Infertility affects 1 in 6 couples globally, and is linked to all major health challenges of the 21st century, from cardiovascular disease to mental health. Yet, despite its prevalence and broad consequences, major determinants of infertility remain understudied. The cause of infertility remains unexplained in up to 40% of women and 28% of couples.
For us as geneticists, this presented a clear challenge: how can we treat conditions whose biology we don’t understand?
Large-scale studies of infertility are rare. Infertility is often missing or under-coded in healthcare records, hormone data is patchy, and male infertility remains particularly underreported. Even more challenging, from a genetic standpoint, is the fact that evolutionary selection tends to limit the frequency of risk variants for reproductive traits. If we could overcome these barriers, we believed that studying the genome could uncover not just causes of infertility, but deeper connections to other health conditions, and possibly even reveal new therapeutic targets.
This was the motivation behind our project, which turned into a collaboration across six countries, seven biobanks, and more than a million participants. Our team based in Oxford, UK analysed the UK Biobank dataset of nearly 500,000 participants, collaborated with two other UK-based studies, Genes and Health and the Avon Longitudinal Study of Parents and Children, and teams across the globe from deCode (Iceland), the Estonian Biobank (Estonia), FinnGen (Finland), and the Copenhagen Hospital Biobank and Danish Blood Donor Study (Denmark). This was all made possible through generous funding support from the Wellcome Trust, NIH, and the Gates Foundation - in addition to all the funding bodies supporting each of our collaborators.
The scale of the effort was extraordinary, and so was the collective enthusiasm. Clinicians, geneticists, statisticians, and bioinformaticians brought their expertise together to answer this fundamental question in human biology.
Along the way, the project grew in both size and scope. When we presented initial findings at the American Society of Human Genetics conference, we were approached by a researcher working on the genetic basis of twinning. Her interest in the link between twinning and infertility led us to a striking result of a strong negative genetic correlation between anovulatory infertility (when the ovaries don’t release eggs) and dizygotic twinning (when two eggs are released simultaneously). It was a moment that beautifully illustrated how new insights can emerge from conversation and collaboration.
Later, while our paper was under review, the Million Veteran Program (USA) released their incredible resource of summary statistics for thousands of conditions studied in nearly a million participants. This addition strengthened our analysis of male infertility and led to another intriguing finding. A common variant near UMODL1, a gene expressed along the migratory path of GnRH neurons - the same pathway implicated in Kallmann syndrome, a genetic cause of hypogonadotropic infertility - was associated with common male infertility!
Among the findings we’d like to highlight is the 71% genetic correlation between endometriosis and unexplained infertility. It suggests that undiagnosed endometriosis could be a major contributor to infertility. We also found a much lower genetic overlap between polycystic ovary syndrome (PCOS) and anovulatory infertility than expected, only about 40%, potentially challenging the assumption that PCOS accounts for over 80% of anovulatory infertility cases.
Another surprising result was the limited genetic overlap between female infertility and obesity, despite clinical emphasis on weight loss as a treatment strategy. Our results suggest a more complex, bidirectional physiological relationship. Similarly, we didn’t observe strong genetic correlations between infertility and reproductive hormones. However, this may reflect differences in when and how hormone levels were measured, especially given that many hormone assays came from post-menopausal women, while infertility affects those of reproductive age.
Finally, on an evolutionary note, we found hints of recent positive selection near some infertility-associated variants, suggesting they may persist in the population due to pleiotropic effects. One such example is a variant near EBAG9, linked to infertility but also to the immune response to infection, possibly explaining its evolutionary persistence.
Ultimately, this study underscores how much remains to be learned about infertility and its links to broader health. The correlations between infertility and women’s reproductive health conditions speak to long-standing issues of underdiagnosis and misdiagnosis. We hope these findings will prompt clinicians and researchers alike to continue pushing forward on important research in women’s health.
Prof Ingrid Granne from the University of Oxford, who was not involved in the original study, says “This study offers valuable genetic insights into infertility, which could ultimately transform our approach to diagnosis and treatment. While more research is needed, these findings highlight important steps toward improving clinical care and addressing gaps in our understanding of reproductive health, particularly for women whose health has long been underserved in medical research.”
Indeed, this work is only the beginning. There’s a vast frontier ahead, from understanding long-term health risks, to identifying new drug targets, to better capturing reproductive hormone dynamics across the menstrual cycle. And there are still fascinating biological and evolutionary questions around the coding of infertility risk in our genomes.
Thanks to more than a million participants, and a global team committed to collaboration, we’re now a little closer to the answers.