skip to primary navigationskip to content
 

From big data to big model: a probabilistic approach to infer cancer evolution

last modified Jun 09, 2015 12:51 PM
Ke Yuan, Cancer Research UK Cambridge Institute

From big data to big model: a probabilistic approach to infer cancer evolution

Ke Yuan, Cancer Research UK Cambridge Institute

Big data allows complex models to fulfil their full potential. This has been demonstrated by the success of deep learning in supervised learning. When studying carcinogenesis, however, the problems of interest often belong to unsupervised learning paradigm in which expert labelled training datasets are unavailable.

Here, we present BitPhylogeny, a probabilistic framework to infer cancer evolution. The model automatically adapt its complexity with the scale of the data, which allows one to jointly estimate the number and composition of clones in the sample as well as the most likely evolutionary relationship connecting them. We validate our approach in the controlled setting of a simulation study and compare it against several competing methods. In two case studies, we demonstrate how BitPhylogeny reconstructs tumor phylogenies from methylation patterns in colon cancer and from single-cell exomes in myeloproliferative neoplasm.