The more data, the better, right? When it comes to genetics, it turns out that might not be the case.
As both genetic sequencing has gotten cheaper and computerized data analysis has gotten better, more and more researchers have turned to what are known as genome-wide association studies in hopes of sussing out which individual genes are associated with particular disorders. The logic here is simple: If you have a whole lot of people with a disease, you should be able to tell what genetic traits those people have in common that might be responsible. This thinking has resulted in an entire catalogue of hundreds of research studies that has shed light on the genetic origins of diseases such as type 2 diabetes, Parkinson’s disease, Crohn’s disease, and prostate cancer, while helping fuel the rise of personalized medicine.
Now, though, a new analysis calls the entire approach into question.
Writing in the journal Cell, a group of Stanford University geneticists write that such large studies are likely to produce genetic variants with little bearing on the disease in question—essentially false positives that confuse the results.
“Intuitively, one might expect disease-causing variants to cluster into key pathways that drive disease etiology [the causes of disease],” they write. “But for complex traits, association signals tend to be spread across most of the genome—including near many genes without an obvious connection to disease.”
Their analysis suggests an intriguing new way of viewing the genome in which nearly every gene impacts every other gene. Instead of a system in which you can plug and play different variables to affect different results, it’s a complex, inter-related network. They call this the “omnigenic model.”
Their work has broad, sweeping implications for the entire field of genetics. First off, that all those big, expensive genome-wide association studies may wind up being little more than a waste of time because they turn up genetic variants that, while perhaps interconnected to the disease, may not actually point to a viable target for things like drug therapy.
Indeed, genes that often seem related to diseases have stumped researchers in terms of the role they actually play in the condition. In the paper, for example, the Stanford researchers re-analysed a 2014 study of 250,000 people which found nearly 700 DNA variants linked to height—but only 16 percent of these variants had anything to do with a person’s height. In the paper, the Stanford researchers suggest that the impact of each variant has a teeny impact on height.
Far from solving a problem though, this new research merely opens up an entirely new line of questioning—and shows us once again, that we may not know nearly as much as we thought we did.