My interest in TCGA analyses began at UVA when I would witness lab members misuse cBioPortal to show that their genes were involved in cancer. I began to wonder why we study genes and then retroactively try to show that they may be involved in cancer. Why not just study the genes that are most likely to be involved in cancer? Surprisingly, you could only find lists of the genes that were most commonly mutated in cancer and not genes whose expression correlated with survival. I was also curious if the same genes were implicated in the different cancers and ended up publishing the largest pan-cancer analylsis of prognostic genes. This also made me realize there wasn't a good data portal for survival data, especially for miRNAs and lncRNAs, so I created OncoLnc.
I also was interested in using the RNA-SEQ data from TCGA to identify novel lncRNAs and with Brian Reon developed a pipeline for finding lncRNAs in brain cancer.
Although my main work currently involves meta-science, OncoLnc receives over 100k hits per month and I field questions about the data and send researchers data and code.