Predicting clone genotypes from tumor bulk sequencing of multiple samples.

摘要: Motivation Analyses of data generated from bulk sequencing tumors have revealed extensive genomic heterogeneity within patients. Many computational methods been developed to enable the inference genotypes tumor cell populations (clones) data. However, relative and absolute accuracy available in estimating clone counts is not yet known. Results We assessed performance nine methods, including eight previously-published one new method (CloneFinder), by analyzing computer simulated datasets. CloneFinder, LICHeE, CITUP cloneHD inferred with low error (<5% per clone) for a majority datasets which samples contained evolutionarily-related clones. Computational did perform well mixtures clones different clonal lineages. Generally, number was underestimated overestimated PhyloWGS, BayClone2, Canopy Clomial required prior information regarding AncesTree produce results large Overall, deconvolution single nucleotide variant (SNV) frequency differences among remains challenging, so there need develop more accurate robust software genotype inference. Availability implementation CloneFinder implemented Python https://github.com/gstecher/CloneFinderAPI. Supplementary are at Bioinformatics online.