Merged Reference Panel using WES data

Question

[shamav523]

I have a non-Caucasian population with genotype data that I want to impute to identify rare variants. I can use the Genome Aggregation Database (gnomAD) reference but the populations used to generate this panel do not include my population of interest. However, I have 800 exomes sequenced from my population of interest. From other posts, I can see that merging reference panels might be a useful alternative, but I wonder about the implications of merging gnomAD with only exome data from my population of interest, as opposed to WGS data. How would this affect rare variant detection? Would it make sense to limit my rare variant detection to exomes only?

[mimarcelape]

How can I merged gnomAD with my population of interest? I'm working with tumor-only samples and I'm doing a somatic variant calling with Mutect2. As a --germline-resource GATK recomend used af-only-gnomad.hg38.vcf.gz. However, I want to consider the AF of my population (ABraOM). Is it possible merged this two dataset? how can I do it?

[GenoMax]

Please post as a new question and then come back and delete this.

Answer 1

[jon.klonowski]

Use gnomAD v2 as that is exomes. gnomAD v3 uses primarily WGS v2 still uses some WGS data to fill in gaps, I think. v3 is mean for use when investigating non coding areas. V2 has more depth and genomes for coding areas.

https://gnomad.broadinstitute.org/help/should-i-switch-to-the-latest-version-of-gnomad