An open-source screening platform accelerates discovery of drug combinations (www.nature.com)

Researchers have announced the launch of Combocat, an open-source platform designed to significantly accelerate the discovery of drug combinations. Traditional methods for screening drug combinations have faced limitations in throughput and efficiency, primarily due to the exponential increase in complexity as more drugs are added to tests. Combocat leverages advanced acoustic liquid handling technologies and integrates machine learning to create an ultrahigh-throughput screening framework, allowing for the testing of thousands of combinations with minimal resources. Notably, the platform achieved a milestone by screening 9,045 combinations in a neuroblastoma cell line, marking the largest testing effort in a single cell line to date. The significance of Combocat lies in its dual operational modes: a "dense mode" that generates comprehensive datasets through detailed measurements, and a "sparse mode" which estimates outcomes based on limited experimental data, drastically boosting efficiency. By generating a reference dataset of over 800 unique drug combinations in a 10 × 10 matrix format, and applying machine learning to predict results from sparse measurements, Combocat minimizes the need for extensive experimental inputs. This integration allows researchers to rapidly identify effective drug combinations, paving the way for innovative therapeutic strategies in oncology and beyond.