Transforming Material Design and Drug Discovery? AI × Molecular Simulation Unveils the "Strongest Knot" - The Secret of Proteins That "Don't Come Apart When Pulled"

A research team from Auburn University and Colorado State University has combined molecular dynamics (MD) simulations with AI regression models to analyze the operational principles of "catch bonds," considered among the strongest in nature. Contrary to the traditional hypothesis that these bonds strengthen "after a certain extension," the study shows that bond strengthening begins "almost immediately" after force is applied. The focus was on adhesive complexes derived from bacterial cellulosomes, where AI accurately predicted the breaking time from short snippets of MD "pulling" videos. The results provide new insights for designing materials in areas such as skin and blood vessel adhesion, tissue durability, and mechanobiology-based drug development, adhesives, and soft robotics. Phys.org featured these findings on September 11, 2025, linking to a reliable primary source (a paper published in JCTC). On social media, there is both excitement about AI's ability to detect "early signs" from dynamic data and cautious voices regarding the generalization of these findings to experimental systems.