Statistical Genetics Update

Just a few things from Razib's interview with Tade Souaiaia.

We have long known that empirically the distribution tails of many traits are fatter than they are "supposed" to be.  Adding to the puzzle is that some traits are fat only at one end. The very approximate new rule of thumb is that the regular distribution for traits might apply for 98-100% of the population in the huge "middle," but there is a nontrivial chance that things have gone a bit screwy at one or both ends.

For a polygenic trait such as height, there are people who fall into the extreme 1% on each end in the usual way, that they have a larger number of tallness traits and fewer average or shortness traits. In addition to that group, there are tall people that got there via one gene, usually a mutation affecting the pituitary, like Andre the Giant. At the shorter end, there are those who got there with a polygenic score for shortness, and also some who inherited a single gene for dwarfism or related condition. With a single deleterious mutation - and many are de novo in that generation or a parent's - it is 50-50 whether a child inherits it. Thus you might get two children in the normal distribution, and two others very short, rather than four generally-shorter children if it were all polygenic.

Tangent: I heard for the first time that sitting height is less affected by these single-gene exceptions, while leg length is more affected. I knew of individual examples where this was true - namely, myself - but didn't know it was a recognised statistical phenomenon.  I am 5'8".  My best friend in highschool was 6'7".  We sat at nearly the same height, so the difference was in the legs.  He bought 37" inseam pants which his mother hemmed at the cuff with great difficulty, while mine were nearly 10" shorter.  If you see me in shorts you immediately wonder what is wrong with that man's tibias? Senior year I had a girlfriend who was 4'11", which was comic when the three of use were walking or standing, but she also sat much taller. Sitting in the bleachers, you could tell from the back she was shorter than us, but not that you'd necessarily notice more than the usual girl-slightly-shorter expectation.

Intelligence is only slightly fattened at the upper end.  There is no single "genius" gene. But there are any number of conditions that are intelligence destructive, fattening the lower tail. You will find families with generally depressed scores, just as you find families where all the kids are smart. But a child at the extreme lower end is more likely to have siblings who are average, while children just a bit better off in the IQ department are likely to have other siblings who also lag. Children at the 90th percentile have siblings that cluster around the 73rd, not the 50th. Those at the 99th percentile have siblings around the 84th.

Enter autism. Because there is this pattern in ASD, researchers are increasingly convinced that there are two autisms, though the mechanisms are not clear. Polygenic autism is what I generally talk about here, of engineer's and librarian's kids who actually find personal adaptations that give them an advantage in some domains while still being impaired in others: overliteral, misinterpreting social cues, low thresholds for emotion, stimuli, and chaos. Single-mutation autism would be much more likely to result in the severely impaired. I wish I had coined the term benign autism 10-15 years ago.  Not that polygenic autism is always or even often benign, and not because I would be doing a great service to the autism community by reducing stigma, but because it's a great title for an airport book and I could have retired earlier.

I haven't finished the podcast and will get to it tonight on my way to the Tolkien class.  That part is going to be on race and sports, always controversial and fun. It is one of the clearest issues of things everyone knows but is not allowed to say. Over a billion people in China but only a few marginal players in the NBA (except the intentionally-bred Yao Ming).  About 20 million from former Yugoslavia, but over a 100 players, including two of the top five currently. It's all about culture, sure.