Obtaining gnomAD allele count data

Science
1

Hi
I have downloaded the hail table gs://gcp-public-data–gnomad/release/2.1.1/liftover_grch38/ht/exomes/gnomad.exomes.r2.1.1.sites.liftover_grch38.ht (gnomad version 2 exomes lifted over to build 38) and I’d like to extract allele count data from this.

Using hl.eval(ht.globals.freq_index_dict) I can see a frozendict with several hundred different items in it. One of these is “‘gnomad’: 0”. Is the 0 the index within the freq field, and therefore mean that I should be extracting the AC from ht.freq[0]?

This is quite possibly a daft question but I do want to ensure that the AC I’m extracting is exactly what I think it is.
If there is documentation on this that I’ve missed I’d appreciate links.

Many thanks in advance
Ruth

2

That’s my understanding, yes. The freq field is an array with lots of groups of summary metrics for various sample groups/exclusions. The freq_index_dict is the map for encoding which sample subgroup is which, and the first index freq[0] is all gnomAD samples.

3

Thank you @tpoterba :+1: