Abstract

Single-cell DNA sequencing enables the construction of evolutionary trees that can reveal how tumors gain mutations and grow. Different whole genome amplification (WGA) procedures render genomic materials of different characteristics, often suitable for the detection of either single nucleotide variation (SNV) or for copy number aberration (CNA), but not for both, hindering the placement of both SNVs and CNAs on the same phylogenetic tree for the study of interplay of SNVs and CNAs. SCARLET places SNVs on a CNA tree, a tree derived based on the copy number profiles, while considering SNV loss due to copy number losses. However, SCARLET requires that the SNVs and CNAs are detected from the same sets of cells, which is technically challenging due to the sequencing errors or the low sequencing coverage associated with a particular WGA procedure. Here we presented a novel computational tool, SCsnvcna, that aims at placing SNVs on a CNA tree whereas the sets of cells rendering the SNVs and CNAs are independent, thus is more practical in terms of the technical challenge from single cell WGA process. SCsnvcna is a Bayesian probabilistic model that utilizes both the genotype constraints on the tree and the cellular prevalence (CP) to search the solution that has the highest joint probability. Both simulated and real datasets show that SCsnvcna is highly accurate in predicting the placement of SNVs and SNV cells. In addition, SCsnvcna has a precise prediction of SNV losses due to copy number loss.