Tumor evolution and intratumor heterogeneity

Division Of Basic Sciences - Nci

Investigators

Linked publications & trials

Abstract

Treatment failures in cancer can many times be attributed to intratumor heterogeneity, both in terms of the genomic composition of tumor cells as well as epigenetic composition and especially methylation. We are developing algorithms and machine learning methods to infer the tumor phylogeny given single cell genomic, transcriptomic, exomic or methylomic data through the use of combinatorial optimization and deep neural networks. Out techniques enables researchers to pinpoint the exact genomic and epigenetic alterations that lead to aggressive and possible therapy resistant tumor sub-clones, We are also developing algorithms for detecting common evolutionary patterns, or trajectories to infer common causes and potential therapeutic targets.