Beneath the familiar language of genes and inheritance lies a deeper, less charted territory — the regulatory machinery that decides whether a gene speaks or stays silent. Researchers at the University of Utah have built a tool called FIRE that uses machine learning to map these molecular switches across the human genome, and in doing so have uncovered something quietly profound: roughly 40 percent of the time two genetically identical DNA copies are regulated differently, no genetic explanation exists. The discovery suggests that biology harbors a dimension of variability that operates outsid
Machine Learning Tool FIRE Maps Gene Regulation Beyond Genetics
Genetics isn't destiny—40% of gene regulation differences have no genetic basis
Mark
So this tool FIRE—it's looking at gene regulation, not the genes themselves. Why does that distinction matter?
Mimi
Because having a gene and having it turned on are two completely different things. You can inherit the exact same gene as your parent, but if it's regulated differently, it behaves differently. That's where disease often hides.
Luke
But how confident are we that FIRE is actually identifying the important regulatory regions? The tool uses machine learning to predict which highlighted regions matter, but it's still making predictions, not measuring function directly.
Mimi
That's fair. They validated it with the celiac disease example—they found a regulatory region, predicted it mattered, and then changed one DNA letter in the lab and confirmed the switch actually opened differently. So there's at least some experimental backing.
Mark
And the 40 percent finding—that regulatory differences exist without genetic differences—that seems like it could be huge for understanding disease.
Mimi
It is. It means genetics isn't destiny. Two people can have identical DNA at a spot but regulate it completely differently. And those differences get passed down as cells divide.
Luke
But we don't know what most of these non-genetic regulatory differences actually do yet, right? The source says their function is unknown.
Mimi
Correct. This is mapping the territory. The next step is figuring out which of these differences matter for health and which are just noise.
Mark
So FIRE is really a tool for finding the questions, not answering them.
Mimi
Exactly. It's narrowing the search space from thousands of possible variants to the one or two that likely matter. That's the hard part.
Luke
And the clinical application—editing DNA switches to treat disease—that's still theoretical at this point.
Mimi
Yes. But Vollger is clear about what he's hoping for: a pathway from identifying disease-causing variants to developing therapies. This is the first step.
The Pulse
- For generations, medicine has treated DNA sequence as destiny — FIRE now reveals that nearly half of all regulatory differences between inherited gene copies arise from forces genetics alone cannot explain.
- Tens of thousands of previously invisible genomic regions, long dismissed as biological noise, are emerging as likely disease-relevant switches, demanding urgent reexamination of what we call the 'dark genome.'
- The tool cut through thousands of candidate variants linked to celiac disease to identify a single DNA letter that disrupts a gene-controlling protein binding site — the kind of precision that transforms association into actionable target.
- Stable, heritable regulatory differences that carry no genetic fingerprint persist across cell divisions, suggesting a new class of biological inheritance that current frameworks were not built to see.
- Researchers are now moving to translate FIRE's maps into therapeutic coordinates — identifying which regulatory switches, when edited, might lower disease risk or offer cures for rare hereditary conditions.
Beneath the familiar language of genes and inheritance lies a deeper, less charted territory — the regulatory machinery that decides whether a gene speaks or stays silent. Researchers at the University of Utah have built a tool called FIRE that uses machine learning to map these molecular switches across the human genome, and in doing so have uncovered something quietly profound: roughly 40 percent of the time two genetically identical DNA copies are regulated differently, no genetic explanation exists. The discovery suggests that biology harbors a dimension of variability that operates outside the rules we thought governed heredity, and that the path from genetic knowledge to medical action may run through this unmapped terrain.
For decades, scientists understood that inherited genes shape our vulnerability to disease — but genes, it turns out, are only half the story. What often determines whether a disease takes hold is not the gene itself but whether it is switched on or off. That regulatory machinery, researchers are now learning, is far stranger and more consequential than anyone had imagined.
Mitchell R. Vollger and colleagues at the University of Utah have developed a tool called FIRE, published in Cell, that pairs machine learning with advanced DNA sequencing to map the genome's regulatory switches — the regions that determine whether a gene is active or silent. The human genome is vast, with only 1 to 2 percent coding for proteins; the rest was long called junk, yet it houses the control architecture for all of gene expression. FIRE adds a machine learning layer that evaluates billions of candidate regulatory regions and predicts which ones are genuinely functional. 'Machine learning here is really copying something we can see with our own eyes in the data,' Vollger explains, 'but at a scale we could never manage by eye.'
The tool's practical power was demonstrated through celiac disease research, where FIRE identified a single DNA letter — among thousands of nearby variants — that disrupts a protein binding site in a gene switch adjacent to an immune gene. Changing that one letter in lab experiments altered how open the switch was, transforming a statistical association into a mechanistic explanation.
