Tophat output contains "unmapped"??

[capricy]

Hello, there,

I recently examined the output of tophat (tophat22.0.8) after converting them into sam file, and found that the following results:

grep "HWI-ST514:143982632:C37PRACXX:5:1101:12852:4786" accepted_hits.ns.sam
HWI-ST514:143982632:C37PRACXX:5:1101:12852:4786 89 A_ref-1.0_Cont33 766585 50 92M * 0 0 CTTGTATTGAGTACGATCTCTCCACCTCTCCGGTTCGCAATACAGCTTTGAGAAAGAACTTATTACCCTCTCTACTATATAATTAAATTGTA DDDDEDDDDDDDDDDDBDCDDDAACDDFHHJJJJJJJJIIHHJJJJJJJJJJJJJJJIJJIHFC:JJHGIIJJIJJJJJJJJJJJJJHHHFF MD:Z:92 XG:i:0 NH:i:1 NM:i:0 XM:i:0 XN:i:0 XO:i:0 AS:i:0 YT:Z:UU
HWI-ST514:143982632:C37PRACXX:5:1101:12852:4786 137 A_ref-1.0_Cont33 766582 50 100M * 0 0 TTCCTTGTATTGAGTACGATCTCTCCACCTCTCCGGTTCGCAATACAGCTTTGAGAAAGAACTTATTACCCTCTCTACTATATAATTAAATTGTACTTTG CCCFFFFFHHHHHJGIIJJJJJJJJJJJJIJIIJJJFGIIJJJJJJJJIJHIJJJJJJJJHHHHHFFFFFFFDEEEEDEDEDEFEEEECCEDCCDEEFED MD:Z:100 XG:i:0 NH:i:1 NM:i:0 XM:i:0 XN:i:0 XO:i:0 AS:i:0 YT:Z:UU

The reference is the genome scaffolds here. My question is about the samflag: 89 represents read paired,mate unmapped,read reverse strand,first in pair; 137 represents read paired, mate unmapped, second in pair.

These reads appear to be paired, and both mapped. Then why did samflag say their mates were not mapped? Is it because they did not map to the same scaffold?

Could anyone explain this to me?

Thanks

[westerman]

It is Tophat, not samtools, that sets the flags. So look at Tophat for answers to your question. I presume you mean version 2.0.8 and not 22.0.8; if so a newer version of Tophat may give better results.

I haven't used Tophat in about a year so treate the following with caution. For your specific question you say ".. Is it because they did not map to the same scaffold?..." but to my eye it looks like they did map to the same scaffold.

1st read to A_ref-1.0_Cont33 at base 766585
2nd read to A_ref-1.0_Cont33 at base 766582

Same scaffold with a 3-bp overlap. My guess is that this is why Tophat did not consider the two to have the mate mapped.

[capricy]

I meant that they did not map to the same strand of the scaffold...

[dpryan]

This appears to be a bug that was fixed in version 2.0.9, which was released over a year ago.

[capricy]

@dpryan, could you please explain a bit more? What is the correct information supposed to look like?

[dpryan]

What do you mean "what is the correct information supposed to look like"? You're using an old version with a known bug (the one you're asking about). Just upgrade.

[capricy]

well, upgrade and rerun, might take several days.

So I wonder if the changes for these two reads in new tophat2.0.9 will be in samflag fields?

Thanks

[dpryan]

Yeah, it should just fix the flags. In this example, the flags should become 83 and 163.

BTW, if you switch to STAR you'll get alignments vastly faster.

[capricy]

thank you very much for explanations!