Table 1.
Associations with fat distribution in the exome-wide gene-burden analysis
| Gene | Variants contributing to burden test | Genetic exposure, variant type; frequency cutoff in % | Beta (95% CI) per allele in SD units of BMI-adjusted WHR | P | AAF, fraction of 1 | Genotype counts (RR | RA | AA genotypes) |
|---|---|---|---|---|---|---|
| ACVR1C 2: 157526766 | 173 | pLOF plus deleterious missense (5/5); AAF < 1% | −0.16 (−0.19, −0.12) | 3.1 × 10−20 | 0.0025 | 615,316 | 3057 | 2 |
| CALCRL 2: 187343128 | 311 | pLOF plus deleterious missense (1/5); AAF < 1% | −0.087 (−0.11, −0.06) | 1.5 × 10−10 | 0.0038 | 613,641 | 4730 | 4 |
| PPARG 3: 12287367 | 327 | pLOF plus deleterious missense (1/5); AAF < 1% | 0.14 (0.089, 0.18) | 1.3 × 10−08 | 0.0012 | 616,856 | 1519 | 0 |
| STAB1 3: 52495337 | 970 | pLOF plus deleterious missense (5/5); AAF < 0.1% | −0.065 (−0.086, −0.045) | 2.8 × 10−10 | 0.0067 | 610,105 | 8262 | 8 |
| PLXND1 3: 129555174 | 1425 | pLOF plus deleterious missense (1/5); AAF < 1% | −0.03 (−0.042, −0.019) | 7.3 × 10−08 | 0.0231 | 589,953 | 28,329 | 93 |
| CD36 7: 80369574 | 525 | pLOF plus deleterious missense (5/5); AAF < 1% | 0.048 (0.031, 0.066) | 6.8 × 10−08 | 0.0090 | 607,219 | 11,126 | 30 |
| ABCA1 9: 104781001 | 880 | pLOF plus deleterious missense (5/5); AAF < 1% | −0.056 (−0.074, −0.038) | 5.1 × 10−10 | 0.0087 | 607,690 | 10,668 | 17 |
| AIFM2 10: 70098222 | 321 | pLOF plus deleterious missense (1/5); AAF < 1% | 0.049 (0.036, 0.063) | 2.1 × 10−12 | 0.0145 | 600,542 | 17,782 | 51 |
| PDE3B 11: 14643722 | 281 | pLOF plus deleterious missense (5/5); AAF < 0.1% | −0.18 (−0.22, −0.15) | 1.4 × 10−22 | 0.0020 | 613,713 | 2459 | 0 |
| INHBE 12: 57455320 | 29 | pLOF; AAF < 1% | −0.17 (−0.22, −0.11) | 1.8 × 10−09 | 0.0009 | 614,471 | 1096 | 1 |
| PLIN1 15: 89664364 | 118 | pLOF plus deleterious missense (5/5); AAF < 1% | −0.2 (−0.23, −0.17) | 4.6 × 10−32 | 0.0025 | 615,348 | 3021 | 6 |
| ANKRD12 18: 9136776 | 156 | pLOF; AAF < 1% | 0.31 (0.22, 0.4) | 1.6 × 10−11 | 0.0003 | 615,156 | 412 | 0 |
| PLIN4 19: 4502179 | 195 | pLOF; AAF < 1% | 0.11 (0.079, 0.14) | 3.7 × 10−13 | 0.0031 | 614,492 | 3874 | 9 |
| INSR 19: 7112254 | 215 | pLOF plus deleterious missense (5/5); AAF < 1% | −0.075 (−0.094, −0.055) | 1.2 × 10−13 | 0.0069 | 609,823 | 8514 | 38 |
| KEAP1 19: 10486119 | 396 | pLOF plus deleterious missense (1/5); AAF < 1% | 0.089 (0.066, 0.11) | 3.4 × 10−14 | 0.0051 | 612,099 | 6266 | 10 |
| SLC5A3 21: 34073569 | 191 | pLOF plus deleterious missense (5/5); AAF < 1% | 0.06 (0.041, 0.078) | 4.7 × 10−10 | 0.0077 | 608,903 | 9442 | 30 |
| Women-only analysis | ||||||
| FGF1 5: 142592177 | 153 | pLOF plus deleterious missense (1/5); AAF < 1% | −0.083 (−0.11, −0.051) | 2.8 × 10−07 | 0.0045 | 352,178 | 3183 | 6 |
| MSR1 8: 16107877 | 165 | pLOF plus deleterious missense (5/5); AAF < 1% | −0.071 (−0.096, −0.045) | 9.8 × 10−08 | 0.0068 | 350,564 | 4786 | 17 |
The table reports genes for which the burden of rare nonsynonymous variants was associated with BMI-adjusted WHR at exome-wide statistical significance (IVW meta-analysis, p < 3.6 × 10−7). Analyses were performed in 618,375 individuals from UKB, MDCS and MCPS. Effect sizes in ratio units can be obtained by multiplying the effect sizes in SD units by 0.08 ratio units. Genomic coordinates reflect chromosome and position in base pairs according to the Genome Reference Consortium Human Build 38. AAF was derived by dividing the number of alternative alleles observed for a particular gene-burden by the total number of all alleles observed for that gene-burden. P-values are from two-sided Z-tests from fixed-effect meta-analysis.
CI confidence intervals, SD standard deviations, BMI-adjusted WHR waist-hip ratio adjusted for body mass index, P P-value, AAF alternative allele frequency, RR reference-reference homozygous genotype, RA reference-alternative heterozygous genotype, AA alternative-alternative homozygous genotype, pLOF predicted loss of function, Missense (5/5) missense variants predicted to be deleterious by 5 out of 5 in silico prediction algorithms, Missense (1/5) missense variants predicted to be deleterious by at least 1 out of 5 in silico prediction algorithms.