I created a set of bam files from Poolseq data using bwa -aln, and all of the output files gave the following error when I ran bamdst to get summary statistics on read depth:
"EOF marker is absent. The input is probably truncated."
When I ran a search for the error message, i found nothing specific to bamdst but several boards noted that samtools would return this error on bam files generated with several output arguments, presumably when converting sam to bam, e.g.
https://sourceforge.net/p/samtools/mailman/samtools-help/thread/4EC52844.3090808@broadinstitute.org/
Unfortunately the patch described goes to a dead link.
My main questions are: A. Why is bwa (or samtools conversion from sam to bam) generating bams with this missing EOF/truncation error? B. Is this something that I need to worry about? Apart from not being able to get summary stats from bamdst, the bam files seem to be processed without issue downstream in my pipeline. C. If the missing EOF is a problem, what can be done about it? I re-ran my alignment in case the truncation was due to termination prior to completion, but that wasn't the problem. I also attempted to convert bam to sam and back to bam again, but the error persists.
This is probably too late, but how exactly did you convert between sam and bam?