variation in amygdala activation and variation in a length polymorphism of the serotonin transporter gene (5-HTTLPR) [15]. The ‘short’ allele at this frequently typed polymorphism has a frequency of about 30% in European populations, and is thus a typical common variant, whose effect size on complex phenotypes we would expect to be small (explaining less than one percent of the variation in a quantitative measure). The short allele is reported to lower transcriptional efficacy (hence reduce levels of serotonin transporter protein). To obtain the degree of significance reported in the 2002 paper, the locus must explain about 28% of phenotypic variance (95% confidence intervals 15–53%). This effect size is indeed much higher than anything reported from genetic analyses of disease phenotypes, but is it likely to be true? Given the results of the imaging GWAS [11•,12•,13•,14•], the answer is that it is almost certainly not.
