filtering uniquely mapped reads results in odd number of reads in PE data

[rikrdo89]

When using -F "[XS]==null" to get uniquely mapped reads, I get a odd number of reads, despite having paired-end data which should result in an even number of reads. is there a way to only get the pairs that are uniquely mapped? I tried adding "proper_pair" but it still resulted in an odd numbers of reads

[pjotrp]

It probably requires a new feature. Closing this until someone wants to write the code :)