[datafatmunger]

In this lecture, Sapolsky talks briefly about the genetic component but at a super high level "good/bad" version of a gene.

Depression is a genetic disorder. It has some degree of heritability. 50% identical twins, 25% full sibling. So again we have the whole nature-nurture interaction. Have the bad gene (a serotonin one), add in major stressors and uh-oh. A 30 fold increase in the likelihood of depression at the extreme. Oh, and glucocorticoids regulate the expression of the gene.

Anyone know what gene he is referring to?

[datafatmunger]

So, I'm guess maybe this one: https://www.ncbi.nlm.nih.gov/snp/rs25531#frequency_tab

...and despite: "According to David Hinds of 23andMe on community forums, "nearly everyone (99.97%) is getting called as CC, and there is no clear heterozygote cluster" ... "the genotype calls for rs25531 on our platform are not meaningful." Looking at OpenSNP frequency data, this seems to be universal for direct to consumer genotyping services at this stage.... "

... but went looking for this in my own data anyway, and couldn't even find the rs#.

disclaimer: I have no idea what i'm doing.

[nnd]

My guess would be SLC6A15 gene (check for rs1545843 and rs1031681 SNPs). (A;A) is a "bad version".