mutation may be related to higher AKT phosphorylation level. Again, I should mention, again, this is just a surface. We just scratch the surface of those great data and we think that the data is very interesting and definitely worthwhile digging more. And we found that it's, we have observed very good correlation between gemomics and proteomics data on the level of individual genes and proteins. But the downstream effects of protein level and phosphorylation level is harder to link to genomic events. In the case of the AKT phosphorylation, we can explain some, but not all. Or we're in the progress to systematically investigate those genomic events combinations for all antibodies. And we think that correlations between phosphorylation data may help elucidate, activate signaling pathways. This IPPA data has been integrated in the CBIO GDAC portal and that I would like to thank Nicky, Giovanni, Ethan, Ben and Chris from MSKCC and Gordon Mills, Yiling Lu and the IPPA team from MD Anderson. And thanks for your attention. Thank you.