NEW YORK (GenomeWeb) – Clinical whole-exome sequencing has been used successfully to diagnose thousands of pediatric and adult patients with genetic diseases. A number of clinical testing labs, including at Baylor College of Medicine and GeneDx, have now started to explore exome sequencing in a prenatal setting, which presents some unique challenges because it requires quick turnaround times and because genetic defects may present differently in a fetus than once a child is born.
At the American College of Medical Genetics annual meeting in Tampa last week, Magdalena Walkiewicz from Baylor College of Medicine and Carin Yates from GeneDx presented initial findings of their respective labs from exome sequencing of fetal DNA samples.
Baylor Miraca Genetics Laboratory has been offering a trio exome sequencing test specifically designed for a prenatal setting since the spring of last year, after receiving several requests for exome sequencing on ongoing pregnancies.
For the prenatal application, the lab adapted its standard exome sequencing approach in several ways, optimizing library construction and exome capture, increasing sequencing coverage, and reducing the workflow from 20 days to less than three days, Walkiewicz said. It also validated the test, which has a turnaround time of two to three weeks, on DNA derived from amniocentesis and chorionic villus sampling.
Improvements to the test, currently in its third version, continue, in particular to increase coverage. At the moment, it provides full coverage for more than 3,600 disease genes listed in the Online Mendelian Inheritance in Man database.
After some debate, the Baylor team decided to report two categories of findings for the prenatal exome: disease genes related to the prenatal test indication, and disease genes unrelated to the prenatal presentation but likely to cause severe childhood disorders. The report includes both pathogenic and likely pathogenic variants, and, in certain cases, variants of unknown significance. VUS are included, for example, for recessive disorders where the lab identifies one pathogenic variant as well as one VUS in a trans configuration.
So far, Baylor has performed prenatal exome sequencing for 72 consecutive cases, providing reports within the promised turnaround time for 90 percent of them.
For 18 of the 52 consecutive cases for which Walkiewicz provided data in her presentation, or 35 percent, they were able to obtain a molecular diagnosis. Of these, nine were autosomal recessive conditions, eight were autosomal dominant diseases, and one was an X-linked disorder.
Walkiewicz said that so far, parents and their doctors have mostly used the test results for managing a child's disease post-delivery, but since the lab has started to return results from samples obtained at 14 to 16 weeks of gestation, they have also been used for making decisions about pregnancy termination.
A total of 30 fetal samples were obtained from amniotic fluid; 14 were tissue samples from products of conception, which are placental or fetal tissue remnants in the uterus after delivery, miscarriage, or pregnancy termination; and four each came from CVS and fetal cord blood, respectively.
Thirty-seven cases were proband-only exomes, of which they solved 12, or 32 percent, and 15 were trio exomes, of which they solved six, or 40 percent.
They also analyzed cases by fetal phenotype as observed on ultrasound and found that the diagnostic rate was similar for different types of anomalies. For three out of five cases that showed brain anomalies but no other phenotype, they provided a molecular diagnosis; of 21 cases with anomalies in the brain and other organ systems, they solved seven; and of 26 cases with anomalies other than the brain, they solved eight. For one of those, they also discovered a secondary finding, a pathogenic mutation for a severe childhood-onset disorder. Also, of seven cases with cardiac as well as brain defects, they solved three, and of six cases with cardiac anomalies and defects in other organ systems, they solved another three.
Family history made a difference to some extent. Of 32 cases where the proband was the first in the family with the presentation, they solved 13, and of 20 cases where there were other affected family members, they solved five.
Parents were not always required to obtain a diagnosis. In one case that involved in vitro fertilization with donor eggs and sperm, neither biological parent was available, and the lab identified a likely pathogenic variant in a gene associated with Kabuki syndrome in the fetus, a diagnosis that was confirmed post-delivery.
According to GeneDx's Yates, whole exome sequencing can be a "useful tool" for identifying the molecular cause of fetal anomalies. Of the more than 12,000 clinical exome tests GeneDx has run so far, 72 were products of conception, either from a fetal demise or pregnancy termination, she said, and the lab was able to provide firm molecular diagnoses for 18 percent of these. Diagnostic yield may increase over time as some of the possible diagnoses are turned into definite ones, she added.
GeneDx is not currently offering exome sequencing in a prenatal setting, she said, but such a test will be "coming soon."
Two-thirds of cases who provided the gestational age came from the second trimester, and a third from the first trimester.
Fifty-four cases had prior diagnostic tests performed — karyotyping, microarrays, or both — which did not yield a diagnosis. Of these, 20 percent received a definite diagnosis from exome testing, and 43 percent a possible result.
Of the exome tests performed, 40 were parent-fetus trios, 23 were proband-only exomes, five were quads that also included a sibling or a previous pregnancy, and four were mother-fetus duos. Analyzing family members increased the diagnostic yield — 24 percent of trios or quads obtained a definite result, whereas only a single case of proband-only exome received a molecular diagnosis.
For 13 of these cases, or 18 percent, the lab was able to obtain a definitive molecular diagnosis, meaning a pathogenic or likely pathogenic variant that explained the fetus' phenotype.
In another 42 percent of cases, the genetic finding partially explained the phenotype, or the researchers found one pathogenic variant and one VUS for a recessive condition. In 12 percent of cases, they found a novel candidate gene that does not have much human data available but for which animal or functional studies suggest it is disease-related. In about a fifth of cases, the lab did not identify any reportable variant.
On ultrasound, fetuses had a variety of phenotypes, including brain anomalies, fluid accumulation in several compartments, and cardiovascular defects. Yates cautioned that accurate phenotyping is difficult because of the limitations of imaging and fetal age. While it is often unknown how a genetic variant presents in a prenatal setting, more exome analyses of fetal samples will allow laboratories to learn about how gene defects correlate with prenatal presentation, she said.
Several genes popped up in more than one case, for example FGFR2, which had pathogenic variants in two cases, both of which had skeletal abnormalities. PTPN11 turned up in two cases, and RYR1 in three, but these cases had different clinical presentations, Yates said, and they had additional variants in other genes. This presents challenges, both for making molecular diagnoses and for counseling families about future pregnancies, she said.