The most unexpected finding, however, was structural. When FIRE examined regions regulated differently between maternally and paternally inherited chromosomes, genetics explained most cases — but not all. In roughly 40 percent of instances, the two copies were genetically identical yet still regulated differently, and these differences persisted as cells divided. 'The most surprising result was finding the limits of how much our genetics can determine our gene regulation,' Vollger says. The discovery points to a stable, heritable layer of biological variability that operates entirely outside DNA sequence.
FIRE also flagged tens of thousands of previously overlooked genomic regions enriched for disease-associated variants, opening new investigative territory. Vollger's broader hope is that by connecting genetic changes to functional regulatory differences, FIRE will help identify which variants actually cause disease — and ultimately guide DNA editing therapies toward the switches that, when corrected, might treat or cure hereditary conditions. It is a preliminary step, but one that redraws the map between what we inherit and what medicine can reach.
For decades, scientists have known that the genes we inherit matter enormously—they shape our risk for heart disease, cancer, diabetes, and countless other conditions. But genes are only half the story. What truly determines whether a disease takes hold is often not the genes themselves, but how they are switched on and off. That regulatory machinery, it turns out, is far more complex and mysterious than anyone realized.
Mitchell R. Vollger, an assistant professor of human genetics at University of Utah Health, and his colleagues have developed a tool called FIRE that begins to illuminate this hidden layer of biology. The technology pairs machine learning with advanced DNA sequencing to map which regions of our genome control gene regulation—the switches that determine whether a gene is active or silent. The work, published in Cell, reveals something unexpected: roughly 40 percent of the time genes are regulated differently between the copy inherited from your mother and the copy from your father, there is no genetic difference between them at all. This finding upends a long-held assumption that genetics determines everything about how our genes behave.
The human genome is vast and mostly dark. Only about 1 or 2 percent of it codes for proteins. The remaining 98 percent was long dismissed as junk, but it contains the regulatory machinery—the switches and dimmers—that control when and how genes turn on. Identifying which of these regions actually matter has been nearly impossible. Researchers could highlight physically accessible regions of DNA, but they couldn't easily tell which ones were genuinely important for gene regulation and which were simply open by chance. FIRE solves this by adding a machine learning layer that assesses billions of highlighted regions across the genome and predicts which ones are likely to be functionally important. "Machine learning here is really copying something that we're able to see with our own eyes in the data," Vollger explains. "But it's doing it at a scale that we could never manage by eye."
The practical payoff is immediate and concrete. The researchers used FIRE to investigate a genetic difference that had been associated with celiac disease risk, but no one knew whether it actually affected how genes worked or was simply a bystander. FIRE revealed that the variant sits in a gene switch next to an immune gene also linked to celiac disease. The DNA change disrupts a spot where gene-controlling proteins normally attach. When the team changed just that single DNA letter in lab experiments, it altered how open the switch was. Out of thousands of nearby genetic variants, this appears to be the one that matters—a crucial distinction for anyone trying to develop a medicine.
But the most striking discovery was the one that defied expectation. When the researchers searched for regions regulated differently between maternally and paternally inherited chromosomes, they found that in most cases, genetic differences could explain the regulatory differences. Yet in a large minority of cases—about 40 percent—the two copies were genetically identical at that spot, yet still regulated differently. These differences persist as cells divide, suggesting they are stable, heritable changes in gene regulation that have nothing to do with DNA sequence. "I think the most surprising result of this work was finding the limits of how much our genetics can determine our gene regulation," Vollger says. The discovery points to a new dimension of biological variability, one that operates outside the traditional rules of genetics.
The team also identified tens of thousands of previously overlooked DNA regions likely to be important for gene regulation. Notably, these regions are more likely than other parts of the genome to contain DNA changes associated with human disease, suggesting they have critical, undiscovered functions for health. This opens a vast new territory for investigation. Vollger's hope is that by linking changes in the genetic code to functional differences in gene regulation, FIRE will help researchers identify which genetic variants actually cause disease—a crucial first step toward therapies that edit DNA to treat or cure hereditary conditions. "This might show us how we could edit the DNA switches that control a person's genes to help lower their risk for a common disease, or even cure them of a rare disease," he says. "It's a small step toward turning the genetic variants we find into something doctors can act on." The work is preliminary, but it suggests a path forward: from understanding how genes are regulated, to identifying which regulatory changes cause disease, to editing those switches to restore health.
Notable Quotes
Only about 1 or 2% of our genome codes for proteins. This gives us a much sharper lens on the other 98%, which includes the switches that turn our genes on and off.— Mitchell R. Vollger, assistant professor of human genetics at University of Utah Health
About 40% of the time that we see a regulatory difference between the copies we inherited from mom and dad, there's no genetic difference between them at that spot.— Mitchell R. Vollger
Want the full story?
Read the original at Newswise ↗
Health & Medicine