consensusRegions - Consensus Regions from Replicated ChIP-seq and ATAC-seq Experiments

Merges the evidence carried by replicated ChIP-seq, ATAC-seq or CUT&RUN experiments into a single set of consensus regions. When several replicates agree on a position their p-values are combined, so that a peak too weak to survive on its own can still be retained. Fisher, Stouffer, Lancaster and rank-product combinations are available; the weighted ones accept per-replicate weights taken from FRiP scores, library sizes, or the peak sets themselves when no alignment file is at hand. Input can be anything from a full narrowPeak to a bare BED3: scores are rank-transformed when p-values are missing, and a weighted k-of-n presence rule takes over when even the score is absent. The threshold on the combined significance can be calibrated against a permutation null rather than fixed a priori. Quality control plots and export to SummarizedExperiment are included.