Genome-wide meta-analysis identifies novel risk loci for uterine fibroids within and across multiple ancestry groups

Abstract

Uterine leiomyomata or fibroids are highly heritable, common, and benign tumors of the uterus with poorly understood etiology. Previous GWAS have reported 72 associated genes but included limited numbers of non-European individuals. Here, we identify 11 novel genes associated with fibroids across multi-ancestry and ancestry-stratified GWAS analyses. We replicate a known fibroid GWAS gene in African ancestry individuals and estimate the SNP-based heritability of fibroids in African ancestry populations as 15.9%. Using genetically predicted gene expression and colocalization analyses, we identify 46 novel genes associated with fibroids. These genes are significantly enriched in cancer, cell death and survival, reproductive system disease, and cellular growth and proliferation networks. We also find that increased predicted expression of HEATR3 in uterine tissue is associated with fibroids across ancestry strata. Overall, we report genetic variants associated with fibroids coupled with functional and gene pathway enrichment analyses.

Fig. 1. Cohort sources and case/control counts for each meta-analysis.

AFR African Ancestry, EUR European Ancestry, EAS East Asian Ancestry, CSA Central South Asian Ancestry, Multi Multi Ancestry.

Fig. 2. Multi-ancestry genome-wide meta-analysis identified eight genes associated with fibroids unpublished in previous literature.

a Manhattan plot, gold labels = novel, black labels = previously unpublished. b Sentinel variant at novel gene results, asterisk = novel gene association. Logistic regression statistical tests; multiple testing correction p-value threshold used (5 × 10⁻⁸).

Fig. 3. European ancestry genome-wide meta-analysis identified four genes associated with fibroids unpublished in previous literature.

a Manhattan plot, gold labels = novel gene associations, black labels = previously unpublished genes. b Sentinel variant at novel gene results, asterisk = novel gene association. Logistic regression statistical tests; multiple testing correction p-value threshold used (5 × 10⁻⁸).

Fig. 4. East Asian/Central South Asian ancestry genome-wide meta-analysis results.

Significant genes were replications of previously identified associations. Logistic regression statistical tests; multiple testing correction p-value threshold used (5 × 10⁻⁸).

Fig. 5. African ancestry genome-wide meta-analysis identified one novel gene associated with fibroids.

a Manhattan plot, gold labels = novel gene associations. b Sentinel variant at novel gene result, asterisk = novel gene association. Logistic regression statistical tests; multiple testing correction p-value threshold used (5 × 10⁻⁸).

Fig. 6. The most enriched network from Ingenuity Pathway Analysis of significant colocalized genes was cancer, organismal injury and abnormalities, reproductive system disease.

Hypergeometric distribution testing with right-tailed Fisher’s Exact Test; p-value threshold used (0.05).