negative bwa mapping quality

[Anney]

Hi,

Does anyone know what negative bwa mapping quality means?

I am using the aligned data from 1000 genome project (I look at this sample NA12716.mapped.ILLUMINA.bwa.CEU.low_coverage.20101123.bam). For a properly mapped pair reads, I found that the mapping quality for one read is 226 and the other is -226. Does anyone know why?

When I do pileup in samtools, it excludes that read with negative mapping quality.

Here is that pair reads data:

ERR000573.12285810 pPR2 X 154402571 60 51M = 154402756 226 GATGCAATAAGAGATAAAGCTAGAGAGGTTAATAGAGGCCAGAACTCATAG =@%?>:@@?@>=>=@?A@>>@3>:=<=>:A>@@>=1=>;>929?>$>4@2> X0:i:1 X1:i:0 MD:Z:51 RG:Z:ERR000573 AM:i:37 NM:i:0 SM:i:37 MQ:i:60 XT:A:U BQ:Z:@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
ERR000573.12285810 pPr1 X 154402756 60 9S42M = 154402571 -226 GCGAAGAAGTCAATTAGAAAGTCTTTTCAAGTTATCCAAGCAGGAGGTCTC 27=AA=8A;<?A?@>@>@@@=;>=<>>?A@>@>@:?>AA>?@>??>>=?9? X0:i:1 X1:i:0 XC:i:42 MD:Z:42 RG:Z:ERR000573 AM:i:37 NM:i:0 SM:i:37 MQ:i:60 XT:A:U BQ:Z:@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

Thank you very much!!

Anney

[poisson200]

Hi.
Interesting and confusing to me too.
Sam manual says that field five is the mapping quality; From manual;
field 5 = MAPQ Int [0,2^8-1] MAPping Quality

However and interestingly, the MQ:i: tag is apparently the mapping quality of the mate/next fragment, which for your alignment are both 60
MQ i = Mapping quality of the mate/next fragment

So I would also be interested to know where -226 comes from too and what exactly is MQ reporting if 226 and -226 are genuine.

Another point is, if I understand and can do the math right, Mapping quality -10 x log10(probability of an error). So you would need a probability of about 1x10^25???????

Last thing; I assume the sam output reports the genome sequence not the reads, as this maps uniquely, at 100% to human genome hg19 to a unique position. So it should have a fantastic mapping score?

An interested bystander.

[Anney]

Thank you for the replay poisson200.

I checked BWA on its website (http://bio-bwa.sourceforge.net/). It says that it is possible one read in a pair has high mapping quality, but the other read has zero ("one read can be mapped unambiguously, but its mate falls in a tandom repeat and thus its accurate position cannot be determined"). So, I guess the negative mapping quality score means the undetermined mapping read. If so, then the flag shouldn't be "pPR2", which mean properly mapped pair-reads !?

And, samtools excludes the reads with negative mapping score (in default filter) when doing pileup.

Anney

[jts]

I believe you are looking at the template length field (field 9). The mapping quality is the 5th field and is reported to be 60 for both reads.

[Anney]

oh...yeah....thanks for pointing this out