This morning my attention was brought to an interesting piece in The New York Times, The Feel-Good Gene. It marshals an impressive array of scientific disciplines, genetics, biochemistry, and psychiatry. Concurrently Linda Avey on her Facebook page pointed to where you could find your genoptype for the SNP in question in 23andMe . Because it is geared toward a popular audience without scientific training the original article in The New York Times is a bit vague about the allele and SNP code. But Avey seems correct in her inferences, because the piece does cite an article in Nature Communications, FAAH genetic variation enhances fronto-amygdala function in mouse and human. So as the article notes the gene is FAAH, but it makes clear that the SNP is Rs324420. If you look up this variant you see that it is associated with many variations in phenotype. The results for the recent paper from last week, cited in the article, indicate that the mutant/derived allele, A, confers a reduction in anxiety in both humans and mice (the gist of the article). In certain contexts because of the reduced anxiety in individuals who carry the A allele they are better at learning, because they more rapidly “update” to conditions where their initial fear response was an overreaction. At this point I will tell you that I am “wild type.” In other words, I carry the ancestral allele, C, and am of homozygous genotype for that allele. In fact both my parents are homozygotes for the C allele. Since this is the majority allele in most populations that is not entirely surprising.
In passing the author of The New York Times piece observes that “roughly 21 percent of Americans of European descent, 14 percent of Han Chinese living in China and 45 percent of Yoruban Nigerians have been found to carry this gene variant.” These data are from the HapMap. According to them 45 percent of the Yoruba Nigerians in the sample carry the A allele, which means that presumably they are less anxious, on average, than the Chinese. Perhaps this aligns with some expectations or stereotypes you might have?
Well, I decided to check a wider range of populations, because it literally takes about 2 minutes. First, I looked in the HGDP browser. You can see that there is wide variation in African populations. The Mbuti Pygmies of Central Africa have a far lower proportion of the A allele than most European groups, in line with a few East Asian groups, as well as Papuans. Across much of Western Eurasia you see many populations at intermediate frequencies, but it looks as if in Europe there is a cline of A from north to south. One can confirm this with the much larger sample sizes of the 1000 Genomes browser. The highest proportion of A, which purports to be associated with reduced anxiety, is found among Finns, at 29 percent. Then there is the British sample at 24 percent. Finally, both the Spanish and Italian (Tuscan) samples return 16 percent. The South Asian groups all exhibit frequencies of A between 15 and 25 percent (at 18 percent for Bangladeshis, it makes sense why my family would be homozygote for C).
What about the phenotype in question? I’ll skip over the biochemistry, though it isn’t too difficult in terms of pathways. The original article states that “one community-based study of almost 2,100 healthy volunteers found that people with two copies of the mutant gene had roughly half the rate (11 percent) of cannabis dependence than those with one or no mutant gene (26 percent).” That’s not a trivial sample size, but it’s one study. But, it dovetails with the overall thesis, that the biochemical priors of individuals with the AA genotype might allow them to avoid addictions to “abused drugs, like cocaine, opiates and alcohol.” I’m a little confused about this assertion because I looked up the SNPedia page for this variant, and the first major association of an AA genotype is with alcoholism and drug use! Additionally, I did not mention one thing when surveying populations earlier: the groups with the highest frequency of the A allele are Amerindians. This is clear in the HGDP and the 1000 Genomes results.
The point of this post is not to suggest that variation within the FAAH locus is not relevant to phenotypic differences in individuals or populations. There’s a lot of epidemiological, and now molecular, biochemical, and neurological, evidence that this missense mutation is important in a functional sense. It is likely to make a difference in outcomes. In The New York Times piece the author speculatively suggests that variation at this SNP somehow perpetuates personality heterogeneity in our species, and is a boon to a society. Granted, this doesn’t seem to be true in all cases, as the Mbuti Pygmies and Papuans may lack polymorphism here. But, it is interesting to me that the derived mutation is found at variable frequencies all across the world. There’s probably a evolutionary and biomedical story here to be told about some sort of balancing selection. But, as with many narratives which are fixated upon endophenotypes, the scientific conclusions aren’t quite cut and dried, and rather are still developing, because the endophenotypes themselves are at the end of a long causal chain.