Selecting embryos for complex traits – what to expect?

Shai Carmi ¹ Ehud Karavani ¹ Or Zuk ² Danny Zeevi ³ Gil Atzmon ^4,5 Nir Barzilai ⁵ Nikos Stefanis ⁶ Alex Hatzimanolis ⁶ Nikolaos Smyrnis ⁶ Dimitrios Avramopoulos ⁷ Leonid Kruglyak ³ Max Lam ^8,9 Todd Lencz ^8,9,10

¹Public Health, The Hebrew University of Jerusalem, Israel
²Statistics, The Hebrew University of Jerusalem, Israel
³Human Genetics, University of California, Los Angeles, USA
⁴Biology, University of Haifa, Israel
⁵Medicine, Albert Einstein College of Medicine, USA
⁶Psychiatry, University of Athens Medical School, Greece
⁷Psychiatry, Johns Hopkins University School of Medicine, USA
⁸Psychiatry Research, Zucker Hillside Hospital, USA
⁹Psychiatric Neuroscience, Feinstein Institute for Medical Research, USA
¹⁰Psychiatry, Hofstra Northwell School of Medicine, USA

Background: Polygenic score predictors with increasing accuracy have been recently developed for human complex traits. Recent progress in the ability to genotype IVF embryos thus suggests the possibility of screening embryos for complex traits such as height or cognitive ability. However, no study has evaluated the expected outcomes of such a procedure, which undermines attempts for a well-informed discussion of the associated ethical, legal, and societal issues.

Methods: We used simulations, theory, and real data to evaluate the potential gain of embryo selection, defined as the expected difference in trait value between the top-scoring and the average embryo.

Results: The gain increases slowly with the number of embryos, but more rapidly with the variance explained by the score. Given currently available predictors and typical IVF yields, the gain would be ≈2.5cm if selecting for height, and ≈2.5 IQ points if selecting for cognitive ability. These mean values are accompanied by wide confidence intervals due to random assortment, unaccounted-for genetic factors, and environmental factors. In 28 nuclear families with up to 20 adult offspring each, the offspring with the highest score for height was the tallest only in 7 families, and was shorter than the family average in 5 families.

Discussion: With current predictive accuracy of quantitative traits, embryo selection has limited utility; however, gains will increase with GWAS sample sizes. In practice, utility will be limited by assortative mating, aneuploidy, implantation failure, lower accuracy in non-European populations, antagonistic pleiotropy, and selection for multiple traits.