We plotted OXTR methylation as a function of substance use initiation at age 13 ranging from 2 SD below to 2 SD above the sample mean. As can be seen in Figure 2B, the significant interaction effect resulted from a significant slope for the association of early initiation of substance use with methylation of OXTR among those carrying the “s” allele (b = .20, p = .01), but a negatively-signed and non-significant slope among those with only the long allele (b = −.05, p = .51). The grey area in the figure shows that for those with an “s” allele, earlier initiation was associated with greater methylation at OXTR.
