Genome-wide analyses identify 25 infertility loci and relationships with reproductive traits across the allele frequency spectrum

Abstract

Genome-wide association studies (GWASs) may help inform the etiology of infertility. Here, we perform GWAS meta-analyses across seven cohorts in up to 42,629 cases and 740,619 controls and identify 25 genetic risk loci for male and female infertility. We additionally identify up to 269 genetic loci associated with follicle-stimulating hormone, luteinizing hormone, estradiol and testosterone through sex-specific GWAS meta-analyses (n = 6,095-246,862). Exome sequencing analyses reveal that women carrying testosterone-lowering rare variants in some genes are at risk of infertility. However, we find no local or genome-wide genetic correlation between female infertility and reproductive hormones. While infertility is genetically correlated with endometriosis and polycystic ovary syndrome, we find limited genetic overlap between infertility and obesity. Finally, we show that the evolutionary persistence of infertility-risk alleles may be explained by directional selection. Taken together, we provide a comprehensive view of the genetic determinants of infertility across multiple diagnostic criteria.

Fig. 1. Overview of study cohorts and analyses for infertility genetic association studies.

a, The case numbers in each cohort contributing to GWAS meta-analyses (MA) for female (left) and male (right) infertility. The prevalence of all-cause infertility in each cohort (%) is noted on the bar plots. Danish, Danish Blood Donor Study/Copenhagen Hospital Biobank. Total case and control counts for each type of genetic analysis: all-ancestry GWAS meta-analysis, EUR-only GWAS meta-analysis and WES analyses (discovery, UKBB and replication, G&H and deCODE) are displayed. Male infertility in deCODE, with <100 cases, was excluded from GWAS meta-analysis. Note the different y-axis scales in each subplot. b, Downstream analyses performed for each type of genetic analysis: lead variants were identified via distance-based pruning for all-ancestry and EUR-only GWAS meta-analyses; colocalization, genetic correlations (genome wide and local), genetic overlap and selection analyses were only performed for EUR meta-analyses due to the need for ancestry-matched LD information; rare-variant and gene-burden discovery tests were performed with WES data for the UKBB EUR-ancestry subset and replicated in individuals with WES data in G&H and whole-genome sequencing (WGS) data in deCODE.

Fig. 2. Miami and Manhattan plots for selected infertility meta-analyses.

a, Genetic variants associated with F-ALL (top) and idiopathic infertility (unknown causes) defined by exclusion of known causes such as anatomical or anovulatory causes, PCOS, endometriosis and uterine leiomyomas (bottom). b, Genetic variants associated with M-ALL. Each point depicts a single SNP. The triangles represent SNPs that only reach genome-wide significance in all-ancestry GWAS meta-analyses. SNPs are annotated with the mapped gene. *The lead variant is reported in only one cohort. Summary statistics from whole-genome regression analyses were meta-analyzed using fixed-effect inverse-variance weighting in the METAL software to produce the displayed P values. The dashed line represents the multiple testing-corrected P value threshold of P < 5 × 10⁻⁸, accounting for ~1 million independent variants in the genome.

Fig. 3. Genetic correlations between female infertility and other phenotypes.

SNP-based genetic correlations (r_g) between significantly heritable phenotypes (Z > 4) were estimated using LD-score regression, performed using the LDSC software on a subset of 1 million HapMap3 SNPs. The points are colored by r_g estimate, scaled by significance (−log₁₀(P)), and labeled with the associated r_g estimate if nominally significant without correction for multiple testing (P < 0.05). a, Genetic correlations among three definitions of female infertility (F-ALL, F-ANOV and F-INCL). b, Genetic correlations between female infertility traits and reproductive hormones testosterone, FSH and AMH (publicly available summary statistics) in female-specific analyses and TSH (publicly available summary statistics) from sex-combined analysis. c, Genetic correlations between female infertility traits and female reproductive conditions, with summary statistics generated from the largest available EUR-ancestry studies for each trait (Methods). d, Genetic correlations between female infertility traits and selected heritable phenotypes (Z > 4) in the UKBB, as generated by the Neale laboratory. Correlations with all heritable phenotypes can be found in Supplementary Table 12.

Fig. 4. Local genetic correlations and polygenic overlap between female infertility and other phenotypes.

a, Local genetic correlations, estimated using LAVA, at 2,495 blocks across the genome. Each point represents a local bivariate genetic correlation between an infertility trait (F-ALL, F-ANOV or F-INCL) and reproductive hormone, reproductive condition or obesity-related trait. The dashed lines indicate significance (sig.) thresholds. The dashed line represents FDR-adjusted or Bonferroni-adjusted P values of 0.05. b, MiXeR estimates of polygenic overlap. The Venn diagrams indicate the estimated number (s.e.m.) of causal variants (in thousands) that explain 90% SNP heritability per component. The circle size reflects the degree of polygenicity. The bars outline the genome-wide genetic correlation (rG) and correlation in the shared polygenic component (rho). The colored portion of the bar is sized by the proportion of causal variants in the shared polygenic component as compared with all causal variants involved and colored by rho. Comp., comparative.

Fig. 5. Directional selection scores at infertility-associated EBAG9 locus.

Recent directional selection, as measured by trait-aligned SDSs (tSDSs) at the EBAG9 locus. The window of ±10 kb around the lead variant associated with F-ALL is displayed, along with the location of nearest gene TSSs. The tSDSs are aligned to the infertility risk-increasing allele, wherein a positive tSDS indicates positive selection for infertility risk-increasing allele at the locus. The dashed lines indicate 2.5th percentile (%ile) and 97.5th %ile of SDSs. Left: a locus plot depicting genomic position on the x axis and tSDS on the y axis. The lead variant rs1964514 (open circle) is not present in the tSDS dataset and thus is assigned a score of 0. Right: a scatter plot depicting relationship between −log₁₀ of the GWAS P value for the variant association with F-ALL on the x axis and tSDS on the y axis.

Fig. 6. Rare variants associated with testosterone and infertility in UKBB WES analyses.

a, The mean effect size versus allele frequency of genetic variants associated with total testosterone estimated using regression analyses. Variants discovered at genome-wide significance (P < 5 × 10⁻⁸) in GWAS meta-analyses (n_female = 235,579, n_male = 235,096) and exome-wide significance (P < 1 × 10⁻⁷) in the UKBB WES analyses (n_female = 197,038, n_male = 197,340) are plotted. The effect sizes are aligned to the minor allele, plotted against MAF on the log x axis. b, The effects of testosterone-associated rare variants (chr:pos:minor allele:major allele) on infertility in females (left: n cases/controls for F-ALL = 3,746/260,413; n cases/controls for F-EXCL = 3,012/261,147) and males (right: n cases/controls for M-ALL = 650/222,393) estimated using regression analyses. The effect sizes are aligned to the minor allele. Per gene, the variant with the lowest P value of all variants that reach exome-wide significance in UKBB WES analyses for testosterone is displayed, for all variants with nominally significant effects on infertility. Effect sizes (β and 95% CIs) for the variant effect on testosterone are to the left of each plot and effect sizes (ORs and 95% CIs) for the variant effect on infertility are to the right of each plot.