One possible confounder: methylation capability.

Once ingested B6/12/folate needs to be methylated into an active form. Different people have different methylation capabilities, pretty much dividing the population into quartiles.

As I understand it, MTHFR is controlled by 2 SNPs so the combinatorics 'fits' quartiles.

On top of that, people with poor methylation capability could be getting sufficient in their diet (e.g. phosphatidylcholine from egg yolks, trimethyl glycine from beets) and have no issue with the conversion.

Those B vitamins are also available in already-methylated form. I looked in the methods section of the linked study and couldn't find if that was what was used - or not. Just something to be aware of.