"The reason the effects are so small is because what the gene does is so far removed and from whatever phenotype you're studying."
... unless you study a medically relevant phenotype, in which case the genes are often just a few steps away (namely, transcription and translation, with processing thrown in for good measure).
I would agree that heritability is not all that meaningful (though technically valid) when talking about behavioral traits, though.
There's no reason that medically relevant phenotypes would be that much less complicated (and lets remember that mental health, is, in fact, medically relevant.)
Yes, there are a select few, rare, single nucleotide polymorphisms that are responsible for congenital genetic defects. But plenty of disease involves the interactions of thousands of genes, like cancer, and for those it's really just as meaningless.
... unless you study a medically relevant phenotype, in which case the genes are often just a few steps away (namely, transcription and translation, with processing thrown in for good measure).
I would agree that heritability is not all that meaningful (though technically valid) when talking about behavioral traits, though.