samtools -view region (overlap?) question

[dietmar13]

hello,

does somebody know how overlap is definded for the

samtools view region

function.

given the defined region # and four reads a-d:

                    ##############################
                aaaaaaaaaaaaaaaaa
                    bbbbbbbbbbbbbbbb
                               ccccccccccccccccccc
                                        ddddddddddddddd

which one is extracted: b and c or a-d?


thank you,

dietmar

[syfo]

No idea, but it should not be difficult to test by browsing a real bal file with the "samtools tview" option and selecting the first region with a couple of overlapping reads. I bet on a-d

[maubp]

I think it should give you all of them - any reads within or overlapping the region requested.

[dietmar13]

further question

I think it should give you all of them - any reads within or overlapping the region requested.

this seems true, all overlapping were extracted.

does somebody know a script to fetch only those which start AND end inside a given range (###).

                    #################################
                aaaaaaaaaaaaaaaaa
                    bbbbbbbbbbbbbbbb
                               ccccccccccccccccccc
                                        ddddddddddddddd
eeee----splice------eeeeeeeeeee

here only: b and c

[swbarnes2]

You might consider using BEDTools instead of samtools view. BEDTools can be given a .bed file of intervals, and a .bam file, and you could limit the reads to only those which completely overlap the intervals.

[dietmar13]

@swbarnes

thank you, works great, but too slow with big bam files:

therefore preselection with samtools...

samtools view -b xxx.bam chr1:1000-2000 | bedtools intersect -f 1.0 -b region.bed -abam - | samtools view - > yyy.sam