Abstract Study question What is the proportion of the genome where the egg and the first or second polar bodies descend from the same maternal homolog? Summary answer The egg is about 30% identical to a random haploid genome from the first polar body and about 40% identical to the second polar body. What is known already After crossing over, human female meiosis generates three products: the first polar body (PBI) splits first with two sister chromatids per chromosome; the second polar body (PBII) and the egg split next, each with a single set of chromatids. When meiosis proceeds normally, the egg must be identical (i.e., same maternal homolog) to PBII at the centromeres. However, the proportion of the entire genome where the egg is identical to PBII (or to a random chromatid set from PBI), while being a fundamental property of meiosis, is not intuitively clear and has not been empirically studied. Study design, size, duration We used both theoretical and empirical approaches. We first derived formulas and performed simulations for the identity between the egg and the polar bodies, with or without modeling crossover interference. We then analyzed genotype data from seven egg/PBI/PBII trios and 43 triploid embryos. Finally, we analyzed 8,476 dizygotic twin pairs from research-consented 23andMe, Inc. customers in an attempt to find PBII twins, who would (hypothetically) share maternal genetic material in all centromeres. Participants/materials, setting, methods The egg/PBI/PBII trios came from three female donors and were genotyped at about 300k SNPs (Ottolini et al, Nat Genet, 2015). Embryos were sequenced to depth about 0.05x for routine PGT-A and identified as triploid upon retrospective analysis with LD-PGTA (Ariad et al, Genome Res, 2023). Coverage-based analysis confirmed triploidy predictions for XXY embryos. The DZ twin pairs were genotyped at about 640k SNPs and self-reported dates of birth within two days of each other. Main results and the role of chance In an infinitely long chromosome, the egg would be identical to PBII, or to a random chromatid from PBI, in only a third of its length. Modeling recombination using the human female genetic map and assuming no interference predicted that the egg would be on average 53% identical to PBII and 24% identical to a random haploid genome from PBI. We confirmed these estimates with simulations. However, in seven egg/PB1/PB2 trios, the egg was on average only 40% identical with PBII. Similarly, in 17 triploid embryos with an MII (likely-maternal) error, the two transmitted maternal chromosome sets were on average only 37% identical. Including crossover interference in the simulations reduced the identity between the egg and PBII to 42%, closer to the empirical observations. Hypothetical PBII twins would occur by two sperm from the same father fertilizing both the egg and PBII. They would share, based on the results above, an average of about 45% of their genomes. Such twins would also have identical maternal genetic material in all centromeres. In 8,476 genetically dizygotic twin pairs, only 67 shared at least one haplotype in all centromeres, not significantly more than expected by chance for dizygotic twins. Limitations, reasons for caution The empirical results are based on a small number of egg/PBI/PBII trios (n = 7) from only three donors. The number of triploid embryos was larger, but they were sequenced to very low depth, and the results for MI-error embryos showed much less sharing between the two maternal chromosomes than expected. Wider implications of the findings We show the fundamental and perhaps counter-intuitive result that the egg is only 30-40% identical to the polar bodies. This provides a reference against which deviations from normal crossover patterns could be evaluated in triploid and other aneuploid embryos. Polar body twins, if they occur, must be extremely rare. Trial registration number Not applicable
Read full abstract