bedtools_intersect files sorting problem

Question

[victoria_aleks]

Dear all, I am running the command: bedtools intersect -sorted -g [GENOME_FILE] -abam [BAM_FILE] -b [BED_FILE] > [OUTPUT_FILE] - and am getting an error about wrong sorting. Apparently my input BAM files have strange sorting, that goes: chr1 chr2 ... Chr9 X Y Chr11 chr12 ...

While BED and GENOME files are: chr1 chr2 ... chr22 X Y

As I understand it is easier to "re-sort" BED file than BAM. Still, I have troubles with doing this :) Could any one advice on this subject? Thank you!

[IP]

How did you sort the bam file?

[victoria_aleks]

I did not. I have sorted bed and genome files according to bam (as I thought). only later I found out that bam file has this strange sorting with XY chromosomes, and, honestly, i dont know how to sort bam file :)

[Matteo Schiavinato]

samtools sort [-l level] [-m maxMem] [-o out.bam] [-O format] [-n] [-T tmpprefix] [-@ threads] [in.sam|in.bam|in.cram]

This is what you're looking for if you don't know how to sort a bam file: http://www.htslib.org/doc/samtools.html

[victoria_aleks]

I still dont understand how i can reshape the order of the chromosomes, so that X and Y will be not between chr9 and chr10, but at the beginning/or at the end...

[IP]

Depending in your error, you may need to sort the bam file. You will need to add more info to the question in case you want some help

[victoria_aleks]

the error is that genome file and bed file have sorting "1 2 3 4 5 6 7 8 9 10 11... x y" and bam file has " 1 2 3 4 5 6 7 8 9 x y 11 12..."

[Sinji]

You need to sort your bed_file and your genome_file appropriately. Bedtools has a 'sort' function that's rather slow, but you can do this to fix the problem:

LC_COLLATE=C sort -k 1,1 -k2,2n a.bed

[victoria_aleks]

I actually sorted the bed file in this manner, but it puts the XY at the end, while my bam file has XY chromosomes in between single-number chromosomes and double-number chromosomes (i.e. between chr9 and chr10)

[Sinji]

So you did LC_COLLATE?

[victoria_aleks]

yes, and XY chromosomes a placed at the end :)