AI Helps Scientists Discover a Mysterious New DNA System

Scientists are using AI to search through huge amounts of genetic data—and it just helped uncover something they had never seen before. Researchers working with Anthropic’s Claude AI say the system helped identify a previously unknown biological system hidden inside the DNA of viruses that infect bacteria. The discovery is interesting because part of the system looks somewhat similar to CRISPR, the famous technology scientists use to study and modify DNA. But there is an important difference: researchers do not yet know whether this new system can edit genes like CRISPR. For now, scientists are trying to understand what it actually does. So, what did AI find? Our DNA contains an enormous amount of information. Scientists have already discovered millions of genes, but much of the genetic material found in nature is still poorly understood. The researchers asked Claude to search through massive genetic databases looking for unusual combinations of biological components. The AI investigated hundreds of thousands of possible candidates and eventually identified a strange group of genes that appeared to belong together. The system was given the name array-associated reverse transcriptases, or ARTs. These systems were found in bacteriophages—viruses that infect bacteria. Why does it look like CRISPR? CRISPR systems are famous because they can use pieces of genetic information to recognize specific DNA sequences. The newly identified ART systems contain unusual repeating DNA sequences alongside an enzyme called a reverse transcriptase. That combination caught the researchers' attention because the repeating sequences have some similarities to structures seen in CRISPR-related biology. However, similarity does not mean the systems work in the same way. Scientists still need to perform experiments to determine exactly what ART systems do inside cells. Where does AI come into this? This is where the discovery gets especially interesting. Instead of asking AI to simply explain existing research, researchers used Claude as a tool for exploring enormous biological databases. Anthropic says about 950 AI agents worked on the search for more than 21 hours. The agents examined hundreds of thousands of reverse-transcriptase candidates and searched for unusual patterns that might have been overlooked. Eventually, the AI highlighted a promising candidate. Human scientists then took over and investigated the discovery using laboratory experiments. In other words, AI didn't simply announce a new biological technology out of nowhere. It helped researchers search through an enormous amount of information and identify something worth investigating. Could this replace CRISPR? Not at this point. There is currently no evidence that ART systems can replace CRISPR or even perform the same job. Researchers still need to understand how the system works, what biological purpose it serves, and whether it can be controlled or used for biotechnology. That could take considerable time. But the discovery shows why AI could become useful in biological research. There are enormous genetic databases containing information from organisms and viruses around the world. Finding unusual patterns manually can be extremely difficult. AI systems can help researchers search those databases much faster and point scientists toward interesting possibilities. Why this matters The most important part of the discovery may not be the new enzyme system itself. It is the possibility that AI can help scientists discover biological systems that were hiding in plain sight. Nature contains an enormous number of genes and molecular mechanisms that humans haven't fully understood. As AI becomes better at recognizing patterns in biological data, researchers may be able to uncover many more of them. For now, ART remains a mystery. Scientists know that the system exists. They have identified its unusual genetic structure and begun studying it in the laboratory. But the biggest question is still unanswered: What can it actually do? That answer could determine whether this discovery becomes an interesting biological curiosity—or eventually leads to a useful new biotechnology.