error in wellington_footprints.py

[billgreenwald]

Hi!

I just downloaded wellington, and when running wellington_footprints.py I was getting the error:

error at line 85:
'<' is not defined for GenomicInterval objects

I determined this was happening only when a BED file with multiple chromosomes was passed. The full line 85 was

orderedbychr = [item for sublist in sorted(regions.intervals.values()) for item in sorted(sublist, key=lambda peak: peak.startbp)]

but, after some testing, it seemed like the error was coming from

sorted(regions.intervals.values())

Based on the comment above that the goal is to iterate by chromosome, then position, I edited line 85 to be

orderedbychr = [item for sublist in sorted(regions.intervals.values(),key=lambda genomicIntervalList: genomicIntervalList[0].chromosome) for item in sorted(sublist, key=lambda peak: peak.startbp)]

This has overcome the bug and the program runs. I am posting this to both confirm that this is the intended result and that my output should be correct, and to give a bugfix if so.

Thanks!

[davidhbrann]

Hi,
I'm trying to run wellington_footprints.py as well and was encountering the same error. I checked quickly and it looks like your fix also prevented the error on some data I downloaded from encode wellington_footprints.py -p $SLURM_NPROCS wgEncodeUwDgfHaspPk.bed wgEncodeUwDgfHaspAln.bam pydnase_out, but I'd also like to know if the fix that Bill posted is correct.

Thanks!

[billgreenwald]

Hey @dhb2128, with my data, Wellington hangs with the last 80 threads finishing up and I get empty output. Do your output files have entries in them?

[davidhbrann]

I got the same message, "Waiting for the last 80 jobs to finish...", but it still produced a WellingtonFootprints.FDR.0.01.bed file, the WellingtonFootprints.wig file, and the files in the p value cutoff folder in the outdir.

[billgreenwald]

was your WellingtonFootprints.FDR.0.01.bed empty?

[davidhbrann]

Not when I ran wellington_footprints.py. That one seemed to work. However, I get no output when I run wellington_bootstraps.py. Not sure if it's a threshold thing; self.scores is less than the FDR limit but self.findDiffFP() seems to not be returning any results.

Was it always empty for you?