Hello,
After filtering a BAM with paired end reads, there are frequently orphan reads whose mate is no longer in the BAM file itself.
I cannot find any means in which this is tracked in the record for the remaining orphaned read, which now has information about a read that no longer exists in the file.
I tried running Picards FixMateInformation after removing a read from a pair and the record was unchanged. I thought the 0x1 flag might get switched to indicate it was a single end read, but more careful reading of the SAM specification indicates that 0x1 indicates that the read was "paired in sequencing", not paired in the alignment.
Is this correct? My concern is that downstream applications that require PE reads (e.g., for structural variant discovery) might not handle BAMs with orphaned reads.
Thanks.
Jonathan
After filtering a BAM with paired end reads, there are frequently orphan reads whose mate is no longer in the BAM file itself.
I cannot find any means in which this is tracked in the record for the remaining orphaned read, which now has information about a read that no longer exists in the file.
I tried running Picards FixMateInformation after removing a read from a pair and the record was unchanged. I thought the 0x1 flag might get switched to indicate it was a single end read, but more careful reading of the SAM specification indicates that 0x1 indicates that the read was "paired in sequencing", not paired in the alignment.
Is this correct? My concern is that downstream applications that require PE reads (e.g., for structural variant discovery) might not handle BAMs with orphaned reads.
Thanks.
Jonathan
Comment