Options available for deciphering variants of unknown significance in genetic testing

admin
10 Min Read

Richard Gawel , 2025-05-13 12:13:00

Key takeaways:

  • Search engines and chatbots are easy to use.
  • Databases such as gnomAD and ClinVar are comprehensive but come with caveats.
  • About a thousand immunologists follow the VUSserve listserv.

Resources are available to help physicians when they encounter variants of unknown significance during genetic testing, Troy R. Torgerson, MD, PhD, said during Updates in Primary Immunodeficiency 2025.

“This is a challenge, and it remains a challenge,” Torgerson, director of experimental immunology, Allen Institute for Immunology, and affiliate clinical professor of pediatrics, University of Washington, said during his presentation.

Troy R. Torgerson, MD, PhD

Google and chatbots

If testing yields a variant of unknown significance (VUS), Torgerson said, the first step is to search for it online.

“It’s easy, and I’m sitting in front of my computer or my phone most of the time,” he said. “I keep calm, and I Google it.”

Torgerson simply enters the name of the gene and the protein variant, such as STAT3 R397W, into the search bar.

“It’s not all the time, but sometimes you hit gold with a simple Google search,” he said.

“The other thing is to ask your favorite chatbot,” he added.

Torgerson suggested OpenEvidence and Perplexity, which he tested by asking about a variant in FOXP3 that is known to be pathogenic.

“It nailed it,” he said. “It gives you all the references that it uses.”

Also, he said, Perplexity does not hallucinate like other artificial intelligence platforms.

“It’s great,” he said. “Use your chatbots to look for these things.”

Online databases

Online databases such as the Genome Aggregation Database (gnomAD), Infevers and ClinVar also are available.

“Let gnomAD become your friend,” Torgerson said.

Version 4 of the gnomAD data set includes 730,947 exomes and 76,215 genomes of diverse ancestries, including all the genomes from the UK Biobank, which were added last year.

“You can go to gnomAD and you can look for every variant,” Torgerson said.

Although gnomAD removes pediatric cases from disease cohorts, Torgerson cautioned, it includes adult cases from disease cohorts such as inflammatory bowel disease.

“Just know that it is contaminated a little bit with variants from these patients that actually have disease,” he said.

Torgerson also noted that most of the exomes and genomes in gnomAD hail from Caucasian populations with northern European ancestry, but coverage of other populations has improved.

“If you’re seeing African American patients, or Hispanic patients, you really need to drill down,” he said.

The database describes the frequency of variants by population, Torgerson continued, adding that variants may be common in, say, the Hispanic population but still be rare in the overall population.

“That’s one of the challenges of gnomAD,” he said.

Other databases collect pathogenic variants for specific diseases, Torgerson further said. For example, Infevers focuses on hereditary autoinflammatory diseases.

“This is curated by the world’s experts in autoinflammatory diseases,” he said. “This is like getting an immediate consult with the top 10 autoinflammatory disease experts in the world.”

Curation and algorithms

ClinVar is another popular database, Torgerson said. It assesses the certainty of whether variants are pathogenic, he continued, but it does not provide the evidence behind these assessments.

“All of the genetic testing companies who are finding all of these VUSs are doing what they should be doing, which is getting that information into a database to let people know that these variants are out there,” he said. “But those variants haven’t been tested yet, and so there’s no evidence.”

ClinGen has assembled more than 2,000 experts from around the world and across multiple specialties to work in panels to curate genes and variants for various diseases in ClinVar, Torgerson said, using the International Union of Immunological Societies gene list as a roadmap.

These panels will review the literature to discover which genes classify associations between genes and immune disorders as definitive, strong, moderate, limited, disputed or refuted.

“It’s manual, it’s very labor intensive, and it’s a labor of love for everybody involved,” Torgerson said.

The Gene Curation Expert Panels are now curating gene associations with severe combined immunodeficiency (SCID) and common variable immunodeficiency.

“We felt like those were the most urgent to curate,” Torgerson said. “Providers need to make decisions on those quickly.”

The panels have completed their curation of primary immunodeficiency diseases and are now focused on primary immune regulatory disorders.

“We’re working our way through them,” Torgerson said.

The Variation Curation Expert Panels are curating specific gene variants with disease, ranking them as pathogenic, likely pathogenic, uncertain, likely benign, and benign, so clinicians can make decisions about care in a timely manner.

“Their goal is, within the next few years, to essentially clean up ClinVar for the top seven genes that cause SCID both in this country and globally,” Torgerson said.

Torgerson also noted that 99 variant effect prediction algorithms are now available that can help physicians figure out whether a VUS is pathogenic by predicting whether it has some kind of effect.

For example, the Atlas of Variant Effects Alliance has created a publicly available Google sheet where physicians can look up specific variants, and the tool will describe the effects that they have.

Additionally, AlphaMissense models the protein structures of every missense variant that is possible in the human proteome and scores them on a range of pathogenic to benign. But Torgeson cautioned that it only predicts whether a mutation will impact predicted protein folding.

“As a result, it’s not very good at predicting gain of function mutations, because in gain of function mutations, usually, the protein folding is fine,” he said. “This is really a great resource for loss of function mutations, not so much for gain of function mutations.”

Phone a friend

When physicians encounter a VUS, Torgerson continued, they can pick up the phone or email another expert.

“The clinical immunology and pediatric rheumatology communities are small. They’re closely knit,” he said. “We’re really good at a really narrow thing.”

So, when someone asks them a question, Torgerson said, these specialists are eager to help.

“Reach out,” he said. “I’ve not met one malicious person that would not be happy to help out with a variant in the field.”

Torgerson also praised listservs such as the Clinical Immunology Society’s VUSserve, which has international reach and about a thousand immunologists following it, ready to read about submitted cases.

Finally, Torgerson said physicians can turn to functional testing. These tests are a great option, he said, but many of them only are available as research tests. Still, he added, they should be more clinically accessible due to the pause in FDA regulation of lab developed tests.

“Again, phone a friend. Find out who’s doing the testing,” he suggested. “Send them your mutation. Let them test it. And if they call you back and say, ‘Hey, that thing is very strong gain of function,’ you can decide whether or not you want to treat the patient accordingly.”

For more information:

Troy R. Torgerson, MD, PhD, can be reached at allergy@healio.com.

Source link

Share This Article
error: Content is protected !!