Friday, January 8, 2010
Deus ex Machina Genetics
Since I began speaking out about the genetics of violence, the entire field of genetics has reached a tipping point of high stakes and high drama. The stakes are that this field, which I would characterize as being presided over to some extent by ivory tower liberals, may be on the verge of proving once and for all that black people have a greater genetic predisposition to violence, or the field may wipe away this evidence along with a couple of decades of knowledge from gene association studies.
Prior to this scientific cataclysm, I received some interesting feedback from a presentation on the genetics of black violence that I created. When the blog Half Sigma reported on the recent study that associated gang violence with the 3-repeat allele in the promoter of the violence gene MAOA, the readers immediately wanted to know about racial prevalences of the alleles. Henry Harpending, the famous researcher who has studied recent human evolution and who co-wrote a new book with Gregory Cochran, pointed to the Sabol et al study, which determined that two-thirds of white people have the 4-repeat allele that is not associated with violent behavior and that 60% of African-Americans and Asian-Americans have the more violence-associated 3-repeat allele.
As one of the few studies of MAOA that lays out such racial prevalences, this study is quite an enigma to me. One might expect a study of nearly 1,500 subjects to more accurately reflect reality than smaller studies, but the study apparently discovered no subjects with the 2-repeat allele, which has a markedly greater association with violence than the 3-repeat allele that has received so much attention. Other studies, like one of 1,100 American males by Guo et al, found that 1% of American males possess this allele. MAOA is on the X chromosome, of which females have two, giving the Sabol et al study well over 2,000 chromosomes to examine. Therefore, I would have expected about 20 or so chromosomes to have the 2-repeat allele, instead of none.
I point all this out because Harpending, for one, relied on Sabol et al to describe the allele frequency of the MAOA promoter in black people. Other studies have replicated the allele frequency of Sabol et al for white males, always with about 1% having the 2-repeat allele. However, two smaller studies offered evidence of a significant proportion of black people having the 2-repeat allele. After I showed the Half Sigma readers my video, someone who goes by the pseudonym, Nanonymous, drew my attention to an important hidden fact about one of the two smaller studies. It was the study by Widom and Brzustowicz, detailing prevalences of rare MAOA alleles, including the 2-repeat allele. The study considered one of its major findings to be a lack of an association between violence and the 3-repeat allele just in non-whites, but it only had 98 non-white male subjects. Then, it claimed to re-analyze the data just for black subjects with “no change in the results.” We do not receive any hints as to how many of the 98 non-whites are black. However, as Nanonymous pointed out, some of the subjects in this 2006 study also took part in a 1989 study by Widom. That study did not look at genetics, but it did reveal that its 1,575 subjects were 67% white and 31% black, leaving just 2% for other groups. For the 2006 study, just over 600 agreed to undergo DNA analysis out of the original 1989 cohort plus 1,196 additional subjects. They were 60.9% non-Hispanic white and 39.1% “non-white.” Having successfully covered up the proportion who are black, Widom and Brzustowicz felt comfortable revealing that 6% of the non-white men had the ultra-violent 2-repeat allele, and none had the other rare alleles. In 2006, Rosenberg et al found that 5 of 37 (13.5%) African-Americans had rare MAOA alleles, which was corroborated by a greater frequency of associated rare haplotypes of the MAOA gene, itself. In general, Africans would be expected to have more rare alleles because non-Africans descend from only a subset of the African population.
Gene Expression, picked up the story, with the writer known as P-ter declaring that the majority of gene association studies, including that of MAOA, are likely wrong. He later posted that correct associations found in genome-wide association studies tend to require thousands or tens of thousands of subjects, making studies with just hundreds of subjects essentially worthless. P-ter singled out for criticism Avshalom Caspi, the researcher who first published an association study for the serotonin transporter gene and depression and who first published an association study comparing the MAOA 3- and 4-repeat alleles’ influence on violent delinquency. “[E]verything he’s published is probably wrong,” P-ter sneered.
I was the first to comment, and I conveyed my shock. “So, I guess you don’t believe in Brunner syndrome, either,” I said. Brunner syndrome was first identified in 1993 as a point mutation that caused complete MAOA deficiency, and it is associated with aggressive outbursts, arson, attempted rape, and exhibitionism. P-ter tried to deny that the existence of Brunner syndrome is sufficient to lend credence to the influence of the promoter alleles. I reminded him that the MAOA Caspi et al study was not a study of the 2-repeat allele (since only one subject had it), and this allele has consistently shown a greater association with violence than the 3-repeat allele. Therefore, MAOA expression has a continuum from Brunner syndrome, to the 2-repeat allele, to the 3-repeat allele, and, finally, to the 4-repeat allele. However, P-ter may be right that the many other phenotypes studied in association with the MAOA promoter may be disproved.
