An AI researcher in the Cathedral of molecular biology (chaitjo.substack.com)

An AI researcher has made significant strides in RNA design by utilizing an AI model, gRNAde, to engineer catalytic RNA functions, specifically targeting the RNA Polymerase Ribozyme. Previously, the model demonstrated an ability to match human experts in complex RNA structure design, but the latest challenge focused on creating functional RNA machines. By leveraging fitness landscape data from earlier directed evolution experiments, the team crafted a "probability mask" that guided their AI's design process, enabling it to navigate mutation tolerance effectively. This approach led to an impressive success rate of 31.5%, dramatically outperforming the standard rational design method, which achieved just 3% success. This breakthrough underscores the potential of AI in synthetic biology, particularly in addressing the complexities of tertiary structure interactions within ribozymes that traditional methods often overlook. The researcher highlighted the importance of interdisciplinary collaboration between AI experts and wet-lab biologists, emphasizing the need for mutual understanding and shared communication styles. By integrating generative AI with experimental biology, the project not only showcased improved performance in designing functional RNA but also illustrated the transformative possibilities of AI in advancing molecular engineering for medicine and sustainability.