Relation, GSK Partner in $110M AI Drug Discovery Collaboration

Relation and GSK have formed a strategic research collaboration to generate large-scale human cellular perturbation data and integrate those datasets into advanced artificial intelligence models, including Relation’s cellular biology foundation model, MORGAN. The partnership combines high-throughput experimental biology with machine learning to improve understanding of disease mechanisms and accelerate the discovery of novel therapeutic targets.

Under the agreement, Relation is eligible to receive up to $110 million in upfront and success-based milestone payments. The companies will jointly create a comprehensive resource of human cellular response data that could strengthen confidence in early-stage target identification and drug discovery.

Leveraging Human Biology and AI for Target Discovery

The collaboration focuses on generating high-quality perturbation datasets that capture how human cells respond to both genetic and pharmacological interventions. These experiments will use physiologically relevant human disease models and integrated laboratory automation to produce reproducible, time-resolved datasets at substantial scale.

Each experiment will incorporate multi-omics readouts, enabling researchers to examine molecular changes across multiple biological layers following cellular perturbation. The resulting datasets will serve as training material for foundation models capable of identifying complex biological patterns associated with disease.

A deeper understanding of cellular responses may help researchers identify disease-driving pathways, prioritize therapeutic targets, and improve the efficiency of drug discovery before compounds enter clinical development.

MORGAN Foundation Model at the Center of Collaboration

The datasets generated through the collaboration will support the continued development of MORGAN (Multi-Omic Regulatory Genomics using Artificial Neural Networks), Relation’s foundation model for predicting cellular perturbation responses across different disease contexts.

By combining experimental human biology with large-scale machine learning, MORGAN analyzes complex multi-omics data to model cellular behavior and uncover biological relationships that conventional analytical approaches may overlook. The expanded dataset generated through the partnership is expected to improve model performance across multiple disease areas and enhance target validation efforts.

Executive Perspective

David Roblin, Chief Executive Officer of Relation, said that improving understanding of disease biology begins with richer experimental datasets capable of informing more reliable computational models.

He noted that the collaboration will generate novel datasets from physiologically relevant human disease systems, helping reveal important biological mechanisms while supporting future therapeutic target discovery. Roblin added that integrating advanced experimental platforms with computational modeling will create a valuable resource for understanding disease biology.

Future Development

The collaboration reflects a broader industry trend toward combining human biology, automation, multi-omics technologies, and artificial intelligence to improve early-stage drug discovery. Rather than focusing on a single therapeutic candidate, the partnership establishes a scalable research platform that could generate insights across multiple disease areas.

As the perturbation datasets continue to expand, both companies expect the resulting foundation models to improve target identification, biological validation, and future therapeutic discovery programs, supporting a more data-driven approach to developing next-generation medicines.

Reference

Relation Announces Expanded Collaboration with GSK