Author: Nicole Ferraro, MS
Predicting phenotype from genotype has long been one of the main goals of human genomics. The completion of the Human Genome Project sparked hope that realizing this goal was just around the corner. However, current consensus is that most traits are influenced not just by one gene, but by many to most , and predicting them from an individual’s genetic composition requires data from across the genome. Enter polygenic risk scores (PRSs).
PRSs were covered in a past Nascent Transcript article , but briefly, the general idea is to learn weights on genotypes from a large cohort of individuals based on the correlation of that genotype with a trait of interest and then aggregate the weights x genotypes across a new individual’s genome to calculate risk of the same trait. Several papers have described potential clinical uses [3-5], but one application in particular has raised several ethical concerns [6-7]. Since our genetic code is essentially the same throughout life, an individual’s predicted genetic risk would be the same at age 2 as at age 50, assuming the same equation is used, which opens the door to calculating genetic risks in embryos.
Several companies are already offering PRS calculations to prospective parents, either as an embryo screening service for couples using in vitro fertilization (IVF) [8-9] or through simulating embryos from the genetic data of two partners , with embryo screening via IVF presented as a potential outcome, regardless of fertility. It is not unprecedented for parents to consider genetic risks they may pass to their children, but that is most routinely applied via carrier screening for recessive monogenic disorders like Cystic Fibrosis or Tay-Sachs disease where we can be quite certain that a child will develop the disorder if they inherit two copies of the risk allele. In contrast, many of the diseases included on the polygenic risk panels are usually developed later in life due to a combination of genetics and environment, like Alzheimer’s, breast cancer or heart disease.
Despite the nascent commercialization of PRSs in embryo screening, there are several limitations to this technology. A study assessing the utility of selecting embryos by the PRS-predicted height or IQ of the subsequent human  found that the adult trait had substantial variance, even for the top-scoring embryos and in evaluating large families, the offspring with the greatest predicted height was most often not the tallest. Furthermore, there are concerns about the generalizability of PRSs to populations that differ from the training population . These results are somewhat unsurprising, given the substantial environmental contribution to polygenic traits, but beg the question: why not instead dedicate resources to changing non-genetic factors we know also influence common disease risk across all people, like access to routine medical care, good nutrition, and clean air and water?
Limitations notwithstanding, there is at least some demand for these technologies as is, and so perhaps it doesn’t really matter if the individual predictive power is limited. But there can still be concerning implications. The American College of Medical Genetics and Genomics recommends reporting genetic test results if there exists a possible intervention , but does embryo selection constitute an intervention in all cases? If such testing gains in popularity, what will be the consequences both legally and socially of not participating? Embryo screening requires IVF, which is expensive, often both physically and mentally taxing, and does not guarantee a successful pregnancy. At what point does this outweigh the benefits of a few percentage point decrease in predicted risk for disease “X” for a couple that could conceive naturally? And how will parents be guided in making these choices, especially if presented with risk estimates varying in uncertainty across many traits? There are also strong concerns about further stigmatization of those living with various conditions that might be included on a screening panel .
These questions may not have obvious answers, but regardless, PRSs are beginning to be applied in reproductive decisions. As a graduate student in a human genetics lab, I have begun to consider more and more the implications of my work in this context. Does the chance that publishing on predicting risk of a polygenic disease from genetics could lead to its adoption in embryo screening impact how we approach studying the genetics of that disease? I hope to continue seeing discussion on what regulations and protections should be in place regarding PRS applications in reproduction and encourage my fellow ASHG trainees to be active participants in those discussions.
 Boyle, E. A., Li, Y. I., & Pritchard, J. K. (2017). An expanded view of complex traits: from polygenic to omnigenic. Cell, 169(7), 1177-1186.
 Wray, N. R., Lin, T., Austin, J., McGrath, J. J., Hickie, I. B., Murray, G. K., & Visscher, P. M. (2020). From basic science to clinical application of polygenic risk scores: a primer. JAMA psychiatry.
 Yanes, T., McInerney-Leo, A. M., Law, M. H., & Cummings, S. (2020). The emerging field of polygenic risk scores and perspective for use in clinical care. Human Molecular Genetics, 29(R2), R165-R176.
 Lewis, C. M., & Vassos, E. (2020). Polygenic risk scores: from research tools to clinical instruments. Genome medicine, 12, 1-11.
 Lewis, A. C., & Green, R. C. (2021). Polygenic risk scores in the clinic: new perspectives needed on familiar ethical issues. Genome Medicine, 13(1), 1-10.
 Lázaro-Muñoz, G., Pereira, S., Carmi, S., & Lencz, T. (2020). Screening embryos for polygenic conditions and traits: ethical considerations for an emerging technology. Genetics in Medicine, 1-3.
 Karavani, E., Zuk, O., Zeevi, D., Barzilai, N., Stefanis, N. C., Hatzimanolis, A., … & Carmi, S. (2019). Screening human embryos for polygenic traits has limited utility. Cell, 179(6), 1424-1435.
 Martin, A. R., Kanai, M., Kamatani, Y., Okada, Y., Neale, B. M., & Daly, M. J. (2019). Clinical use of current polygenic risk scores may exacerbate health disparities. Nature genetics, 51(4), 584-591.