Our new paper is on biorxiv! "Predicting master transcription factors from pan-cancer expression data" biorxiv.org/content/10.110…. We set out to find master TFs in ovarian cancer, and went on quite a journey... A few highlights in the thread 1/
Prioritizing candidate MTFs without ChIP-seq data is tough, so we developed a gene expression-based approach which we loving called the Cancer Core Transcription factor Specificity (CaCTS) algorithm 🌵. (We then decorated the lab in cactuses) /2
Using CaCTS we identified candidate MTFs for 34 major tumor types and 140 molecular/histologic subtypes.
Candidate MTFs tend to be associated with super-enhancers, and many are lineage-specific essential genes - woo CaCTS works! 3/
Candidate MTFs tend to be associated with super-enhancers, and many are lineage-specific essential genes - woo CaCTS works! 3/
If we cluster tumors by the union set of 273 candidates, we see both anatomic (e.g. renal) and functional (e.g. squamous) groupings. Squamous tumors contain both squamous and lineage-specific TFs - we wonder how these dual circuitries cooperate (theres another paper there!) 4/
Candidate MTFs contain more somatic mutations than non-candidates that are also highly expressed, suggesting possible mechanisms of deregulation in cancer 5/
As a proof-of-concept we validated putative MTFs for ovarian cancer. Our circuitry consists of PAX8, SOX17 and MECOM. These 3 factors are associated with large SEs, are important for ovarian cancer cell survival, and cobind the genome to regulate key ovarian cancer genes 6/
PAX8, SOX17 and MECOM are very sensitive to general inhibitors of transcription and we hope this work will pave the way to a much-needed new treatment for ovarian cancer (and other gyn tumors driven by these 3 factors). 7/
This fun project was jointly led by @jessgenomics @mihaell and @ivycorona, with contributions from the whole Lawrenson lab including talented student @RNameki, @AbbasiForough 💪🦔 @AndrewlinB and Felipe (Felipe, get a twitter account) and last years Gyn Onc fellow Heidi Chang
And as always, we can't thank our collaborators enough - this work was done with critical contributions from many including @youngricka @TheAbraLab @IsaacAKlein @ronny_drapkin @MoreiraPharmD @simon_gayther @HoutanNoushmehr @beth_karlan De-Chen Lin, Matt Freedman
