Re: [Samtools-help] samtools flagstat mapped read count

[Kate Im <kat...@gm...>]

Thank you so much for clearing that up. I was unaware that was the
convention when only one paired end is mapped.


On Tue, May 27, 2014 at 6:56 AM, John Marshall <jm...@sa...> wrote:

> On 27 May 2014, at 10:58, Wolfgang Maier <
> wol...@bi...> wrote:
> > On 22.05.2014 20:59, Kate Im wrote:
> >> the number of unmapped read (estimated by
> >> subtracting the reported number of mapped reads from the reported number
> >> of total reads) is always higher than the number of sequences with an
> "*"
> >> in the third column of the SAM file. Shouldn't these be the same?
> >
> > Ideally, yes, but the SAM/BAM format specifications
> > (http://samtools.github.io/hts-specs/SAMv1.pdf) say that:
> >
> > "Bit 0x4 [in the FLAG field] is the only reliable place to tell whether
> > the segment is unmapped.
>
> In particular (see §2, 4.1 of that document), there is the common
> convention for pairs in which just one end is mapped, of giving both reads
> the RNAME and POS (3rd and 4th) columns of the mapped end.  This has the
> useful side-effect of bringing the unmapped end alongside its mate when the
> file is coordinate-sorted.
>
>     John
>
> --
>  The Wellcome Trust Sanger Institute is operated by Genome Research
>  Limited, a charity registered in England with number 1021457 and a
>  company registered in England with number 2742969, whose registered
>  office is 215 Euston Road, London, NW1 2BE.
>