Yesterday (February 28) the Lancet published a study called “Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis.” The study was conducted by the Cross-Disorder Group of the Psychiatric Genomics Consortium.
The Psychiatric Genomics Consortium (PGC) was formed in 2007. Its purpose is “…to undertake meta-analyses of genome-wide association studies (GWAS) for psychiatric disorders…” The authors state that they have no conflicts of interest.
On their home page, the PGC state:
“The PGC has received funding from many sources. Before participation in the PGC, establishing and genotyping of the primary case-control collections were funded by a wide range of national, international, and commercial funders. Funding for the PGC was for data analytical efforts. The PGC has relied heavily on the goodwill of its members and their donated effort. We are deeply grateful to the National Institute of Mental Health (NIMH), the Netherlands Genetic Cluster Computer, and Hersenstichting Nederland for their sponsorship of the PGC.”
It is noteworthy that some of the finances for the base studies came from “commercial” sources. In addition, Hersenstichting Nederland, on its website acknowledges that it receives donations, but doesn’t identify any major donors.
So it’s possible that there’s been some contamination of the original research studies by pharmaceutical funders. This point is important, because we know that pharma contamination of research is not uncommon in the psychiatric field.
In the present meta-analysis, the PGC focused on five “mental disorders:” autism, ADHD, bipolar disorder, major depressive disorder, and schizophrenia. They accumulated genetic and “diagnostic” data on 33,332 subjects and 27,888 controls. They then looked for single-nucleotide polymorphisms (SNPs – pronounced ‘snips’) that were present in the subjects but absent in the controls. (An SNP is a DNA variation.) These were genome-wide scans. As of June 2012, 187,852,828 SNPs had been identified on the human genome. Some of these SNPs are associated with disease; others are associated with normal traits.
The authors were looking specifically for SNPs that were present in all five “diagnostic” categories as opposed to earlier studies that had focused mainly on single “diagnoses.”
The researchers found that: “SNPs at four loci surpassed the cutoff for genome-wide significance…” The researchers conclude that: “…specific SNPs are associated with a range of psychiatric disorders of childhood onset or adult onset.”
Two of the four SNPs are in locations known to be associated with voltage-gated calcium channels in the brain.
DISCUSSION
Firstly, the opening statement on the PGC’s home page is: “The purpose of the Psychiatric Genomics Consortium (PGC) is to conduct meta-analyses of genome-wide genetic data for psychiatric disease.” In other words, they are pre-committed to the notion that the categories listed in the DSM are diseases. Accordingly, and perhaps inevitably, they view the four SNPs which reached statistical significance as pathological variations, rather than normal trait variations.
Secondly, genes determine anatomical structure. Biological function, of course, arises from structure, but is modified by environmental factors. Behavior also draws from structure, but the link is becoming increasingly tenuous and unreliable. For instance, we humans walk upright because of the configuration of our legs, arms, torso, etc. (structure), but the fact that a person might walk in a military style, say, is more a function of his training (i.e. environment) than his genes. Indeed the statement: ‘he walks in a military style because of his genes’ would, I think, generally be considered silly. As behavior becomes more complex the genetic impact is increasingly diluted.
Thirdly, if you study the genetic make-up of almost any group of people who frequently engage in a common activity, you will probably find a significant correlation to a genetic factor if your sample size is large enough. Consider the example of people who play American football. It’s obvious to anybody with even a passing familiarity with this sport that the players have two noteworthy characteristics: they are above average in size, and they are extremely brave. Now I know very little of biology or genetics, but I would be surprised if there were not SNPs, or more likely groupings of SNPs, that correlate with these two traits. The association might be quite small, but if your sample size is 33,000 (as in the PGC’s study) even very small effects will achieve statistical significance. But nobody would conclude from this that playing football is an illness or even a symptom of an illness. Nor would we say that people play football because of their genes. And we would recognize that lots of big, brave people don’t play football.
Fourthly, the associations found by the PGC researchers, although statistically significant, were apparently quite modest. “As in almost all G(enome) W(ide) A(ssociation) S(tudies) of complex disorders reported so far, the effect sizes of genome-wide significant loci are individually quite small and the variance they account for is insufficient for predictive or diagnostic usefulness by themselves.” (p 8 of the study)
Fifthly, there is an underlayment of genetic involvement in all behavior. You can’t pick up a pencil or stab someone with a knife if your genetic material doesn’t provide for an opposable thumb. You can’t become depressed if your genetic code has no provision for the production of emotional apparatus. And so on. To find slight associations between SNPs and behavior simply reflects the fact that the genome is the organism’s blueprint.
Sixthly, calcium plays an important role in regulating neuronal excitability. The voltage-sensitive calcium channels modulate the amount of calcium available to the neuron which, in turn, modulates the neuronal excitability which, again in turn, could lead to behavioral excitability in certain circumstances. But neither behavioral excitability nor its absence is an illness. In and of themselves they are not even problems. They are variations of normal, though in certain circumstances they could become problems.
A headline in yesterday’s New York Daily News said: “Five most common mental illnesses share a common genetic basis….”
It might more accurately have said: A non-homogeneous group of people who are defined by the presence of vaguely-defined behaviors have been found to have certain genes with greater frequency than people who are not members of this group.
A complication in this whole area arises from the fact that some (in my view a very small minority) of the individuals who attract psychiatric “diagnoses” do indeed have something wrong with their brains. I can’t actually prove this, but I would be very surprised if it were not so. So we’re always going to find slight correlations between particular “diagnoses” and physical factors such as genes, neurotransmitters, etc… But the vast majority of “diagnosed” individuals have nothing wrong with their brains. It is fallacious to assume an underlying physical pathology based on a person’s behavior. If I walk with a limp for instance, it may be because of a genetic defect. But it might also be because my father walked with a limp, and I modeled my walking behavior on him; or because I’m trying to find an excuse to avoid work; or because I’m acting in a play; or because I’m a physiotherapist and trying to get a feel for my clients’ difficulties; or because my shoes are too tight; or because I’m intoxicated, etc., etc…