Is CAG sequence length in the androgen receptor gene correlated with finger-length ratio?

Manning, Bundred, Newton, and Flanagan reported a significant correlation of .29 in a sample of 50 British males between the length of a repeated sequence on the androgen receptor gene and 2D:4D finger-length ratio on the right hand. We report a 2nd failure to replicate this result. Ours was a sample of 182 Australian male twins studied for other purposes, for whom both measures were available. The result was a nonsignificant correlation of .055. A similar result was obtained for female twins, and for comparisons within sibling pairs. Correlations are also reported for left hands and right–left differences––the last showed a weak tendency toward replication. 2011 Elsevier Ltd. All rights reserved. John Manning and his colleagues (Manning, Bundred, Newton, & Flanagan, 2003) reported a significant positive correlation of .29 in a sample of 50 British men between the secondto fourthfinger ratio (2D:4D) on the right hand and the length of a repeated CAG sequence in the androgen receptor (AR) gene. Longer CAG sequences have been related to less effective transcription of testosterone (Chamberlain, Driver, & Miesfield, 1994). Thus the Manning et al. finding suggests that differences in sensitivity to androgens, as well as differences in androgen levels themselves, may be playing a role in determining the 2D:4D ratio. This would not explain sex differences, for which CAG sequence lengths do not differ, but might be related to individual differences within a given sex. In a recent review, Breedlove (2010) makes Manning’s correlation a centerpiece in his argument that digit ratios reflect prenatal androgens: ‘‘The strongest evidence that androgens affect digit ratios is the report that normal polymorphism in the androgen receptor (AR) gene correlates with digit ratios in men.’’ (p. 4117) The 2D:4D literature is replete with failures to replicate correlational relationships discovered in small samples (e.g. <100 cases). We are aware of one other effort to replicate this particular one (Hurd, Vaillancourt, & Dinsdale, 2011). It did not succeed: the authors found a correlation of only .006 between AR and righthand 2D:4D in 180 male Canadian undergraduates. At the time we learned of the Hurd et al. article, we had been carrying out an attempt at replication in an Australian sample of twins and their siblings that had been assembled for other purposes by Martin and his colleagues (Wright & Martin, 2004). For a subgroup of this sample, both digit ratios and AR gene repeats were available. Thus we could ask: What is the correlation between right-hand digit ratios and AR repeats for the males in this group? The Australian males were younger than Manning’s British males (mean age 17.9 as against 32.6 years)––as were the undergraduates in the Hurd et al. study (mean age 19.2 years)––but for a stable effect determined prenatally, age differences should not matter. It should be emphasized that we are not here addressing the broader issue of whether prenatal androgens are related to individual and sex differences in finger-length ratios. We are here concerned with a more limited question: are differences in the sensitivity to androgens, as reflected in the length of the repeated CAG sequence on the AR gene, associated with (and hence possibly responsible for) individual differences in finger-length ratios? In addition to right hands, Manning et al. (2003) reported for their sample the correlation of AR sequence length with left-hand 2D:4D (a trivial .005) and with the difference in 2D:4D between right and left hands (a significant .36). Hurd et al. obtained non-significant correlations for these of .12 and .13, respectively. These correlations also are available in the Australian sample. Because the Australian sample also contains females, we can compare results for them. Males are exposed prenatally to higher levels of androgens than females, and this is hypothesized to lead to their lower average 2D:4D ratios, but if differing finger-length ratios for individuals within each sex result from differing prenatal 0191-8869/$ see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2011.09.009 ⇑ Corresponding author. Tel.: +1 512 475 7008. E-mail addresses: [email protected] (J.C. Loehlin), [email protected] (S.E. Medland), [email protected] (N.G. Martin). Personality and Individual Differences 52 (2012) 224–227