Should I switch to the latest version of gnomAD?

The gnomAD v2 call-set contains fewer whole genomes than v3.1, but also contains a very large number of exomes that substantially increase its power as a reference in coding regions. Therefore, gnomAD v2 is still our recommended dataset for most coding regions analyses. However, gnomAD v3.1 represents a very large increase in the number of genomes and may be more suitable if your primary interest is in non-coding regions, or if your coding region of interest is poorly captured in the gnomAD exomes. This can be assessed using the coverage plots in the gnomAD browser. Most genomes in v2.1.1 are included in v3.1 and should therefore not be considered independent sample sets. We are currently assessing the v3.1 dataset, but do not have a defined deadline for implementation in AlamutTM Visual Plus.

Another consideration when choosing which gnomAD dataset to use is the ancestry of the samples that you are interested in. gnomAD v3.1 contains a substantially larger number of African American samples than v2 (exomes and genomes) and for the first time provides allele frequencies for the Amish population. gnomAD v3.1 also has a fully genotyped call-set available from the Human Genome Diversity Project and the 1000 Genomes Project, representing >60 distinct populations.

Finally, gnomAD v3.1 was mapped to GRCh38. So, if your data is on this build, it probably makes sense to switch to v3.1. There is also a liftover version of gnomAD v2.1.1 onto GRCh38 available. gnomAD plans to produce a larger GRCh38-aligned exome call set in 2022.