zhou-lab/biscuit

biscuit markdups

Opened this issue · 4 comments

will-NYGC commented

I'm testing out your pipeline and am getting an error on BAM files produced after running biscuit markdups.

[E::bam_hdr_read] invalid BAM binary header

Further, progress output suggests no dups are marked although dup rates based on picard MarkDuplicates are in the order of 20%.

do you know why my BAM header is invalid? related to stderr suggesting no dups marked?

will-NYGC commented

Hi, I thought I'd bump this since I'm still having issues. I've tried running biscuit markdup infile.bam out.file.bam on my data for a couple of different aligned BAMs and keep getting the above error. They're BWA-meth aligned so should work. Have you seen this issue before?

zwdzwd commented

Hi Will,
Thanks for reporting. This is a bug with the bam header. See if the following solved the problem.
release.zip
I will include the change in the new release.

Wanding

will-NYGC commented

looks to be fixed. thanks, wanding!

ttriche commented

nb. I find that I have to re-sort the markdup'ed BAM in order to index it and proceed with pileup'ing.