Dev compute matrix strand 20220301

[joshtburdick]

Welcome to deepTools GitHub repository! Please check the following regarding
your pull request :

• Does the PR contain new feature?
• Does the PR contain bugfix?
• Does the PR contain documentation changes?
• Does the PR contain changes to the galaxy wrapper?

This is my sketch of a way of allowing computeMatrix to deal with stranded pairs of bigWig files (e.g. "...forw.bw" & "...rev.bw"), without having to run it for each strand.

Please note that this is much more a "concrete suggestion" than a stable feature. I haven't added Travis tests; indeed, I haven't sent any pull requests before, to anything.

It adds a class, pyBigWigStranded, which is just a wrapper for a pair of pyBigWig objects. (It might well make sense for this to be part of pyBigWig, but I wasn't sure how to do that.) Having done that, the main other changes to computeMatrix are that:

• I sketched a command-line option for what names to use for stranded pairs. For things like RNA-seq / PRO-seq data, I've seen .forw.bw & .rev.bw, or .p.bw & .m.bw, or .pl.bw & .mn.bw, or .plus.bw & .minus.bw . So I like the idea of "the user gives the + strand file, and deepTools guesses the - strand file". I realize that may be too complicated.
• the computeMatrix code guesses whether a file is stranded or not, based on the name, and either creates a regular pyBigWig object, or a pyBigWigStranded object
• the inner loop just uses whichever object; if the file is stranded, the pyBigWigStranded object forwards the .values() call to the pyBigWig object for that strand

Again, to reiterate: this is very much a sketch, but to me it seems possibly convenient.