About

The critical bottleneck in drug discovery remains poor understanding of the biology underlying disease. As a result, often we don’t know why patients become sick; and many drug candidates fail in trials while numerous devastating diseases remain untreated.  Historically, successful drugs have mostly been discovered by sheer luck. Relation delivers a radically different approach that can better understand the biology of disease and rationally discover new therapeutics.

Relation’s platform uses the power of ActiveGraph ML. The technology has been successfully employed by technology companies to solve problems in computer vision and product recommendations, but never before in drug discovery at this scale. With ActiveGraph machine learning, Relation can understand the huge number of combinatorial functional relationships between genes, proteins, and drugs.

Relation is pioneering a “Lab-in-the-Loop” that can integrate active learning at every step of drug discovery, from predicting cell states to the validation of new targets. An important challenge for any therapeutics company using machine learning is “ground-truth data,” or information known to be true. Working from real cells provided by proprietary biobanks, Relation’s technology generates genomic data that provide direct insights into critical biological relationships that are fed directly into its ML systems. The platform then requests new experiments to improve its predictive ability, cutting through an otherwise intractable combinatorial space.