SAN FRANCISCO (GenomeWeb) – Fast, point-of-care preimplantation genetic screening could improve the success of in vitro fertilization procedures by not requiring embryos to be frozen, according to researchers from Columbia University Medical Center.
Zev Williams, chief of the division of reproductive endocrinology and infertility at Columbia, and colleagues have developed a PGS method on Oxford Nanopore's MinIon device that could reduce testing time to around four hours, from one week to 10 days. In a proof-of-principle study published on the preprint server BioRxiv this month, the team showed that the method yielded the same results as Illumina's VeriSeq PGS test for chromosomal aneuploidies.
In the study, the researchers tested only nine embryos, and Williams said that the next step would be to do a larger trial to validate the method before eventually offering it as a clinical service. The work adds to two other studies that Williams' team has published on nanopore sequencing over the last two years, including a 2016 study published in Genetics describing an approach to do short-read sequencing on the MinIon and a study published this month in G3: Genes, Genomes, Genetics in which the group described a modified library prep and multiplexing for the MinIon.
Williams said the goal of the work is to continue to improve the success rate of IVF. NGS-based methods have helped by identifying which embryos are chromosomally normal, he said, but current methods take around one week or longer to deliver results. "Because of that delay, embryos are frozen," he said, and the "woman has to wait for the next menstrual cycle before an embryo is implanted."
Williams added that although it's unclear how big of an impact freezing and thawing embryos has on their viability, it likely contributes at least somewhat to IVF failure. In addition, he said, a method that can be done at the bedside also has the potential to be less costly, since it avoids freezing costs and enables the whole procedure to be performed during the same visit.
Williams' team first developed a method for rapid library preparation and sequencing of short DNA fragments on the MinIon. The key to the library prep method is that the researchers were able to optimize barcoding and the addition of sequencing adapters, performing both in one ligation reaction. That helped reduce the time to around 45 minutes from two to four hours. In addition, the library is barcoded, so it can be pooled with other samples and sequenced, Williams said.
In total, the entire workflow takes between four and six hours, which includes whole-genome amplification, library prep, and between 20 minutes and two hours of nanopore sequencing. Sequencing time will vary depending on the level of multiplexing.
In the BioRxiv paper, the researchers tested the protocol on nine blinded samples. Trophoectoderm biopsy samples were taken from fresh, day-five embryos and sent to a reference lab for clinical PGS testing using the VeriSeq PGS assay, with excess DNA used to run the MinIon-based workflow.
The nine samples included both diploid and aneuploid samples and one normal reference male sample. Five samples were multiplexed and assessed in one MinIon run. The researchers developed a modified Z-score method for aneuploidy detection — a bioinformatics strategy that assesses the relative proportion of each chromosome to determine whether it is diploid or aneuploid.
After sequencing, the researchers used only reads that were assigned to a barcode and uniquely matched to the reference genome for analysis. More than 70 percent of reads were assigned to a unique barcode, and of those, between 75 percent and 95 percent were uniquely assigned to the reference.
The researchers also determined that to detect whole chromosome aneuploidy, 30,000 reads were needed. In the study, their results were concurrent with the VeriSeq PGS test. Of the nine samples, five were abnormal, including a female monosomy 22, female trisomy 19, female trisomy 22 and monosomy X, female trisomy 13 and monosomy 14, a male with an extra X chromosome and trisomy 15, and a female with trisomy 15 and monosomy 18 that was also mosaic for trisomy 6.
The researchers discovered that although 30,000 reads were sufficient for detecting all of the whole chromosome aneuploidies, about 60,000 reads were required to detect mosaicism. In addition, they estimated that around 60,000 reads would be needed to detect sub-chromosomal deletions and duplications.
John Zhang, founder and CEO of New Hope Fertility, an IVF clinic in New York, said that the study was "promising," but "it needs a larger clinical trial following all the way through pregnancy and birth to determine efficacy." In addition, he said, the current VeriSeq PGS test is "fairly fast and accurate," so it would be important to do a "side-by-side comparison to honestly determine the benefits of nanopore sequencing."
Williams said the next step is to scale up the study and test larger number of samples. "There's always a challenge when you go from the lab to the clinic and determining the test parameters under real-world conditions," he said.
One key aspect the team will be evaluating is the accuracy and robustness of the MinIon for clinical testing. "One limitation of nanopore sequencing has been its accuracy," Williams acknowledged. However, since aneuploidy detection does not rely on calling point mutations, base-level accuracy is not as important. "We want to have an accuracy of over 99 percent, and to test enough samples to be sure that's what it is will take time," he said. Nonetheless, "I think the potential is there."
Aside from being able to do PGS testing faster and without requiring embryos to be frozen, Williams said, he anticipates that test costs could be lower. Sequencing and reagents costs would likely be less, he said, and the costs of freezing and thawing embryos, which can add thousands of dollars, would be avoided. In addition, he said, it would enable some laboratories to do testing in house versus having to send samples to a reference lab for testing. At some large, experienced laboratories like at Columbia, MinIon-based PGS testing would be feasible and desirable, but for some smaller labs, "having the ability to send samples to a large reference lab is helpful," he said.