Table 3.
Significant novel associations in the PheWAS meta-analysis
| Novel association in meta-PheWASa | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Gene | SNP | EA (EAF)b | Known associated phenotypec | Phenotype | OR (CI95) | P value | Directiond | N cases | N controls |
| CD226 | rs763361 | T (0.47) | IBD | Hypothyroidism | 1.05 (1.04–1.07) | 8.11e−11 | ++?+? | 35,428 | 412,577 |
| GDF15 | rs17724992 | A (0.73) | BMI | Heart metabolic diseasee | 1.03 (1.02–1.04) | 3.08e−09 | +???? | 275,944 | 209,302 |
| High blood pressuree | 1.03 (1.02–1.04) | 7.64e−09 | ++??? | 151,511 | 465,686 | ||||
| Blood pressure medicatione | 1.03 (1.02–1.04) | 1.76e−07 | +???? | 125,406 | 394,753 | ||||
| GERD | 1.03 (1.02–1.04) | 6.11e−07 | +???? | 130,654 | 384,572 | ||||
| Any CVDe | 1.03 (1.01–1.04) | 1.40e−06 | +???? | 148,577 | 388,405 | ||||
| IFIH1 | rs1990760 | T (0.61) | T1D | Asthma f | 0.96 (0.95–0.98) | 1.11e−07 | − − − − − | 57,101 | 269,659 |
| IRF5 | rs10488631 | C (0.11) | SLE | Hypothyroidism | 1.08 (1.05–1.12) | 5.78e−07 | ++?+? | 23,182 | 236,240 |
| PNPLA3 | rs738409 | G (0.33) | ALT | Severe acne | 0.91 (0.88–0.93) | 1.47e−11 | −???? | 14,812 | 187,018 |
| High cholesterol | 0.96 (0.94–0.97) | 1.59e−07 | − −??? | 101,646 | 180,947 | ||||
| TYK2 | rs34536443 | G (0.89) | Psoriasis | Any immune disorder | 1.10 (1.07–1.13) | 4.27e−12 | +???? | 112,148 | 173,986 |
| Hypothyroidism | 1.14 (1.08–1.20) | 1.19e−06 | ++?−? | 23,145 | 233,757 | ||||
ALT: alanine aminotransferase, BMI: body mass index, CVD: cardiovascular disease, EA: effect allele, EAF: effect allele frequency, GERD: gastroesophageal reflux disease, IBD: inflammatory bowel disease, SLE: systemic lupus erythematosus, T1D: type diabetes, T2D: type 2 diabetes
aAssociations reaching P < 1.8e−6 (Bonferroni-corrected significance threshold) in the meta-analysis of PheWAS results with GWAS results. The full list of potential novel SNP-phenotype pairs reaching FDR < 0.1 is provided in Supplementary Table 5. Novel associations with direction of effect opposite to the known associated disease(s) effect, predicting potential adverse drug events, are highlighted in bold
bThe effect allele is the risk allele for known associated disease(s) related to the therapeutic hypothesis
cKnown associated disease related to the therapeutic hypothesis (surrogate for efficacy). The strongest association reported in the literature is indicated. The full list of known associations is provided in Supplementary Table 1
dDirection of effect in 23andMe, Genomics plc UK Biobank, FINRISK, CHOP, and GWAS
eCorrelated phenotypes
fMeta-analysis results including the 23andMe, Gplc/UK Biobank, FINRISK, CHOP, and GWAS Gabriel cohorts. When further including the independent GWAS EVE study, the association reaches P = 6.7 × 10−8