PeakSegJoint: fast supervised peak detection via joint segmentation of multiple count data samples

摘要: Joint peak detection is a central problem when comparing samples in genomic data analysis, but current algorithms for this task are unsupervised and limited to at most 2 sample types. We propose PeakSegJoint, new constrained maximum likelihood segmentation model any number of To select the peaks segmentation, we supervised penalty learning model. infer parameters these two models, use discrete optimization heuristic convex learning. In comparisons with state-of-the-art algorithms, PeakSegJoint achieves similar accuracy, faster speeds, more interpretable overlapping that occur exactly same positions across all samples.