DNAStringSet as genome input for RunChromVAR and AddMotifs

[rpolicastro]

Currently the documentation for AddMotif and RunChromVAR states that a BSgenome object is required for input to the genome argument. To avoid having to forge a BSgenome for new/custom assemblies or unsupported organisms, a DNAStringSet would be preferred as an alternative input. It appears that without modification to the code both AddMotif and RunChromVAR work with a DNAStringSet in a preliminary test run.

For AddMotif the internal call is to motifmatchr::matchMotifs. A DNAStringSet is a supported genome input according to their documentation.

For RunChromVAR although the chromVAR documentation only states that a BSgenome object is used for input, their generic functions support DNAStringSets.

There also don't appear to be any input checks or BSgenome specific function calls in Signac that would throw an error in an initial perusal of the code.

Thanks!

[timoast]

You're right, the requirement for a BSgenome object in Signac::RunChromVAR() comes from what's stated in the documentation for chromVAR::addGCBias() but it seems like other inputs are supported (based on https://github.com/GreenleafLab/chromVAR/blob/master/R/utils.R#L320-L358). I'm happy to update the documentation to state that DNAStringSet and FaFile objects are also supported.

For AddMotifs() I will also update to match what's stated in motifmatchr::matchMotifs()

[timoast]

Updated on the develop branch