Accurate and Efficient Cell Lineage Tree Inference from Noisy Single Cell Data: the Maximum Likelihood Perfect Phylogeny Approach

摘要: Cells in an organism share a common evolutionary history, called cell lineage tree. Cell tree can be inferred from single genotypes at genomic variation sites. inference noisy data is challenging computational problem. Most existing methods for assume uniform uncertainty genotypes. A key missing aspect that real usually has non-uniform individual Moreover, are often sampling-based and very slow large data. In this paper, we propose new method ScisTree, which infers calls genotype Different most approaches, ScisTree works with probabilities of (which computed by callers). assumes the infinite sites model. Given uncertain individualized probabilities, implements fast heuristic inferring calling allow so-called perfect phylogeny maximize likelihood Through simulation, show performs well on accuracy trees, much more efficient than methods. The efficiency enables applications including imputation doublets. program available download at: https://github.com/yufengwudcs/ScisTree.