I did my own review of the scientific literature that specifically sought to replicate the Caspi et al study of the 3- and 4-repeat alleles of MAOA in white males. Twelve total studies sought to replicate the study in one manner or another. Only three of these had unequivocally non-significant results. Confirmatory studies raised questions about the methods of those studies, such as the narrow definition of the environmental exposures. A study by Sjoberg et al found an association between the 3-repeat allele of MAOA and aggression when testosterone levels are high. Thus, even though this study did not try to replicate Caspi et al, it gives additional indirect support by using an aggravating environmental exposure (testosterone), which can be objectively measured. All told, the studies that looked at the interaction between the 3- versus 4-repeat allele genotype and maltreatment had a combined sample size of over four-thousand white male subjects. Granted, a serious meta-analysis would exclude some studies. Kim-Cohen et al attempted a meta-analysis, but it included its own cohort, and rather than look at an outcome of aggression in that cohort, it noted an association with composite mental health problems, and ADHD, in particular.
During the discussion on the Gene Expression blog, the Australian researcher, Daniel MacArthur, first hedged, “As for MAOA: I wouldn’t bet my house on this not turning up in [a] well-powered GWAS [genome-wide association study] for violent behaviour (as P-ter predicts), but I’d say the odds of this happening are much better than 50%.” He later conceded “that the evidence for an association between the VNTR [variable number of tandem repeats] variant and antisocial behaviour is substantially more consistent than most of these associations. This may well be one of the rare cases of genuine associations.…” I am pleased if a rank amateur, like myself, can help to convince a group of researchers not to glibly dismiss the past seven years of research on a gene for violent delinquency just due to an embarrassing failed association in the study of depression. I noticed that the original study on Brunner syndrome found that dexamethasone treatment increased MAOA expression of female carriers of the mutation. I really believe that this line of research will save lives. I was also happy to see contributions to the discussion from writers that I admire like Steve Sailer and Gregory Cochran.
This may indeed be a watershed moment that will lead to far more skepticism of future candidate gene association studies. Hopefully, replication studies will become more standardized. I look forward to reading any potential large genome-wide association studies for violent delinquency. P-ter envisioned that such studies would uncover multitudes of genes with each having very small effects. However, I agree with Cochran that violent behavior can be biologically advantageous. Plus, I have long been intrigued by the possibility of “master genes” or other ways of coordinating evolutionary change, so that life can adapt to changing environments.
I also think that it is important to be mindful of the politicization of science. Much of the articulation in the popular press of the notion that race is a social construct that I have come across was actually in the defense of the Human Genome Diversity Project. This project has sought to decode the genomes of all the diverse branches of our human race. However, some activists tried to shut it down out of a fear that biological weapons against specific peoples would result. Therefore, the defense that there, in fact, is no one gene for being black was put forth. Furthermore, as MAOA demonstrates, genetic racial differences are almost always just disparities of allele frequencies. Even so, those differences are enough to cause hysteria. Consider this statement at the end of Ellis and Nyborg’s study of racial differences in testosterone levels: “scientists should be on guard against even the hint of any misuse of research findings in this area.” I suppose it is a sort of misuse of science for one to childishly berate a group of people and identify with a “superior” group to compensate for one’s own inadequacies. The irony is that this statement was followed by an acknowledgement for the help given by J. Philippe Rushton, a scientist well-known for his courageous disregard for suppressive sensitivities concerning everything from brain size to penis length.
Other examples of science politicization abound. Last year, the blogger known as "gc" was concerned that the Obama administration might censor human biodiversity science. Harpending responded that he received an invitation from the National Human Genome Research Institute to attend a workshop about the ethics of researching and publishing evidence of natural selection in humans. Another example is how Bruce Lahn seemed to be under pressure when he decided to forswear IQ-related genetics research. Now I am wondering if the actions taken in the Widom and Brzustowicz study and the gratuitous linking to the MAOA-aggression association by Risch et al were further examples.
The Gene Expression blog has been an invaluable ally of human biodiversity research. The writers are highly intelligent freethinkers. The fact that some of them have some liberal-leaning views and were early supporters of Howard Dean is no more an incongruity with their race realism than is the Christian faith of Steve Sailer and Charles Murray. It would only help this line of research to make it a little more acceptable in polite (that is, liberal) company. On the other hand, I am noticing eroding anonymity for certain writers, which could be a problem for them because liberal academia, I believe, is their home. (Full disclosure: nooffensebut is not my real name.)
If, somehow, MAOA promoter alleles should ever be proven meaningless, as P-ter suggested they would be, I shall follow the scientific ideal described by Richard Dawkins in The Root of All Evil. I shall rejoice at having been proven wrong, such that scientific truth is advanced. Thankfully, we all believe in science.