Table 1.
Novel low frequency variants at fasting proinsulin loci previously identified by genome-wide association studies
SNP | Gene | Variant | Chr | Position | Minor/major allele |
MAF | β̂ ± SE | Effect size in SD units ± SE |
Proportion of trait variance explained |
P value | Conditional P value |
---|---|---|---|---|---|---|---|---|---|---|---|
rs61741902 | SGSM2 | V996I | 17 | 2,282,779 | A/G | .014 | .126 ± .021 | .41 ± .07 | .0047 | 8.7 × 10−10 | 4.8 × 10−10 |
rs35233100 | MADD | R766X | 11 | 47,306,630 | T/C | .037 | −.100 ± .013 | −.32 ± .04 | .0075 | 7.6 × 10−15 | .0001 |
Chr, chromosome; MAF, minor allele frequency; SE, standard error; SD, standard deviation. Positions are from NCBI Build 37 with allele labels from the forward strand. Fasting proinsulin was log-transformed and adjusted for fasting insulin, BMI, age, and age2. Effects are reported for the minor allele.β̂ coefficient units are ln(pmol/l). Conditional P values are reported after adjusting for the lead SNPs from GWAS signals (rs4790333 at SGSM2; rs7944584 and rs1051006 at MADD). Full results of conditional analysis are provided in Supplementary Table 4. For rs35233100, effect size in SD units (± SE) and proportion of trait variance explained after adjusting for rs7944584 and rs1051006, are −0.17 (± .05) and 0.0007, respectively. Based on analysis of 8,224 non-diabetic males.