> but a bitter lesson from the last decade for those who believed it was the major driver of disease is that the environment has a stronger effect.

It doesn't make sense to talk about whether genes or environmental factors have stronger effects in general: they're very very very strongly interacting. It's not phenotype = genes + environment, it's phenotype = stupidly_complicated_function_we've_only_just_begun_to_understand(genes, environment). You can talk about the effects of genetic variation in a particular environment, or of environmental variation given fixed genetics, but "the contribution from [genetics/the environment]" without any further qualification is not a thing that exists.

I know, but the dominant trend was to (try to) predict diseases by calculating genetic risk scores, ignoring any environmental contributions.

This trend is even known as genetic determinism. There is a whole generation of senior academics who are fiercely attached to this line of thought.

Personally I think nyssos's explanation (that disease is a complex function of the joint distribution of genotype and environment) is still more convincing, and that the reason environment temporarily got more attention is simply that our ability to understand causality of disease in non-mendelian contexts is still very limited (it being an extremely complex function).

I've worked in this field for a while and continuously get into arguments with geneticists, part of that is because I simply don't understand the explanations genetics gives (I'm a molecular biologist and computer scientist; the difference is that genenticists treat genotype to phenotype as a black box, while molecular biologists see it as an event system), but another part of it is that most scientists are highly susceptible to the streelight effect, and genetics is something tangible you can easily measure in the lab, while environment is... much more complicated.

I agree with your perspective, but would add a little nuance. I think the molecularly-mediated version of genetics is promising (e.g., Transcriptome Wide Association Studies, or TWAS). It breaks out of the black box a little, but still enriches for causal associations that purely mechanistic approaches miss. Pure molecular mechanism has a hard time tracing from cells to organismal phenotype at age 60 in a varied environment.

Do you do computational biology? I think that field is fascinating. I’m working on human biomolecular data viz tools in a consortium of academic labs (Harvard, CMU, etc.) funded by the NIH. Really enjoy working with computational biologist.

How much better could you predict if you sequence both DNA and samples of expressed RNA from various parts of the body?

Yes, exactly. Genetics is the difference between humans and fruit flies. All the radical reshaping we’ve done to the environment over the history of civilization would not have been possible without the intelligence given to us by our genes.

When people talk about genetic factors in life outcomes they always point to these very simple Mendelian diseases. They don’t point to more subtle traits that are caused by a wide variety of genes interacting. Just because we don’t fully understand these genetic mechanisms doesn’t mean they’re not there.

And as far as “nurture” goes, genes play a big role in that too. If a parent has a greater predisposition towards impulsivity or violence or substance abuse then that’s going to affect their fitness as a parent. It’s the Anna Karenina principle: everything has to go right if you want the best outcome.