Which file is af-only-gnomad.hg38.vcf.gz?

Answered

March 09, 2020 19:33

Hi Team!!

I have been searching for a file named :

af-only-gnomad.hg38.vcf.gz

I went to gnomad website and downloaded the file All chromosomes VCF (.tbi) from the gnomADv3 tab. The file is overwhelmingly large and I wonder how much time will the PONs take if I used it. Am I referring to the correct file.

After reading various GATK blogs, I came across this bucket that contains the af-only-gnomad.hg38.vcf.gz. but the size is very small. I don't think the two files are the same? Are they?

Can you please point me to a GATK blog where the difference between two files is clearly underlined?

You mentioned that you don't provide the aforementioned file for hg19 here. Does it mean the file I downloaded from the bucket is reliable or was it reducted for workshop purposes?

4 comments

David Benjamin

March 10, 2020 04:43
rohit satyam The gnomAD VCF is enormous because it contains a lot of INFO field annotations, none of which Mutect2 needs except for AF (allele frequency in the population). The AF only gnomad that we provide in the best practices google bucket is the gnomAD VCF with all extraneous annotations removed. In principle you could use gnomAD with all the annotations, but it would waste a lot of CPU time parsing the VCF.
1

Comment actions Permalink
rohit satyam

March 10, 2020 17:31
Wow. Thanks a lot. David Benjamin
0

Comment actions Permalink
Brian Wiley

January 23, 2021 10:35
David Benjamin, I just did a quick check for first 5 variants from af-only-gnomad.raw.sites.vcf for b37 for chromosome 1 and 17. All 5 I could see from chr17 on gnomAD site but none of first 5 from chr1 (positions 10067, 10108, 10109, 10114, 10119) are listed (see https://gnomad.broadinstitute.org/region/1-10067-10128?dataset=gnomad_r2_1). The first on gnomAD is 10128 for chr1. How was af-only-gnomad.raw.sites.vcf created for b37?
0

Comment actions Permalink
Genevieve Brandt (she/her)

January 29, 2021 17:26
Hi Brian Wiley, from one of the other threads you commented on I found a WDL script that builds the resources. If you didn't already see it, you can check out this script for more detailed information about how the devs build the Mutect2 resources.

https://github.com/broadinstitute/gatk/blob/master/scripts/mutect2_wdl/mutect_resources.wdl
0

Comment actions Permalink