samtools: mpileup -r <regionspec> or -l <regionfile.bed> includes wrong areas

[jblachly]

samtools mpileup -r chrA:nnn-mmm includes region chrA:nnn-mmm correctly, but for some combinations of BAM files it also returns pileups from the same coordinates on other contigs, for example chrF:nnn-mmm and chrQ:nnn-mmm

It is not yet clear to me under what circumstances this behaviour is triggered. I have only experienced this when passing a list of bamfiles with -b, and notably when passing the following
-b <(head bamfiles.txt | tail -n [1..x])
to narrow down to a specific file that could trigger it, I triggered the behaviour at tail -n 4. However, passing the 4th bamfile ALONE does not trigger the error.

This behaviour happens with both -r but NOT -l; see also my other Issue#304

[jblachly]

I need to correct the above: the error was NOT triggered when using -l, but as I mentioned -l is abysmally slow.

[jkbonfield]

Closing this as it doesn't sound like an error. The -l being slow issue is a different thing (streaming vs random access).

Feel free to reopen the ticket if I've misinterpreted your last comment.