Relation Therapeutics and GSK announced an expanded research collaboration to generate large-scale cellular perturbation data for cellular foundation models. Under the terms, Relation could receive up to $110 million in upfront and milestone payments as the companies build general-purpose models intended to improve target discovery. Relation also unveiled MORGAN (Multi-Omic Regulatory Genomics using Artificial Neural Networks), a multi-cell-type model meant to predict how cells respond to genetic and pharmacological interventions. GSK and Relation did not specify disease focus areas, but the approach centers on petascale multiomic perturbation datasets produced through automated laboratories. Relation positions the work as a data-first strategy—using high-throughput, time-resolved, multi-omic readouts to increase confidence in target selection and validation. Relation’s CEO David Roblin framed the collaboration as addressing the “data” problem as much as the modeling challenge.