Quantifying not-normal reported variant frequencies

Hi! I’m trying to use gnomAD 3.1 to quantify the not normal frequencies of variants. In that sense, my general idea is that given one variant, e.g. A827G, I want to substract the counts corresponding to the studies coming from individuals that are assigned to haplogroups that are defined by that variant. I thought I could do these by checking phylotree variants assigned to each haplogroup, but I don’t know if this is a fine approach, plus there is the fact that in the case of A827G, this variant appears to be defining a subclade of B instead of being a top level marker and I’m unable to find deeper level of haplogroup detail in the reported counts coming from individuals assigned to the B haplogroup inside the gnomAD database.

Could you give me a hand to figure this out? Thanks in advance.