An AI-assisted genomic census of 193,342 assemblies finds a GFP-barrel family in unreported lineages, carrying pseudogenes in both fungi and cnidarians

Life on Earth relies on a vast, invisible library of genetic instructions, a code written in four chemical letters that tells every cell how to build itself. For decades, scientists have focused on the parts of this code that produce active, working proteins, the molecular machines that power everything from digestion to vision. But a significant portion of the genome is often dismissed as "junk," containing broken or incomplete instructions that no longer function. In the world of marine biology, one of the most famous genetic discoveries is the green fluorescent protein, a molecule that glows in the dark and has become a standard tool for watching cells work in a lab. This protein folds into a sturdy, barrel-like shape that traps a glowing chemical center. For years, researchers believed that if this barrel shape existed without the glowing center, it was likely just a broken, useless remnant of evolution.

However, a new study challenges this assumption by looking much deeper into the genetic code than ever before. Instead of relying on the list of known, working proteins, the researchers scanned the raw genetic blueprints of nearly two hundred thousand different organisms. They were hunting for a specific, hidden pattern: a barrel-shaped protein that kept the complex internal machinery of the glowing protein but had lost the glowing part itself. What they found was a widespread family of these non-glowing barrels, carrying a specific network of electrical charges inside them, existing in lineages of life that had never been suspected of having them. The study reveals that these proteins are not just random genetic debris; in many cases, they are being carefully preserved by evolution, suggesting they perform a vital, yet-to-be-discovered job in the dark.

The research began with a simple but massive question: how far does this specific family of barrel proteins reach across the tree of life? Previous studies had only looked at the proteins that cells had already built and cataloged, missing any genetic instructions that were broken or incomplete. To find the full picture, the team used a computer-assisted search that read the DNA of 193,342 different genome assemblies in every possible direction. This approach allowed them to spot genetic sequences that looked like the barrel protein even if they contained errors that would stop a cell from making the final product. The search returned nearly four hundred locations in the genome that matched the pattern. When the researchers filtered these results to find the ones that carried the complete, hidden electrical network inside the barrel, they identified 140 distinct locations. These locations belonged to 77 different types of organisms, expanding the known range of this protein family significantly.

Among the new discoveries were organisms that had never been linked to this family before. The search uncovered these barrel proteins in ancient groups of sponges, a type of jellyfish known as a coronate, and even in single-celled organisms called archaea that live in extreme environments. Perhaps most surprisingly, the team found these proteins in fungi, a group of organisms that includes mushrooms and molds. In the fungal samples, the researchers found that a substantial number of these genetic instructions were broken, carrying errors that would prevent them from working. This pattern of broken copies was also seen in the jellyfish and coral relatives, suggesting that while the protein is useful in some lineages, it is falling apart in others. The study used a method of evolutionary reconstruction to show that in a specific group of black molds, the gene has been lost independently many times, indicating it is not essential for their survival.

Despite the broken copies, the parts of the genome that remained intact told a different story. In the organisms where the gene was still whole, the DNA showed signs of being under strict evolutionary pressure to stay the same. This means that any random changes to the protein's structure were being weeded out by natural selection, a strong signal that the protein is doing something important. In one type of coral, the gene had even been copied eleven times in a row, a sign that the organism is relying heavily on it. The researchers also found that in two different groups of bacteria, these barrel genes were sitting inside mobile genetic elements, such as viruses that infect bacteria or jumping genes that move around the genome. This suggests the gene might be moving between different species of bacteria, a process that could help it spread to new environments.

The study also looked closely at the specific chemical makeup of these proteins. In the famous glowing protein, a specific amino acid called tyrosine is essential for creating the light. In this newly discovered family, that tyrosine is replaced by different, non-glowing amino acids. In the fungal and bacterial versions, the spot is usually filled by valine, while in the animal versions, it is often phenylalanine. This consistent replacement suggests that the protein has evolved a new function that does not involve light. The researchers measured the distances between the internal electrical charges and found that the arrangement in this family is distinct from the glowing proteins, further confirming that they are a unique group. While the exact job of these proteins remains a mystery, the evidence that they are being preserved in some species and lost in others points to a specialized role that varies across the tree of life.

One of the most intriguing findings was the presence of these proteins in organisms that do not have them in their protein catalogs. Because the search looked directly at the DNA, it found genetic instructions that had been missed by traditional methods. For example, the study found these barrels in a group of microscopic algae that live inside other marine animals, and in a type of sponge that had never been reported to carry them. The researchers were careful to rule out the possibility that these were just contaminants from other organisms, using multiple checks to ensure the genes truly belonged to the host. They also found that in some cases, the broken versions of the gene were still being read by the cell, even though they could not make a working protein, which adds another layer of complexity to how these genetic instructions are used.

The work does not claim to have solved the mystery of what these proteins do. Instead, it provides a clear map of where they are found and how they are changing over time. The researchers suggest that the next step is to actually build these proteins in a lab and test them to see what chemical reactions they can perform. Until then, the study stands as a powerful demonstration of how looking directly at the raw genetic code, rather than just the finished products, can reveal hidden chapters of evolutionary history. It shows that nature often keeps complex molecular machines even after they lose their original function, repurposing them for new, unseen roles in the dark corners of the living world.