For organisms that have a blacklist, a config mechanism should be provided to specify it (by URL or file:// URI) to use it when calling peaks directly with the peak caller (like macs2), or to be provided as a step after peak-calling for those peak callers that don't support it (spp).
For organisms that don't have a blacklist, provide an alternative mechanism for configuring which samples should be provided to the GreyListChIP package to generate a greylist to be used as above
These lists should be optional
The GreyListChIP package describes the algorithm, which seems straightforward. It may or may not make sense to reimplement into a snakemake rule using other tools already installed in the environment. Specifically, it does this:
generating a tiling of the genome,
counting reads from a BAM file for the tiling,
sampling from the counts and fitting the samples to the negative binomial distribution to calculate the read count threshold,
filtering the tiling to identify regions of high signal, then
The GreyListChIP package describes the algorithm, which seems straightforward. It may or may not make sense to reimplement into a snakemake rule using other tools already installed in the environment. Specifically, it does this: