Different coverage from bedtools and in vcf file - HELP PLEASE

Question

[Paul]

Dear all,

I have trouble to understand how bedtools computing coverage. I have vcf file (generated from Illumina somatic caller) - I was created bed file from my vcf.

Then I try to compute coverage with bedtools and I have for the same coordinate different coverage. Could you help me to explain why? Probably some reads where filter out during vcf calling algorithm.

Output is:

coverage in vcf:

chr13    32890572    DP=45048
chr13    32893197    DP=1494
chr13    32899359    DP=12809
chr13    32900176    DP=57850
chr13    32900177    DP=61728

and in my bedtools output:

chr13    32890572    32890573    47857    1    1    1.0000000
chr13    32893197    32893198    1686    1    1    1.0000000
chr13    32899359    32899360    15673    1    1    1.0000000
chr13    32900176    32900177    65461    1    1    1.0000000
chr13    32900177    32900178    65461    1    1    1.0000000

Note, that my vcf coverage is before filtering.

Can anybody now why is this different? And if you can explain (mathematically) how bedtools computing coverage? Is it sum of all coverage in the interval and then divide length of interval? Of it is just sum of all coverage in interval?

Thank you so much for clarification.

Paul.

[Sukhi Singh]

Its important to know, how the coverage was calculated exactly, which sub-tool and which parameters were used. Just write the whole command used to generate the file.

[Paul]

Thank you for response. VCF was generate automatically from Illumina MiSeq Reporter - Illumina Somatic Variant Caller. And I am use coordinates of my variants to generate bed file to calculate bedtools coverage.

My code was:

bedtools coverage -abam my_input_bam -b my_bed_file > coverage.xls

I am not able to find how bedtools calculating coverage - mathematically meaning.

[Sukhi Singh]

Did you read this, it should answer your question.