New Online Tool Could Bring Genome-based Risk Scores to More Patients

Scientists have reported a big step forward in bringing genome-based answers to more patients for a variety of clinical applications. To properly understand this achievement, we need to brush up on a few key points — and I promise to keep the technical jargon to a minimum.

The new report comes in the form of a preprint, which means it’s a research paper that has not yet run the gauntlet of scientific peer review. That peer review process helps to stress-test each paper, and at the end of it papers can look a little different, or not be published at all. So it’s important to remember that we’re looking at a preliminary report that still needs third-party validation.

This particular preprint focuses on polygenic scores (sometimes known as polygenic risk scores). Don’t run away! I’ll call them multi-gene scores, because that’s really what they are: a calculated score based on a number of genetic variants located across several different genes. When genetic data first made its debut in the clinic, everything was based on single genes, and often single variants within those genes. As their understanding of the human genome progressed, scientists found tons of individual variants associated with major health risks: a gene linked to susceptibility for Alzheimer’s disease, mutations associated with rare diseases such as Alport syndrome, a genetic region that causes fragile X syndrome.

When you’re looking at a genome that spans 3 billion DNA base pairs, it’s a lot easier to tease out the function of single genes or variants that have a major biological effect. But after 25 years of combing through our shared DNA to identify these single-gene disorders and diseases, scientists have realized that there is still a lot of genetic diversity left to explain. And it turns out that some of the remaining answers lie in combinations of genes or variants. Each individual variant might only have the tiniest effect on physiological function, but when you bundle up all of the relevant variants, together they produce a signal large enough to explain risk for a certain disease.

As you might expect, it’s the more complex diseases that often have these more complex genetic risk profiles. Risk for coronary artery disease, stroke, schizophrenia, and various cancers, among other conditions, has been more accurately explained through multi-gene scores than through any single gene or genetic variant. Because these diseases are far more common than the rare diseases usually associated with a single genetic change, the use of multi-gene scores for clinical purposes stands to bring the benefits of precision medicine to a lot more people.

But what we gain from multi-gene scores depends on their scientific validation. They were hotly debated among scientists when they first emerged; a growing consensus now accepts the value of these scores but there is still much to be learned about which genes and variants to include, and how heavily to weight each one in any risk scoring calculation. Honing those details has to happen for every condition or disease that is linked to a multi-gene score before it can be clinically useful.

And that’s where this new preprint comes into play. Scientists at Nationwide Children’s Hospital in Ohio, the Institute of Molecular Medicine Finland, and other institutions have developed the PGS Browser, a publicly available online platform that facilitates analysis and interpretation of polygenic scores. This is admittedly a pretty technical advance — the interactive tool will be used primarily by scientists and clinical researchers — but it represents a critical step in ushering these scores into clinical use. The initial data in the browser is based on a deep analysis of thousands of multi-gene scores across nearly half a million participants in the FinnGen large-scale research program.

Every type of genetic data that has made it into mainstream (or even mainstream-adjacent) medicine has been propelled there by a tool like this. A decade or so ago, the launch of ClinVar gave clinical laboratory teams a place to report and share interpretations of single genetic variants linked to disease. Not only could teams that excelled at variant interpretation have a greater impact with their work by sharing it publicly, but it also gave teams that weren’t as advanced a free public resource to help check their analyses and improve their interpretation skills. Around the same time, the Exome Aggregation Consortium (originally known as ExAC, and now renamed gnomAD) let scientists share troves of exome and genome sequencing data. By sharing information from large-scale sequencing studies, the database makes it easier to harmonize results both for shared projects and for independent projects that can be compared to the public data.

These and other tools allow the entire community to improve by establishing a kind of lookup table and making it possible for scientists to compare the results they generate to those produced by others. Time will tell if the PGS Browser has the staying power of ClinVar or gnomAD, but its launch is a good indicator that multi-gene scores are getting closer to widespread clinical use. If that’s the case, they could do a lot of good for a lot of people.