br Results The AR genotype frequencies are shown in Table
Results The AR genotype frequencies are shown in Table 1. Alleles at the two loci are in linkage disequilibrium, as expected for such closely linked loci and consistent with previous findings (Kittles et al., 2001). The genotype frequency for the DBH 1021 polymorphism is similar to previous findings reported by Zabetian et al. (2001). The minor allele (DBH 1021T) had a frequency of 17% in our sample, which is close to the reported frequency in the Japanese population of 16%. Table 2 shows the mean scores on the personality measures for each DBH genotype. It can be seen that there were no significant differences, but there was a non-significant trend for P, with lower scores in the T/T group. For the AR polymorphisms, Table 3 shows the female means for each genotype and Table 4 the male means. The only significant effect for females was for the CAG repeat polymorphism and BAS Drive, where the long/long group had lower scores. For males, there was a significant effect for the CAG repeat polymorphism and P, with short AM679 being associated with higher P scores. There was also a significant effect for the GGC polymorphism and BAS Drive, with short alleles being associated with higher BAS Drive scores. There were no significant interaction effects involving the AR polymorphisms when both were considered together.
Discussion Zabetian et al. (2001) have reported that homozygosity for the T allele of the DBH C-1021T polymorphism is associated with very low DβH activity. Taken together with Zuckerman’s (1994) hypothesis that low plasma DβH is associated with P, we would predict lower P scores in people with the T/T genotype. However, we in fact found a non-significant trend in the opposite direction (P=0.099). The T/T genotype was uncommon, comprising only 2.1% of our sample (or 19 participants), which limits statistical power. However, given the trend in the opposite direction, it is unlikely that the small sample size prevented us from detecting the hypothesized effect. With the AR gene, previous research has found associations of the short alleles of both the CAG and GGC repeat polymorphisms with behavioural characteristics typical of high P. In the present study, the short CAG genotype was significantly associated with high P in males only. In females, the trend was for higher P in the group homozygous for short alleles, but this did not even approach statistical significance. For the GGC polymorphism, there were no significant associations with P, although the trend was for the short alleles to be associated with higher P. If there is a real association of AR polymorphisms with P, then the effects must be quite small given the sample size of the current study. For males, the power to detect a small difference (0.2 standard deviation units) between long and short alleles is approximately 81% for the CAG repeat polymorphism and around 74% for the GGC repeat polymorphism (P<0.05, 2-tailed) (Cohen, 1977). For females, the power to detect a small difference between the short/short and other genotypes is approximately 81% for CAG repeats and 80% for GGC repeats. In addition to these findings with P, there were some significant associations of AR polymorphisms with BAS Drive. For females, the homozygous short CAG group had higher Drive scores, while for males the GGC short allele was associated with higher scores. Neither of these associations was predicted. Given that 28 statistical tests were carried out, it is possible that these are Type I errors, particularly as associations were not observed with other BAS scales which are correlated with the Drive scale. Although the present results are negative, we believe it is important to report all findings to avoid biases in the literature. The present sample has also been genotyped for several other polymorphisms which might plausibly be associated with personality traits, but again there have been largely negative results. While some of these results have been previously published, others have not. In order to provide a complete record of the results, we have presented them as Appendix A to this paper. Only 2 of the 33 statistical comparisons in this appendix are statistically significant. One of these did not replicate across sub-samples (Henderson et al., 2000) and the other involved heterozygotes having a higher score, which is not biologically plausible.