What prevents us from developing tools, whether statistical techniques or otherwise, from settling these questions once and for all? And why do the answers need to be universal? Most papers I've read do not universalize their results, at least outwardly, but (some) popular science writers seem to do that a lot.
I can imagine, and several other comments allude to this, that individuals with different genetic makeup will experience a different set of dynamics around nutrition. I don't mean to be facetious or to suggest these aren't real challenges, I'm just curious on what the big obstacles are, in 2020.