The headline from the largest genetic study of human energy metabolism ever run was a single rare gene, FNIP1, whose carriers appear substantially protected against heart and metabolic disease. Underneath that finding sits a set of results that got far less attention, including a measurement choice that reframes what the study is actually about and one gene whose effect was noticeably stronger in women.
The analysis, published in Nature on August 5, 2026, sequenced the protein-coding DNA of 1,032,116 people across 11 cohorts in North America, Europe, and Asia. It was led by the Regeneron Genetics Center.
The Measurement Was Not Weight, It Was a Ratio
The team did not screen for obesity. They screened for the triglyceride-to-HDL cholesterol ratio, two values that come from any standard lipid panel.
Before hunting for genes, they validated that ratio against a battery of physical measurements. A higher ratio was associated with greater total and visceral fat across MRI, DEXA, and bioimpedance; greater fat deposition in the liver and skeletal muscle; higher liver enzymes and biopsy-confirmed liver steatosis, inflammation, and fibrosis; and higher fasting insulin, HbA1c, blood pressure, and C-reactive protein. A higher baseline ratio predicted later type 2 diabetes, heart attack, fatty liver disease, and cirrhosis, consistently across African, admixed American, East Asian, South Asian, and European ancestries.
Crucially, the fat measurement was not a waist circumference. Using whole-body MRI, the team computed a visceral-to-gluteofemoral fat volume ratio, essentially the ratio of fat packed around the organs to fat stored in the hips and thighs. Carriers of FNIP1 loss-of-function variants had a lower ratio, lower body fat percentage, higher lean percentage, less liver fat, and lower liver enzyme levels.
That is the reframe. The variant is not simply associated with being thinner. It is associated with a different storage pattern.
One Gene Behaved Differently in Women
The exome scan turned up 59 independent genes associated with the ratio, heavily enriched for genes expressed in liver and fat tissue. Forty-four of them had never been linked to this biomarker before, according to the study record in PubMed, and the associations were replicated in a separate group of 114,942 participants from the All of Us Research Program.
When the researchers split the analysis by sex, the results were remarkably consistent for almost all of them. One gene was not. Rare variants in PDE3B showed a significant sex interaction, with an effect of 0.40 standard deviation units per allele in women versus 0.28 in men.
PDE3B is not a random hit. It has already been implicated in fat distribution in earlier work, which makes the sex difference more interesting than a lone statistical outlier would be.
It also deserves proportion. This is one interaction signal among 59 genes in a single study, and it needs independent replication before anyone builds a theory on it. What it does suggest is that the genetics of where fat goes may be more sex-dependent than the genetics of how much energy the body burns.
The Same Study Found Reasons Not to Switch This Pathway Off Everywhere
FNIP1 works with a partner protein, folliculin, encoded by FLCN, to restrain mitochondrial activity and energy expenditure. Break one copy of FNIP1, and the brake loosens. The obvious drug idea is to loosen it deliberately.
The same dataset shows why that has to be done carefully. Carriers of one broken copy of FLCN had an 18-fold higher risk of collapsed lung and a 9-fold higher risk of kidney cancer, consistent with Birt-Hogg-Dubé syndrome, an inherited disorder the gene is known to cause. People who inherit two copies of the FNIP1 gene develop immunodeficiency-93, a recessive condition characterized by absent B cells and thickened heart muscle. A 2025 mouse study found liver injury and bile duct cancer after knocking out Flcn in liver cells, though the authors' own long-term mouse experiments showed a protective liver picture instead.
Carriers of a single FNIP1 variant showed no significant association with the immunodeficiency features, consistent with that condition's recessive inheritance. But the pattern is why the authors propose silencing FNIP1 only in liver cells, using the same delivery approach behind several approved RNA therapies, rather than shutting the pathway down body-wide.
What Anyone Reading This Should Actually Do
Nothing is the honest answer.
The protective association rests on 155 carriers. The odds ratio for combined cardiometabolic disease was 0.39, with a 95% confidence interval of 0.22 to 0.69, meaning the true reduction could be anywhere from about 31% to 78%. Coronary artery disease and cirrhosis, examined individually, did not reach statistical significance.
These are genetic associations supported by cell and animal experiments, not clinical evidence. No drug targeting this pathway exists; none is in human trials, and body weight is shaped by many genes plus environment and behavior. Regeneron-affiliated authors receive salaries from and hold stock or options in the company that led the work and would develop any resulting therapy.
There is no test to ask for and no advice to follow. What the study offers, as an accompanying Nature summary frames it, is a map of 59 genes, 23 of which already encode approved or clinical-stage drug targets, and a hint that the fat-distribution portion of that map may not read the same way in men and women.
Key Questions Answered
What did the study measure? The ratio of triglycerides to HDL cholesterol in 1,032,116 people, validated against MRI, DEXA, liver biopsy, and later disease outcomes, is a marker of the body's energy state.
Why does fat location matter more than fat quantity? Fat packed around the organs is associated with worse metabolic outcomes than fat stored in the hips and thighs. The study measured that directly as a visceral-to-gluteofemoral volume ratio on MRI.
What was the sex difference? Rare variants in PDE3B, a gene already tied to fat distribution, were associated with a larger favorable shift in women than in men. It was the only clear sex interaction among 59 genes.
Does carrying an FNIP1 variant prevent obesity or diabetes? No. This is an association in 155 carriers with a wide confidence interval. It does not establish that the variant prevents any condition in an individual.
Could this become a drug? Possibly, but nothing is in human trials. Researchers propose silencing the gene only in liver cells because disabling the broader pathway is associated with serious conditions, including kidney cancer and immunodeficiency.
Who funded and ran the work? Regeneron Pharmaceuticals through its Regeneron Genetics Center, with academic collaborators. Regeneron-affiliated authors receive salaries from and hold stock or options in the company.
