Test Quality: New Study Casts Doubt on Whole Exome Sequencing’s Accuracy in Diagnosing Genetic Disorders
Depending on the reporting laboratory, patients who undergo whole exome sequencing may receive false negatives or incomplete test results. That is the finding of a new study published in Clinical Chemistry
Depending on the reporting laboratory, patients who undergo whole exome sequencing may receive false negatives or incomplete test results. That is the finding of a new study published in Clinical Chemistry on Jan. 6, 2020. Conducted by researchers from the University of Texas Southwestern Medical Center, the study reviewed clinical tests from three major U.S. laboratories and concluded that whole exome sequencing routinely fails to adequately analyze large segments of DNA, a potentially critical deficiency that could prevent doctors from accurately diagnosing potential genetic disorders ranging from epilepsy to cancer.
Challenges Posed by Whole Exome Sequencing
Whole exome sequencing is a technique for analyzing protein-producing genes. It is used to identify genetic mutations that cause disease in children and adults with rare or undiagnosed diseases. Of course, when parents and children undergo whole exome sequencing, they fully expect the test results to be accurate. So do the physicians who order the tests.
But exome sequencing may not be living up to those expectations. The process of fully analyzing the approximately 18,000 genes in an exome is inherently difficult and susceptible to oversights. This could explain why approximately half the tests do not pinpoint a mutation.
The researchers said they conducted the study because to test their hypothesis that vast differences in testing quality are endemic in clinical genetic sequencing but have not been well documented or shared with clinicians. The research team double checked findings from 36 patients’ clinical exome tests performed from 2012 to 2016 by three major, unidentified U.S. clinical laboratories, noting coverage of genes and nucleotide positions. The study sample included 20 proband patients and 16 parents seen at a pediatric genetic testing clinic.
Although the sample size was admittedly small, the researchers said that the laboratories are representative of clinical genomics laboratories in the U.S. “I consider the three labs in the study to be of generally high quality, and we routinely send patient samples to these labs,” researcher Dr. Jason Park, PhD, an associate professor of pathology at UT Southwestern, said in an email to LabPulse.com.
Insufficient Gene Coverage
The study results raise concerns about the accuracy of whole exome sequencing. First, the reanalysis showed that, on average, each laboratory adequately examined less than three-quarters of the genes—34%, 66% and 69%. It also revealed startlingly wide gaps in the laboratories’ ability to detect specific disorders.
The researchers also found starkly contrasting results and inconsistency in terms of which genes were completely analyzed. A gene was not considered completely analyzed unless the laboratory met an industry-accepted threshold for adequate analysis of all DNA that encodes protein, which is defined as sequencing that segment at least 20 times per test. Notably, less than 1.5 percent of the genes were completely analyzed in all 36 samples. A review of one laboratory’s tests showed that 28 percent of the genes were never adequately examined and only 5 percent were always covered. Another laboratory consistently covered only 27 percent of the genes.
One possible explanation for this wide variance in coverage was that each of the three laboratories used different methods and products for whole exome sequencing. Technology and/or reagent issues can result in variation between laboratories, including the method used to turn a patient’s DNA into a library of fragments and the probes used to capture the DNA of interest, Park explained. But, he added, variation in quality of results within a laboratory is related to laboratory practices, and individual laboratories can and should address the variation between samples being tested.
|Whole Exome Sequencing: Comparison of Test Results from 3 Major U.S. Labs|
|Laboratory A||VCRome v2.0 (Roche Sequencing Solutions) or xGen Exome Research Panel v1.0 (Integrated DNA Technologies)||69% (12,184 of 17,723 genes)|
|Laboratory B||VCRome v2.1 (Roche)||66% (11,687 of 17,723 genes)|
|Laboratory C||SureSelect XT2 All Exon v4 (Agilent) or Clinical Research Exome (Agilent)||34% (5,989 of 17,723 genes)|
Recommendations for Improvement
The authors called on laboratories to continue working to improve the consistency of exome-testing performance. Recommendations include:
Statistical measures: The use of average measurements in large datasets such as exome/genome is only meaningful when the average is further described in combination with other summary statistics such as standard deviation and coefficient of variation,” according to the authors. “Additional statistics such as genes completely covered will also allow a more accurate assessment of the test’s validity and would pressure clinical laboratories to achieve higher consistency.”
Auditing and Inspection: The researchers also recommended increased auditing of coverage quality and overall performance by accreditation and regulatory inspectors. One example would be to incorporate evaluation of coverage consistency between laboratories into the College of American Pathologists (CAP) proficiency testing program for next-generation sequencing.
New evidence suggesting that whole exome sequencing routinely fails to adequately analyze large segments of DNA raises new doubt about whether physicians can rely on the method to accurately diagnosing potential genetic disorders. The problem is exacerbated by wide variation in accuracy from laboratory to laboratory.
Unless and until coverage and consistency improves, ordering physicians need to be aware of the deficiencies in whole exome sequencing. The researchers recommended that physicians ask laboratories about which genes are covered in the tests and consider the alternative of targeted panel tests for some patients, e.g., for a child with epilepsy without other complicating clinical problems. For such patients, ordering a smaller genetic test that completely analyzes a panel of genes associated with that disease may not only be less expensive, but just as likely to help physicians find answers.
This content is exclusive to Diagnostic Testing and Emerging Technologies subscribers
Start a Free Trial for immediate access to this article and our entire archive of over 20 years of DTET reports.