Skip to main content
Premium Trial:

Request an Annual Quote

As Labs Start Returning Incidental Findings, Weighing Variant Pathogenicity Remains Challenge


NEW YORK (GenomeWeb) − As researchers attempt to put last year's guidelines by the American College of Medical Genetics and Genomics for reporting incidental findings from clinical exome and genome sequencing into practice, they are starting to learn about the frequency of such findings and the challenges associated with classifying variations as pathogenic.

In an article published in Genetics in Medicine this month, for example, researchers from the National Institutes of Health Undiagnosed Diseases Program described their experience with finding and reporting incidental – also called secondary – findings in 159 families participating in the project.

Following the ACMG guidelines, they analyzed variants in 56 genes in the exomes of 543 family members. In 27 individuals from 14 families, or about 5 percent of participants, they discovered a total of 14 reportable variants.

These numbers are a little higher but roughly in line with those of others groups, who reported rates of incidental findings ranging from .2 percent to 4.8 percent. Baylor College of Medicine's Whole Genome Laboratory, for example, discovered incidental findings in 3 percent of more than 2,000 patients undergoing exome sequencing when they only looked at the ACMG's gene list, and in 4.8 percent when they assessed additional genes, according to Sharon Plon, a professor of molecular and human genetics at Baylor.

Also, a study published last fall by researchers at the University of Washington that looked for variants in 114 genes in 1,000 individuals from the National Heart, Lung, and Blood Institute Exome Sequencing Project identified pathogenic or likely pathogenic variants in 23, or 2.3 percent, of participants.

Another study, published two years ago by researchers at the National Human Genome Research Institute, analyzed variants in 37 genes in the exomes of 572 participants in the ClinSeq study and found clinically important mutations in eight, or 1.4 percent, of them.

And researchers at the Radboud University Medical Center in the Netherlands, who do not routinely scan genes unrelated to a patient's disorder for secondary findings, said last year that they had discovered merely two incidental findings in about 1,000 exomes from patients and family members, a rate of .2 percent.

Technical differences between the labs, for example in the coverage and completeness of the exomes, as well as differences in the lists of genes they scanned for secondary findings, likely account for some of the variation in their frequency.

But another factor seems to be the lack of consensus for classifying variants as 'expected pathogenic', which the ACMG recommends should be reported along with 'known pathogenic' mutations.

Challenges with classifying variants are partly caused by our lack of knowledge, according to Murat Sincan, a research fellow for the Undiagnosed Diseases Program and one of the lead authors of its recent paper. "We simply don't know what each possible variant's downstream functional effects are, [and] neither can we reliably predict those effects with our current knowledge," he told Clinical Sequencing News.

And even when there is information about a variant available in databases and the literature, it is not always consistent. "One study might say it has a deleterious effect, whereas another study, done in another model organism or in humans, might argue otherwise," Sincan said. "In those cases, it's really a judgment call on the researcher's end to try to determine the expected pathogenicity of this mutation, and it's never an easy task − it takes reading almost all the literature around this gene" by an expert, which is time-consuming.

For their own study, for example, it took about 320 hours of manual labor to analyze variants, including 200 hours to intersect them with locus-specific databases and to flag potentially interesting ones, and 120 hours to review the literature and splice predictions for individual variants.

Plon agreed that analyzing variants for pathogenicity remains a challenge. "I do feel that the current lack of clear guidelines as to what mutations should be reported does make it very difficult to implement [the ACMG guidelines], and I think that difficulty should not be underestimated," she said.

For many genes, there is no list of known pathogenic mutations, and some types of variants tend to be pathogenic in some but not other genes. The RET oncogene, for example, is often activated by missense mutations, whereas the majority of pathogenic variants in the BRCA1 and 2 genes are truncating mutations.

"The laboratory has to do a significant amount of work to try to curate whether they think there is enough evidence for pathogenicity to call something known pathogenic, and to report it as an incidental finding," Plon said. "That is really not a minor test."

The lack of clear criteria for classifying variants as pathogenic likely leads to differences in incidental findings rates between laboratories, according to Gail Jarvik, head of medical genetics at the University of Washington and the senior author of last year's study. Her group, she said, believes in more stringent criteria for incidental findings than the UDP's researchers and weighs the evidence from functional assays, allele frequency, and family segregation differently than that group.

The UDP's study "underscores the need for consensus regarding how much evidence and of what type one needs when classifying a variant as pathogenic," she told CSN, particularly in the context of incidental findings. "Given the low prior probability for a pathogenic variant in a person without the disorder, and often with no family history, we do not support returning variants as pathogenic unless they are highly likely to be," she said.

Clearer criteria might emerge from a revised version of the ACMG recommendations that is currently being drafted. That version "is trying to be much more specific about what type of evidence you need to look at, and makes some attempt to measure the importance of different types of evidence," Plon said, although "it certainly won't provide you with a list" of pathogenic mutations in the 56 genes.

ClinVar and other databases are starting to gather such lists, but in the meantime, labs will still collect evidence manually. "As we continue to build these databases, if you lack definite information either implicating or acquitting mutations … you still need to go to the literature and try to gather more information about the pathogenicity of the mutations," Sincan said.

Lists of pathogenic variants already exist for some genes, for example, certain mismatch repair genes, Plon said, that were gathered and published by an international consortium. Plon is part of the Clinical Genome Resource, or ClinGen, consortium, which was funded with more than $25 million from the NIH last year to develop a framework for evaluating variants for pathogenicity and submitting the information to the ClinVar database.

Another group of scientists recently published their own set of guidelines for classifying sequence variants that were discovered as part of research projects as pathogenic, which they plan to follow up with more specific recommendations in the future.

Besides finding better ways to determine which incidental findings to return, researchers are also starting to look at the clinical implications of the results. As part of an NIH-funded clinical exome sequencing trial, for example, Baylor researchers are following patients for two years, monitoring whether they follow up on incidental findings with their physicians and whether additional family members seek testing.

"I think it is very likely to be a major aim of research moving forward to get some type of longitudinal data about how often a mutation that was reported as an incidental finding is associated with any evidence of disease over long-term follow-up," Plon said, and how useful and cost-effective these findings are. She added that the NHGRI's Clinical Sequencing Exploratory Research consortium is especially interested in these questions.