Protective genetic associations and reactivated viruses
Two Nature studies examine FNIP1 variants associated with lower cardiometabolic risk and dormant-virus reactivation during severe COVID-19, while keeping their populations, mechanisms and causal limits distinct.
Science··Morning
An ultra-rare variant association across one million exomes
One of the peer-reviewed Nature studies used exome sequencing from 1,032,116 people across America, Europe and Asia to estimate associations between rare protein-coding variants and the ratio of triglyceride to HDL cholesterol. The scan identified 59 independent genes tied to that measure; 23 encode a target of an approved or clinical-stage drug. Ultra-rare protein-truncating variants in FNIP1 had a reported carrier frequency of 0.01 per cent. Carriers had a lower triglyceride-to-HDL ratio, less liver fat, lower glycaemia and a more favourable fat distribution, alongside around 60 per cent lower odds of cardiometabolic disease. Those human data establish an association. The mechanistic part used separate experimental systems: knocking FNIP1 down in human liver cells increased lipid breakdown and lysosomal gene expression. In mice on a high-fat diet, hepatic inhibition of Fnip1 with its paralogue Fnip2, or of their interactor Flcn, limited weight gain, reduced liver fat and improved insulin sensitivity. The cell and mouse results investigate the biological pathway around FNIP1; they do not show that the lower risk observed in human carriers arises through the same mechanism or that targeting the pathway is safe and effective in people. The paper reports no treatment trial.[1]
Dormant viruses during acute illness
The other peer-reviewed Nature study followed multi-omic longitudinal data from 1,154 hospitalised COVID-19 patients in the IMPACC cohort. Researchers detected substantial reactivation of Herpesviridae and Anelloviridae during acute illness, with different viral families following distinct temporal patterns. Reactivation was associated with disease severity, clinical outcomes and increased systemic inflammation. The study also shows that this phenomenon is not confined to immunosuppressed people: it occurs frequently in immunocompetent individuals during severe illness. Viral activity persisted into convalescence, and Anelloviridae were associated with the long-COVID group. Although these longitudinal observations can track change within patients over time, they do not determine the direction of cause. Reactivated viruses could contribute to worse illness; the immune environment of severe illness could awaken dormant viruses; or both could accompany a shared process. The authors explicitly state that they do not establish causation. The reported associations therefore cannot be read as evidence that viral reactivation produced the clinical outcome. Because the cohort comprises hospitalised COVID-19 patients, extending the findings to milder infections or other patient populations would also go beyond what this study demonstrates.[2]
A background condition and a concurrent process
Read together, the studies offer separate examples of how observed risk can vary with biological context. In the FNIP1 research, the context is an ultra-rare protein-truncating variant inherited by a small number of people; the measured outcomes are lipid markers, fat distribution and the odds of cardiometabolic disease. In the virus research, the context is viral activity that changes during severe infection and can extend into convalescence; the measured outcomes include disease severity, clinical outcomes, systemic inflammation and associations with long COVID. The first links a large-scale human genetic association to a pathway investigated in cell and mouse experiments. The second uses a longitudinal observational design to show reactivation moving alongside illness in the same patients over time. These evidence structures cannot be collapsed into one mechanism, one patient group or one clinical outcome. Their shared point is narrower: an inherited characteristic or a concurrent process arising during illness can accompany different risk profiles within a population average. In both papers, that distinction precedes any claim about causal intervention; it describes in greater detail which biological characteristic appears in whom and which outcome it accompanies.[1], [2]
