[Samtools-help] Question about samtools mpileup and overlapping reads

[Carla M. L. <cma...@gm...>]

Hi!
I'm using Samtools version 1.3.1. I was analysing some illumina paired-end
sequenced samples and I realised that when running mpileup with this
version, by default, samtools discards all the Minus Strand reads when
there are overlapping reads. Is there any reason why always discards Minus
Strand reads? Does it make sense?  What happens is that I have some real
SNPs in regions where all the reads overlap so, when it discards all the
Minus Strand reads, the SNPs appear only in one strand and then it doesn't
pass the downstream filters that we apply. I know that with the
option --ignore-overlaps I can keep all the overlapping reads so my
question is Why samtools does not discard Positive strands or Minus Strands
in a random way by default?

Thanks for your attention,

Carla