In the meta analysis apart
In the meta-analysis, apart from Kano\'s study (Kano et al., 2012), we found that the significant associations of previous studies were marginal and changed along with the genetic effect models (Porcelli et al., 2015; Terock et al., 2018). Given the inconsistency in genetic effect models and the heterogeneity of meta-analyses, it was possible that the associations of 5-HTTLPR with alexithymia conditionally depended on genetic stratification (Koller et al., 2004) and noised variables such as demographic and disease-related factors of individual studies. As such, the false positive report probability, which yielded serious inferential errors (Lucke, 2009), was greatly increased for the individual studies. It seemed thereby that the superficial association partly attributes to the population stratifications (Koller et al., 2004; Qin and Zhu, 2017). In the previous studies, the genetic stratification in 5-HTTLPR (i.e., the frequencies of LL gamma-Glu-Cys synthesis was ranged from 0.04 to 0.36 in Asians and Europeans) among the individual studies was significant. Because the frequency of 5-HTTLPR varies across subgroups, most studies attempt to solve population stratification problem by controlling for self-reported ethnicity. However, the controlling for self-reported ethnicity may fail to control for population stratifications. The population stratifications therefore confounded the associations reported in genetic-behavioral studies when the empirical model fails to properly control for them. As findings of the replicated study, the deviation from Hardy-Weinberg equilibrium of the 5-HTTLPR was present, the deviation suggests a possible genetic stratification for the polymorphism in this sample. However, when this association was examined by meta-analysis with several types of subsamples, the population stratifications may be controlled with the random model of meta-analysis (Engels et al., 2000; Nakaoka and Inoue, 2009; Yang et al., 2018). As compared with the previous studies that only examined the role of S/L polymorphism (Ham et al., 2005; Kano et al., 2012; Porcelli et al., 2015), we also examined the effects of rs25531 and the combination effects of S/L polymorphism (5-HTTLPR) and rs25531. In the replicated study, we found that neither the main effects nor haplotypes of the two polymorphisms impact alexithymia. Moreover, as compared with the previous studies (Ham et al., 2005; Kano et al., 2012; Mandelli et al., 2013; Porcelli et al., 2015; Terock et al., 2018), we also examined the associations of 5-HTTLPR with alexithymia with different genetic effect models. The findings provide complete profiles on the association of 5-HTTLPR with alexithymia. The association of COMT Val158Met with alexithymia was also mixed (Ham et al., 2005; Min et al., 2016; Swart et al., 2011; Zekioglu et al., 2014). The effect of Met allele is small as well (Fig. 2). The power analysis indicated that a sample of 1500 participants was required when the effect size of genetic polymorphism (i.e., the differences in the score among the Met/Met, Val/Met, and Val/Val genotype groups) reached Cohen\' f = 0.1.The small sample sizes of previous studies would reduce the reliability of their conclusions. As compared with these studies with small sample sizes (Ham et al., 2005; Min et al., 2016; Swart et al., 2011; Zekioglu et al., 2014), we examined the association of COMT Val158Met polymorphism with alexithymia in a larger population. The sample size had 95% power to detect the small genetic effect. In this study, we examined the associations of the COMT Val158Met polymorphism with alexithymia with three genetic effect models. Additionally, as compared with the previous study (Ham et al., 2005), we also examined that the interaction of COMT Val158Met and 5-HTTLPR on alexithymia. It seemed therefore that we ruled out the possible impacts of this polymorphism on alexithymia. Of note, as compared with the 5-HTTLPR genotypes that were not in the Hardy-Weinberg equilibrium, the COMT genotypes were not the case. The distinctive distributions of Hardy-Weinberg equilibrium for the 5-HTTLPR and COMT Val158Met possible resulted from their different chromosome locations.