No, that is the correct interpretation of the flags. 8 is also the flag for "mate is unmapped", which in theory, the first read should have, since it has the 4 flag for being unmapped itself. I'd guess that the 4 flag in the second read is an error, but it's hard to explain. One weirdness that I know bwa has when aligning is it concatenates separate reference sequences, so if a read crosses over two references, it's have a mapping position, and be flagged as unmapped (PICARD yells at you when it sees this), but that wouldn't seem to explain what you have there.
And no, unmapped paired reads might not have an * in the rname. If a read is unmapped, and its mate is mapped, it's supposed to have the rname and position of its mate. That's so they will sort together. It's not a bug, it's a feature, really. The flag is supposed to show this, maybe the ISIZE field will too.
And no, unmapped paired reads might not have an * in the rname. If a read is unmapped, and its mate is mapped, it's supposed to have the rname and position of its mate. That's so they will sort together. It's not a bug, it's a feature, really. The flag is supposed to show this, maybe the ISIZE field will too.
Comment