Abstract
Recent advances in the methodologies of reconstructing cancer evolutionary trajectories opened the horizon for deciphering the subclonal populations and their evolutionary architectures under the cancer ecosystems. An important challenge of the cancer evolution studies is connecting genetic aberrations in subclones to clinically interpretable and actionable target of subclones for individual patients. In this paper, we present a novel method for constructing tumor subclonal progression model for cancer hallmark acquisition using multi-regional sequencing data. We parepare a subclonal evolutionary tree inferred from variant allele frequencies and estimate the pathway alternation probabilities from large scale cohort genomic data. We then construct an evolutionary tree of pathway alternation that takes account of selectivity of pathway alternations by the notion of probabilistic causality. We show the effectiveness of our method using a dataset of clear cell renal cell carcinomas.
