bumphunter

Dependencies & Environment

Package-intrinsic requirements from the Bioconductor landing page, reproduce in any R environment.

• Version: 1.54.0 · Bioconductor: 3.23 · R: ≥ 4.6
• Depends: S4Vectors, IRanges, Seqinfo, GenomicRanges, foreach, iterators, locfit
• Imports: matrixStats, limma, doRNG, BiocGenerics, GenomicFeatures, AnnotationDbi
• Install: BiocManager::install("bumphunter")

When to Use

• Finding continuous, spatially clustered genomic regions ("bumps") that differ significantly between conditions using bumphunter().
• Grouping genomic locations into distinct clusters based on maximum distance using clusterMaker().
• Extracting positive, near-zero, and negative segments from a vector of test statistics using getSegments().
• Packaging segmented regions into a table of bump characteristics using regionFinder().

When NOT to Use

• For basic linear modeling of independent, unclustered genomic features, use limma because bumphunter is specifically designed to share information between nearby clustered locations.
• For end-to-end analysis of Illumina 450k arrays without manual matrix setup, use minfi because it provides a tailored wrapper around the bumphunter engine.
• For whole-genome bisulfite sequencing (WGBS) data requiring specialized smoothing, use bsseq because it adapts the bump hunting methodology specifically for bisulfite data.

Data Requirements

• Signal Matrix: A numeric matrix (y) of genomic signals where rows represent genomic locations and columns represent biological replicates.
• Design Matrix: A design matrix (X) representing the experimental covariates, created with standard R modeling functions.
• Genomic Coordinates: Vectors for chromosome (chr) and genomic positions (pos) corresponding to the rows of the signal matrix.