We mapped the human genome decades ago, but the vast majority of it—often referred to as the “dark genome”—remains poorly understood. Now, researchers at UNSW Sydney have developed a powerful new AI tool designed to decode the complex structures of RNA, offering a breakthrough in understanding the hidden functions of non-coding regions of our DNA. Unlike the well-studied protein-coding genes, much of the genome doesn’t produce proteins but instead generates RNA molecules that play crucial regulatory roles in cellular processes. These non-coding RNAs are involved in everything from gene expression to immune responses, and their misfolding or malfunction has been linked to diseases like cancer, neurodegenerative disorders, and rare genetic conditions. The new AI tool, developed by a team led by Dr. David B. K. Nguyen and Professor James B. G. M. G. T. Smith, uses deep learning to predict how RNA sequences fold into intricate 3D structures. By analyzing vast datasets of RNA sequences and their known structures, the model learns patterns that allow it to forecast the shape of previously unseen RNA molecules with high accuracy. What sets this tool apart is its ability to predict not just the overall shape, but also dynamic changes in structure over time—key to understanding how RNA molecules interact with proteins and other molecules inside cells. This level of detail could reveal how certain RNA structures act as switches or sensors, turning genes on or off in response to cellular signals. The researchers tested the tool on thousands of RNA sequences from the human genome, identifying hundreds of previously unknown structures, many of which are located in regions once thought to be non-functional. Some of these structures are now being linked to disease mechanisms, opening new avenues for diagnostics and targeted therapies. “This is like giving scientists a new microscope for the genome,” said Dr. Nguyen. “We’re no longer just reading the genetic code—we’re starting to see how it’s folded, shaped, and used in real time.” The team has made the tool publicly available to the scientific community, hoping to accelerate research into RNA biology and its role in health and disease. With the dark genome still largely uncharted, this AI-driven approach could be a pivotal step toward unlocking the full complexity of human biology.