Reinforcement

In 1995, another team from the same laboratory as the group that published the Hamer et al. 1993a study, replicated their colleagues’ paper. Their results supported the original conclusion that male homosexuality was transmitted down the maternal line and was linked to the q28 region of the X chromosome (Hu et al., 1995). Sixty seven percent of homosexual male sib-pairs from 33 families shared all of their Xq28 markers, a value significantly greater than the null hypothesis of 50 percent sharing by chance (P<0.05) (Hu et al., 1995).  

The level of allele sharing between the sib-pair probands and their non-homosexual brothers was also determined (Hu , 1995). This followed criticisms (Fausto-Sterling and Balaban, 1993) that no data was provided for this relationship in the original Hamer paper (1993a). Twenty two percent of heterosexual males shared Xq28 markers with their homosexual brothers (the sib-pair), a value significantly less than the null hypothesis of 50 percent sharing by chance (<0.05) (Hu ., 1995). This suggests that homosexual brothers might be in some way genetically distinct from their heterosexual brothers, further reinforcing the argument for the presence of a gene that influences sexual orientation amongst men.  

Hu (1995) acknowledged that, as with the original study by Hamer and colleagues (1993a), the use of sib-pair families places limitations on the possibility of extending the results to the population at large. The value of results that are not representative of an entire population should be questioned. Sib-pair families already fit the pattern of maternal transmission that the researchers hope will be assigned to the gene in question. Moreover, the number of homosexual siblings in sib-pair families is significantly higher than the background population rate. This sampling method can only increase the likelihood of finding maternally-transmitted, genetic similarities amongst homosexual family members. 

Replication

In 1999, Rice et al. published a paper claiming to have found no linkage between the occurrence of homosexuality in males and loci in the Xq28 chromosomal region (represented by four genetic markers). Pedigree analysis was performed using questionnaires to assess the sexual orientation of the probands. Conversely, the sexual orientation of their relatives was assessed not by direct questioning, but via often unreliable information volunteered by the proband. This instantly appears to be a flaw in the experimental design of this study. Additionally, Rice et al. (1999) did not use quantifiable and consistent Kinsey scales to assess the sexual orientation of any of their subjects, unlike both previous studies (Hamer et al., 1993a; Hu et al., 1995).  

The level of allele sharing at position Xq28 was then determined for 52 homosexual male sibling pairs using DNA linkage analysis (Rice et al., 1999). This represented a larger sample size than that of Hamer et al. (1993a). Homosexual brothers shared 46 percent of the alleles at this position, a non-significant result. Using strong, almost accusatory language, Rice et al. (1999) stated that they could provide no explanation for the significant discrepancy (P<0.001) between their allele sharing results and those of Hamer et al. (1993a).

Conflicting Views

A heated written debate was published in 1999, concerning the relevance of the Xq28 region to male homosexuality. Responding to the criticisms made of his earlier study by Rice . (1999), Hamer (1999) used the outcome of a ‘meta-analysis’ to justify his original conclusion that homosexual brothers shared a significant number of alleles in the Xq28 region. The results of four studies into the genetic basis of homosexuality were combined to provide a larger and more reliable sample of families (Hamer 1993a; Hu 1995; Sanders , 1998 cited in Hamer, 1999; Rice , 1999). The outcome of this meta-analysis was significant, with a 64 percent linkage (= 0.0001), suggesting that the total research thus far found linkage between the Xq28 region and male homosexuality (Hamer, 1999).

Rice, Risch and Ebers (1999) retaliated to Hamer’s criticisms (1999), suggesting that his meta-analysis should not combine the results of the four studies as two of them were from his own group and this did not address the issue of replicating his results correctly. Combining the two studies not undertaken by Hamer’s group (Sanders , 1998, cited in Hamer, 1999; Rice , 1999) gave a value of 56.6 percent, not significantly different to the null hypothesis of 50 percent sharing by chance ( >0.05). The existence of a genetic basis to male homosexuality was therefore rejected, but the dispute was still not resolved conclusively: “the basis of sexual orientation remains uncertain” (Rice, Risch and Ebers, 1999).