Why sequence is unmapped by bwa but position is given in Sam file and also can be aligned by blat ?

Question

[ZheFrench]

I have a read where sam flag is 69 (meaning Pair End Read, Unmapped and First in Pair) , Mapping quality is 0 and Cigar string is *. The sequence is not showed in the viewer (samtool tiview,IGV).

I'm wondering why a read is said unmapped whereas a position is given in the sam file. So the read can be mapped? When I tried to align with blat, I found one unique position (the same as in sam file).

So I don't understand ...Why this read is said unmapped whereas a position is given(chr12 112036669)? How to retrieve/use this read in my bam to display in viewer as a "good sequence"? Why is the read filtered? Because of the mapping quality? and why CIGAR string is set to *? Can I use it of one way or an another or is it a better idea to let this read stay in "garbage"? Thanks.

M02792:15:000000000-A98VA:1:1103:24705:18436 69 chr12 112036669 0 * = 112036669 0 GTCGCTGAAGCCCCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCCGCCGCCCGCGG CCCCCGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG8FGGG@FF:F@FE=: RG:Z:A421

Answer 1

[Devon Ryan]

If an alignment is marked as unmapped then any position information is unreliable, since it may simply be a hold-over from an internal method that didn't get cleared. In the case of paired-end reads like this, typically an unmapped mate is given the coordinates of its mapped mate.

[ZheFrench]

Ok so that's meaning the sequence is unmapped for X reasons but the sequence can be still considered as a "good one". If I use another mapper or bwa with different parameters for example, the sequence could be probably set as mapped. The mapping is wrong but the sequence is (probably) good ?

[Devon Ryan]

The sequence is simply the sequence. Whether a read maps or not has no bearing on the quality of its sequence.

Answer 2

[hoxus.bidon]

Long time ago, but I have the same "problem".

Read in the BWA FAQ:

I see a read stands out the end of a chromosome and is flagged as unmapped (flag 0x4). What is happening here? Internally BWA concatenates all reference sequences into one long sequence. A read may be mapped to the junction of two adjacent reference sequences. In this case, BWA will flag the read as unmapped, but you will see position, CIGAR and all the tags. A better solution would be to choose an alternative position or trim the alignment out of the end, but this is quite complicated in programming and is not implemented at the moment.