Tree inference for single-cell data

摘要: Understanding the mutational heterogeneity within tumors is a keystone for development of efficient cancer therapies. Here, we present SCITE, stochastic search algorithm to identify evolutionary history tumor from noisy and incomplete mutation profiles single cells. SCITE comprises flexible Markov chain Monte Carlo sampling scheme that allows user compute maximum-likelihood history, sample posterior probability distribution, estimate error rates underlying sequencing experiments. Evaluation on real data simulation studies shows scalability present-day single-cell improved reconstruction accuracy compared existing approaches.