Devon Ryan

Can you remove the `--smoothLength` option? It's not actually doing anything here (the default is already not to do smoothing).

Nice investigative work. I'll have a look at the BAM file, since this sounds like a weird bug in pysam.

This appears to come down to a possible pysam issue: ``` >>> import pysam >>> bam = pysam.AlignmentFile("BsrDI_rep1_trimmed.bam") >>> bam.lengths (0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...

Maybe you need to invert the scaling factors? What about signal in intergenic regions?

Sorry, what's the issue?

They're the max/mean/etc. of all samples together. Internally, all of the samples are in the same matrix (as ranges of columns), the rows of which are being sorted.

deepTools parallelizes over contig chunks, so that's the primary limiter. This will probably still take a while, but I expect a day would be more than enough.

That version of deepTools is from 6 years ago and the matplotlib version you're using is similarly old. I'd suggest updating both to more recent versions.

Every line corresponds to a transcript in the GTF.

Can you post the bigWig somewhere so I can more easily debug this?