In this issue of the JCI, Chavli, Klaasen, and colleagues (9) provide insight that profoundly affects the debate on the application and benefit of PGT-A. Their single-cell, whole-genome sequencing (sc-Karyo-Seq) analysis provides compelling evidence that more than 80% of high-quality human IVF embryos exhibit chromosomal defects in a mosaic distribution, often affecting only a small proportion of cells.

The power of the single-cell sequencing approach means that the chromosome status of each individual cell of entire blastocyst-stage embryos could be analyzed (9). This method contrasts with prior bulk-sequencing techniques that instead sequence pooled cells from whole embryos or segments of embryos. The bulk approach is subject to dilution effects, so that loss or gain of DNA cannot be detected when only 20% or fewer of the cells are abnormal. It also misses some types of anomalies in which reciprocal changes between cells cancel out one another (10, 11). This means that earlier studies using bulk sequencing have routinely undercounted the incidence of chromosomal anomalies and have classified embryos as entirely normal when they may not be (Figure 1).

Different genetic analysis platforms have a varying capacity to detect mosaic aneuploidy in blastocyst-stage embryos. Embryos may be euploid (100% normal cells), aneuploid (100% abnormal cells), or mosaic aneuploid, with varying proportions of cells in the inner cell mass and/or trophectoderm affected by numerical and/or structural chromosomal defects. sc-Karyo-Seq analysis of all cells recovered from blastocysts offers high sensitivity and specificity, accurately detecting mosaic aneuploidy in even a small minority of cells. Bulk sequencing techniques analyze entire embryos or embryo parts and have lower sensitivity and specificity than sc-Karyo-Seq, only detecting aneuploidy when at least 20% of the cells in a sample are aneuploid. PGT-A is a clinical test utilizing bulk sequencing to analyze small segments of four to six trophectoderm cells recovered from embryos via biopsy. PGT-A frequently fails to correctly classify aneuploid status in mosaic aneuploid embryos due to low sensitivity of bulk sequencing techniques and spatial heterogeneity of chromosome status, meaning that the segment is not informative of the entire embryo. When none or less than 20% of cells in a segment are aneuploid, the test may return a negative or inconclusive result even when aneuploidy exists. When more than 20% of cells in a segment are aneuploid, the test may return a positive result even if most of the cells in an embryo are normal.

The Chavli et al. study (9) convincingly demonstrates that embryos with chromosomal abnormalities are far more common than previously thought. They sequenced approximately half of the cells from each of 55 good-quality surplus blastocyst-stage embryos and found that at least 45 of the embryos (82%) contained a proportion of cells with numerical and/or structural DNA faults consistent with mosaic aneuploidy. The majority of these embryos (n = 32, 58%) exhibited diploid-aneuploid mosaicism, in which, on average, 60% of cells were normal, indicating a mitotic origin. Defects were similarly common in the trophectoderm layer and inner cell mass compartment of embryos and often affected 20% or fewer of the cells in an embryo — an incidence that is not reliably detectable by bulk DNA-sequencing techniques. A range of chromosomal defects was evident, with similar frequencies of numerical and structural abnormalities. Sequencing results were not obtained for all cells because of technical challenges with cell recovery, and since embryos in which aneuploidy was not found often had lower cell recovery rates, the authors considered it likely that the true rate of mosaic aneuploidy in embryos is even higher than 82%.

The study used good-quality blastocysts made using a high standard of clinical protocols, and the experiments and analysis were technically and scientifically robust. Therefore, we find it compellingly reasonable to expect that mosaicism occurs with similar frequency in all IVF clinics around the world. However, more studies using single-cell approaches will now be necessary to confirm the generalizability of these findings.