Why is there such a big difference between gnomAD v2.1.1 and v4.0.0 o/e?

[adriafarres]

Hello,

Thank you for the gnomAD 4 release. I have been checking the pLI and o/e of quite a few genes, and there seems to be pretty big differences for quite a few genes between both versions. Some examples:

	gnomAD 2	gnomAD 4
SKI	0.04	0.59
TAOK2	0.13	0.36
ACTN2	0.12	0.33

Why is that?

Some genes like ZNF804A seem to have way less observed SNVs in gnomAD 4 than gnomAD 2. In the case of ZNF804A, it's 12 vs 2. What could be the reason?

Thank you