Russian AI slashes drug-protein analysis time to 13 seconds per molecule

Russian researchers have developed an AI system that predicts drug-protein interactions far quicker than existing tools. The team, led by Daria Frolova from Skolkovo Institute of Science and Technology (Skoltech) and Ligaand Pro, claims their method is 30 times faster than AlphaFold for analysing a single molecule. This breakthrough could speed up virtual drug discovery and make it more accessible to labs with standard computing power. The new system calculates how a single molecule interacts with a protein in just 13 seconds. In comparison, AlphaFold—the current industry standard—takes around 6.5 minutes for the same task. The researchers achieved these results after three years of development.

The AI operates in three key stages. First, it roughly positions the molecule near the protein. Next, it refines the molecule's orientation. Finally, it filters out impossible configurations to leave only viable interactions. Beyond speed, the team aims to create tools for generating new molecules, predicting their properties, and optimising their design. These features could further streamline drug development by reducing reliance on costly lab experiments. Skoltech's press service announced the breakthrough, highlighting its potential to transform early-stage drug research. The next phase involves experimental validation and encouraging adoption across the pharmaceutical industry.

The system's efficiency may allow virtual drug screening to run on mid-range hardware, lowering costs for research teams. Experimental testing and industry uptake will determine how widely the technology is adopted. If successful, it could become a standard tool in computational drug discovery.