I spend a lot of time in a journal club discussing human behavior genetics with researchers who include researchers on extreme old age. That's a data-mining study, and would not impress other researchers in the field (or even get published in a better journal) without being replicated in an independent dataset. Spurious gene-correlation studies of this kind are a dime a dozen.

Food for thought is that the heritability of longevity as discovered in twin studies is actually not particularly high.