1000 Genomes and ESP Populations in Exome Aggregation Consortium Data

Question

[Vivek]

Is there any documented information on how many of the ESP/1000 genomes samples were included in the ExAC data release? I was under the impression that all samples were included but when I'm trying to annotate a few SNPs I can see some discordance in the allele frequencies.

For example I see this Exonic SNP with 0.119 allele frequency in 1000 genomes Phase 3 dataset but this cannot be found in ExAC data

5    131705587    rs13180043    C    T    100    PASS    AF=0.11901

Another example, present in both ESP & 1000 genomes but not in ExAC

1    156108976    rs7339    G    C    .    PASS    DBSNP=dbSNP_52;EA_AC=246,2936;AA_AC=561,823;TAC=807,3759;

1    156108976    rs7339    G    C    100    PASS    AF=0.185304;

I'd like to know preferably what kind of overlap exists between these 3 population sets and if possible what kind of capture regions were used for ExAC data.

[Ram]

Is there any source to your assumption that all 1000g/ESP samples were included in ExAC? The About page speaks of an analysis from scratch, which would imply that the results are independent of 1000g or ESP.

[Vivek]

Check under contributing projects: http://exac.broadinstitute.org/about

Even if they did variant calling from the scratch if you include a sample set, you expect to see a SNP with high enough allele frequency in the population.

[Ram]

That makes sense. I guess you could always email them for specific details or check if they have a preprint that you could read.

Answer 1

[rbagnall]

Neither of the variants are in the called regions.

5   131701132   131701341   -   hsa-mir-3936
5   131705614   131706109   +   SLC22A5|ice_target_65179

1   156108820   156108949   +   LMNA|ice_target_14401
1   156109511   156109680   +   LMNA|ice_target_14402

See here for the full bed file of called regions: ftp://ftp.broadinstitute.org/pub/ExAC_release/release0.1/exome_calling_regions.v1.interval_list

[Vivek]

Thank you! This is what I was looking for. They don't have this for the latest release (0.2) but I guess its safe to assume it would be the same.