The SNP selection threshold producing the allelic score that explained the most variance in the intermediate differed across the variables. This is not surprising since different phenotypes have different underlying genetic architectures and differ in the extent to which the variants that influence them are tagged by SNPs on genome-wide chips. In general, the best allelic scores for BMI were produced using liberal cut-offs. In contrast, allelic scores for CRP and LDLc tended to perform best when constructed from conservative threshold cut-offs. A possible reason for the discrepancy is that CRP and LDLc are both influenced by loci of major effect in ALSPAC (i.e. the variants rs4420638 (R2 = 2.5%, p = 1.7×10−25) and rs10401969 (R2 = 2.0%, p = 1.1×10−20) explain disproportionately large proportions of the variance in LDLc, whilst the variants. rs2794520 (R2 = 2.4%, p = 5.2×10−24) and rs4420065 (R2 = 2.3%, p = 4.5×10−23) explain large portions of the variance in CRP- see also the last two columns in Table S7). At stringent p value cut-offs therefore, these scores primarily reflect genuine quantitative trait loci of
