Some DNA reading methods skip whole pages
Re-testing of genetic tests from the three largest laboratories in the United States has shown that full-exom sequencing (reading only protein-coding genes) can miss large DNA fragments. Because of this, clinicians can leave up to half of patients without a correct diagnosis, whether it is cancer or epilepsy, the researchers said on the pages of Clinical Chemistry.
Today, medicine is turning to genetic tests more and more often: for example, they help to rule out a diagnosis in children, learn about the risk of developing diseases, or how to change their lifestyle so that they suffer less from their symptoms. Scientists from the University of Texas Southwestern Medical Center rechecked genetic tests in 36 patients from national American laboratories for 2012-2016 and found that they analyzed only 36%, 66% and 69% of genes.
Most often, such gaps do not appear due to deception of customers, but because of the imperfection of sequencing methods. Today we cannot unravel a tightly packed DNA molecule and read it all at once. In order to avoid mistakes and not to confuse the sequence of fragments during assembly, the same sequence must be read at least several dozen times (minimum for commercial companies – 20) and lay them on top of each other. Genetics call this the “coating,” and if it is not enough, then the sequence section is not read clearly enough. Only here physicians usually have no idea about these restrictions, and companies can use it.
“Many patients are young children with neurological impairment and need the most comprehensive diagnostic test,” explains Jason Park, Ph.D. and associate professor of pathology, Southwestern Medical Center, University of Texas. “The whole system falls apart when you start thinking about these tests to rule out a diagnosis.” The negative outcome of exomic sequencing makes no difference when so many genes are left without a thorough analysis. ”
So, in the case of epilepsy, one of the laboratories in some tests missed up to 80% of the genes associated with the disease, the second analyzed 40%, and only one analyzed at least three quarters properly. Only 1.5% of all protein-coding genes were analyzed in all 36 samples. In the rest, large fragments were missed, and in all cases they were different. In one laboratory, 27% of the genes in all tests were well coated, in the other 28%, and in the third only 5%.
“When we saw this data, we made a mandatory question for laboratories about the coverage of specific genes,” commented co-author Garrett Gotvey, Ph.D. and clinical geneticist at the University of Texas. “I don’t think it’s worth hoping for full coverage of all 18,000 genes each time, but it is fair to expect at least 90%.” According to her, to clarify the diagnosis, it is better to rely not on a genome-wide or full-exomic, but on a more narrowly focused sequencing that checks only those genes in which mutations are likely to be important in a particular case. Otherwise, there remains the risk of missing the most important thing, trying to cover everything at once.