The one gene-one trait idea was succeeded long ago by the polygenic model, the idea that many traits, such as height, are complex and controlled by many genes. Now, as bewildering findings accumulate from genome-wide association studies, the polygenic model is starting to show the strain.
For complex traits, scientists at Stanford University have noticed, association signals tend to be spread across most of the genome. For example, signals associated with the propensity for a particular disease may be found near many genes that lack any obvious connection to disease.
These Stanford scientists, led by Jonathan Pritchard, Ph.D., have even come around to the idea that the gene activity of cells is so broadly networked that virtually any gene can influence disease. The implication is that most of the heritability of diseases is due not to a handful of core genes, but to tiny contributions from vast numbers of peripheral genes that function outside disease pathways.
Any given trait, it seems, is not controlled by a small set of genes. Instead, nearly every gene in the genome influences everything about us. The effects may be tiny, but they add up.
The Stanford scientists, then, propose that it may be time to reassess a core assumption in the study of disease-causing genes—that they are clustered in molecular pathways directly connected to the disease. In short, the scientists suggest that the polygenic model may need to be expanded. It may need to become “omnigenic.”
The omnigenic model was introduced June 15 in Cell, in an article entitled, “An Expanded View of Complex Traits: From Polygenic to Omnigenic.”
“Intuitively, one might expect disease-causing variants to cluster into key pathways that drive disease etiology,” wrote the article’s authors. “We propose that gene regulatory networks are sufficiently interconnected such that all genes expressed in disease-relevant cells are liable to affect the functions of core disease-related genes and that most heritability can be explained by effects on genes outside core pathways.”
In any cell, there might be 50 to 100 core genes with direct effects on a given trait, as well as easily another 10,000 peripheral genes that are expressed in the same cell with indirect effects on that trait, noted Dr. Pritchard.
Each of the peripheral genes has a small effect on the trait. But because those thousands of genes outnumber the core genes by orders of magnitude, most of the genetic variation related to diseases and other traits comes from the thousands of peripheral genes. So, ironically, the genes that have the most indirect impact on disease end up being responsible for most of the inheritance patterns of the disease.
“This is a compelling paper that presents a plausible and fascinating model to explain a number of confusing observations from genome-wide studies of disease,” said Joe Pickrell, Ph.D., an investigator at the New York Genome Center, who was not involved in the work, in a press release accompanying the study.
Until recently, commented Dr. Pritchard, he thought of genetically complex traits as conforming to a polygenic model, in which each gene has a direct effect on a trait, whether that trait is something like height or a disease, such as autism.
But last year, while putting together a paper on the recent evolution of height in northern Europeans, Pritchard was forced to rethink that idea.
In the earlier work on the genetics of height, Dr. Pritchard and his colleagues were surprised to find that essentially the entire genome influenced height. “It was really unintuitive to me,” he said. “To be honest, I thought that it was probably wrong.”
“I gradually started to realize that the data don't really fit the polygenic model. We started to think, 'If the whole genome is involved in a complex trait like height, then how does that work?'”
The polygenic model leads researchers to focus on the short list of core genes that function in molecular pathways known to impact diseases. So, therapeutic research typically means addressing those core genes. A common approach to gene discovery is to do larger and larger genome-wide association studies, the paper notes, but Dr. Pritchard's team argues against this approach because the sample sizes are expensive and the thousands of peripheral genes uncovered are likely to have tiny, indirect effects.
“After you get the first 100 hits,” advised Dr. Pritchard, “you've probably found most of the core genes you're going to get through genome-wide association studies.”
Instead, he recommends switching to deep sequencing the core genes to hunt down rare variants that might have bigger effects. For clinical use, Pritchard said, there's still a rationale for genome-wide association studies: to predict the peripheral gene-based risk factors in individual patients in order to personalize medicine.
Dr. Pritchard's omnigenic model promises to take basic biology in new directions and means biologists need to think a lot more about the structure of networks that link together those thousands of peripheral disease genes.
“If this model is right,” said Dr. Pritchard, “it's telling us something profound about how cells work that we don't really understand very well. And so maybe that puts us a little bit further away from using genome-wide association studies for therapeutics. But in terms of understanding how genetics encodes disease risk, it's really important to understand.”