Forum:1000 genomes data?

Question

[datanerd]

Hi guys,

With better and bigger reference data available from more diverse populations through TOPMED, was wondering if people still rely on 1000 genomes or use that data in their research. If so- how?

Thanks,
Mamta

[Ram]

I've moved this to Forum as this is more a discussion than a topic with a finite set of right answers.

Answer 1

[Ram]

Even back in 2015-6, my then-lab had moved on to ExAC, and now my current lab uses gnomAD. I've never seen references to 1000g in the recent past, and although VCF annotation pipelines still add 1000g fields, it's more of a just-in-case-you're-interested scenario than actual functionality. 1000g still has good population frequency data, but even that is addressed a lot better by gnomAD.

[WouterDeCoster]

I guess there are populations in 1000G that are absent from gnomAD?

[Ram]

I gotta check but IIRC, gnomAD is more granular than 1000g (gnomAD has <population>_<sex>_ fields, whereas 1000g only has population fields).

[Ram]

Honestly, comparison is not straightforward (not on the web interface at least). 1000g lists more specific origins on their samples. For example, both GIH and ITU 1000g populations have South Asian Ancestry (SAS super-population) where gnomAD only specifies sas (v2.1)/South Asian (v2, v3) as ancestry for its samples. OTOH, I get to use non-cancer subsets and gender-specific AFs, which is more useful than having to compare to 118 Telugu speaking folks in the UK.