mixNBHMM

Efficient Detection and Classification of Epigenomic Changes Under Multiple Conditions. Differential Peak Calling via Hidden Markov Model With Mixture of Negative Binomial Distributions. mixNBHMM is an R package for the detection of genomic regions exhibiting short and broad differential enrichment under multi-replicate, multi-condition settings. It is applicable to data from ChIP-seq, ATAC-seq, DNase-seq, and related experiments. mixNBHMM detects these regions using a three-state hidden Markov model (HMM) with a finite mixture of negative binomials as emission distribution in the HMM’s differential state. The HMM framework is particularly suitable for the detection of differential broad peaks and the mixture model allows the classification of differential combinatorial patterns of enrichment across conditions. Code to replicate the results from the mixNBHMM paper is available at https://github.com/plbaldoni/mixNBHMMPaper