Review: CORRELATED GENOTYPES!

I'm not sure what the best format is conveying this information, so we will make changes as necessary.

The Introduction
First,  what is being addressed in this paper? Fowler et al. are looking to investigate whether "variation in specific genes might be relevant to the process of friend selection or, distinctly, whether the process of choosing friends might result in correlated genotypes between friends." Specifically, they are examining "phenotypic similarity between individuals connected in a social network is reflected in their genotypes."  For those who are not sure what a "genotype" is, for this article they are referring to a DNA sequence at a particular location in the human genome. Also, a phenotype is the manifestation of having a particular genotype. Phenotype can be mean hair color and height and things like that, but it can also mean intelligence, age to puberty, and other such things.
 The authors point out that there are genetic correlations between kin groups and this has been show in many diverse organisms, but this may not be the "only basis on which natural selection might possibly operate at the group level."  The possibility of  a non-kin selection basis has long been postulated, but no data has previously been reported. If we understand the phenotype to be under the direct action of one's own genome, then it would certainly be remarkable if our friend's genomes was shaping our phenotype.
 It is a known fact that people associate with people that they resemble (phenotypically similar to one another), so one needs to be aware of situations in which phenotypic similarity can result without being caused by influence from friend genomes.
1. A trivial by-product of a group of people being unable to move away from one another. Imagine an isolated tribe in the Amazon.
2. "People may actively choose friends of a similar genotype." Of course, people cannot see their friend's genotypes, but phenotype is a direct consequence of genotypes and as such they could be picking a certain set of genotypes. The authors give an example of skinny people befriending skinny people, which percludes genetic based obesity phenotypes.
3. Behaviorial phenotypes may induce one to seek out environments where they will encounter people of similar phenotypes and by extension similar genotypes. The author's note that this could occur with distance runners who join clubs or go to events where they will meet other distance runners.
4. The environment of similar phenotypes may pull a similar phenotype in as is the case with college admission.

The Data
 Let me upfront about this, I have very little training in statistics and as such this paper is very difficult to analyze. So, I will probably just tell you what they did, but I cannot go into more than that.
 The researchers started with the Add Health dataset (National Longitudinal Study of Adolescent Health) which is data from a study that "explores the causes of health related behavior of adolescents in grades 7 through 12 and their outcomes in young adulthood."  In this study, each student nominated up to 5 male and 5 female friends and the details of those friendships. Every student and their friends were swabbed and genotyped for the following genes:
CYP2A6- cytochrome p450 2A6
DRD2- dopamine D2 receptor
DRD4- dopamine D4 receptor
MAOA- monoamine oxidase A-uVNTR
SLC6A3- Dopamine transporter
SLC6A4- Serotonin transporter

 I cannot show you the social network figures from the actual paper, because of copyright issues and my lack of what is ok to publish, but they show two social networks showing postive genotypic correlation for DRD2 and negative genotypic correlation for CYP2A6. The trend that they saw was replicated by statistical analysis of a different dataset (Framingham Heart Study Social Network) where they saw similar trends of similar magnitude.

The Discussion
 It is possible that the trends seen are because of association between genes, but this is unlikely because associating genes are very rare.
 Why were these genotypes postitively correlated (DRD2) or negatively correlated (CYP2A6)? The authors point to evidence that DRD2 has been associated with alcoholism and it s fairly straightforward to see that alcoholics are avoided by non-alcoholics and that alcoholics associate in environments that non-alcoholics avoid.
 The personality traits attributable to CYP2A6 are not known at this time, but it has been implicated in personality traits.
 This research highlights the overestimation of genetic effects that has occurred because previous research was not aware the effect that friend genotypes could have on one's phenotype. The author's give an example of a woman predisposed to drink because of her own genotype, but also because of her friend's genotypes, which makes them more likely to drink.
 The most interesting idea presented, at least to me, is the idea of humans as "metagenomic." This means that "we could possibly view an individual's genetic landscape as a summation of the genes within the individual and those around him, just as in certain other organisms."
 Overall, I found this to be a great read and would very much like to understand the statistics behind the conclusions.  The field is still young and there are many questions to ask!