Lose of paired-end info after samtools markdup

[ishtiaqueh]

Hi all,

I just did some paired end ATAC-seq and I have been trying to analyse the data. I align with Bowtie, remove mitochondrial reads, sort and then remove duplicate reads. I have noticed that the bams I generate from my fastqs lose the paired end information (samtools view -c -f 1 yields 0 reads) after marking duplicates with samtools. This is not the case when I run the same pipeline on other fastqs from published datasets.

I have added screenshots of the first few lines from the 2 fastqs. It'd be very helpful if someone could point me to any noticeable differences between the two that might be causing this discrepancy and how I might go about solving it.

Thanks!