Harvard Researchers Uncover a 'Dark Genome' of Chimeric mRNAs That Make Functional Proteins

Researchers at Harvard Medical School have discovered that genes located on different chromosomes can combine to produce chimeric messenger RNAs that encode entirely new proteins, a finding that upends a core assumption of molecular biology and reveals what the team describes as a hidden layer of the genome. The study, published in Nature on September 2, 2026, used a technique called direct RNA sequencing to catalogue more than 30,000 chimeric mRNAs in mammalian cells, a collection the authors call the "dark genome" library because it was invisible to the standard methods that have mapped gene expression for decades. Senior author Ruaidhrí Jackson said in a release distributed by EurekAlert on Sep 2 that the discovery means scientists have been reading an incomplete blueprint: "We thought we had a blueprint of every mRNA that is made in the body, and now we're saying that was just page one."

The finding matters because messenger RNA is the molecule that carries instructions from DNA to the cellular machinery that builds proteins, and the prevailing assumption has been that each mRNA is transcribed from a single contiguous stretch of one gene. The Harvard team showed that during an immune response, healthy chromosomes can physically loop together inside the nucleus, bringing distant genes into proximity so that the cellular machinery transcribes a hybrid mRNA that takes sequence from each gene. The result is a chimeric protein that does something neither parent protein does on its own, and the team demonstrated that at least one such chimera plays a real role in fighting infection, a functional proof that moves the discovery from cataloguing exercise to biological mechanism.

Key Facts

The scale of the discovery is best measured by the numbers. Direct RNA sequencing of mammalian cells produced a catalogue of more than 30,000 chimeric mRNAs, according to the Nature paper published on Sep 2, and about 400 of those were shown to be regulated by inflammatory signals, with many conserved across both human and mouse immune cells, a conservation pattern that strongly suggests biological function rather than noise. news-medical.net reported on Sep 2 that the team profiled how these inflammatory-regulated chimeras respond to stimulation, establishing that the hybrid transcripts are not rare accidents but a regulated part of the immune response.

The decisive experiment centered on a specific chimera formed from two genes, GSDMD and TMEM106A. GSDMD alone is known for opening cell membranes during pyroptosis, a form of inflammatory cell death, and the researchers found that a GSDMD-TMEM106A chimeric protein forms naturally in mice during immune responses. Mice engineered to lack the chimeric protein, while keeping the normal GSDMD and TMEM106A genes intact, showed a significantly slower immune response and were roughly 50 percent less able to control a Salmonella infection, according to the study. In sepsis models, the same mice had milder responses to endotoxin, while restoring the chimera by delivering its mRNA, in work done in collaboration with Moderna, made mice more sensitive to endotoxin, confirming that the chimeric protein amplifies the inflammatory cell death pathway that GSDMD normally drives.

The collaboration with Moderna is notable because it shows the discovery has immediate practical tools attached to it. The messenger RNA technology that Moderna helped pioneer for vaccines can be used to deliver specific chimeric mRNAs into cells, which gave the Harvard team a way to test causality: add the chimera and the immune response strengthens, remove it and the response weakens. Technology Networks reported on Sep 2 that this experimental loop, catalogue, regulate, delete, restore, is what elevates the finding from a large genomic survey into a demonstrated new axis of gene regulation, and it opens the door to studying whether other chimeras in the 30,000-entry library have equally important roles in health and disease.

Analysis

What this really means is that the genome may contain far more functional information than the reference sequence suggests, and that our understanding of how genes encode traits has a systematic blind spot. The standard model treats each gene as a unit that produces one protein, with complexity added by alternative splicing, in which a single gene's exons are combined in different ways. The Harvard discovery describes something qualitatively different: genes that are not adjacent on the same chromosome, and in some cases not even on the same chromosome, joining together to create a transcript that never appears in any gene annotation. If even a fraction of the 30,000 chimeric mRNAs turn out to be functional, the effective size of the proteome, the full set of proteins a cell can make, is far larger than textbooks describe, and every disease study that relies on gene expression data has been missing a layer of information.

The bigger picture here is what the discovery implies for the immune system specifically. The chimeras the team characterized are induced by inflammatory signals, which suggests that cells may use this chromosome-looping mechanism as a rapid-response system, generating new protein variants on demand when the body is under threat, without waiting for the slow process of genetic mutation and natural selection. That would make the chimeric mRNA system a form of fast, reversible genome innovation, and it would explain why the immune system, which must adapt quickly to novel pathogens, would be the place where such a mechanism is most active. The Salmonella experiment is the proof of principle: the chimera is not a bystander but a functional component of the anti-bacterial response, and its absence measurably weakens the mouse's defense.

There are important caveats. The functional validation in the paper focuses on a single chimera in mice, and the leap from a catalogued chimera to a demonstrated function is enormous; most of the 30,000 transcripts will likely turn out to be noise or to have subtle effects that are hard to measure. The mouse results also do not yet establish that the same chimeras matter in human disease, though the conservation of inflammatory-regulated chimeras across species is encouraging. And the discovery raises as many questions as it answers, about how chromosome looping is controlled, why some genes are selected for chimera formation and others are not, and whether errors in the system contribute to autoimmune disease, in which the immune system attacks the body's own tissues. If chimeric proteins are involved in autoimmunity, the finding could point to an entirely new class of drug targets for diseases that have resisted treatment.

Why It Matters

For biologists, the discovery expands the map of what the genome can do and will likely send many labs back to their own datasets to look for chimeric transcripts they previously discarded as artifacts. For the pharmaceutical industry, the finding suggests a new class of therapeutic targets: if chimeric proteins amplify inflammation, drugs that block specific chimeras could treat inflammatory and autoimmune diseases with a precision that targeting the parent proteins cannot achieve, and the Moderna collaboration points to a direct way to test such drugs. For the field of immunology, the work reframes how the innate immune system achieves its flexibility, and it may explain longstanding puzzles about why the body's response to infection is so much faster and more adaptable than the fixed set of genes would predict. For patients with inflammatory diseases, the discovery opens a research avenue that did not exist a week ago, though treatments based on it remain years away.

Next Up

In the coming months, watch for follow-up studies that test whether chimeric proteins play a role in specific human diseases, particularly autoimmune and inflammatory conditions where the GSDMD pathway is already implicated. Watch also for the Harvard team's broader functional screen of the 30,000-entry library, since the field needs to know how many of the chimeras are biologically active rather than transcriptional noise. The most important near-term development will be replication: if other laboratories confirm that chimeric mRNAs are a widespread, regulated feature of mammalian cells, the discovery will move quickly from a single striking paper into a new subfield, and the "dark genome" will become a standard chapter in how gene regulation is taught.