Business

Relation expands GSK collaboration to USD 110m with MORGAN foundation model for target discovery

Relation expands GSK collaboration to USD 110m with MORGAN foundation model for target discovery

UK-based Relation Therapeutics Limited and GSK have announced an expanded research collaboration focused on generating large-scale human cellular perturbation data and deploying it into AI foundation models, including Relation's MORGAN platform, to support disease biology understanding and novel therapeutic target discovery. Under the agreement, Relation may receive up to USD 110 million in combined upfront and success-based milestone payments. No specific drug candidate is involved; the deal is a pre-competitive, data and platform-level collaboration with no named clinical programs at signing.

Relation will generate large-scale perturbation datasets using advanced human cellular disease models and integrated automation systems capable of producing time-resolved, multi-omics readouts at scale. The resulting data will be used to train and refine MORGAN — Multi-Omic Regulatory Genomics using Artificial Neural Networks — Relation's foundation model of cellular perturbation response across disease contexts. MORGAN is designed to model how human cells respond to genetic and pharmacological interventions at multi-omic resolution, converting high-throughput experimental data into target hypotheses grounded in human disease biology. No peer-reviewed publication for MORGAN has been identified in the public scientific record; the platform's described capabilities rest on company disclosures and GSK's financial commitment as indirect external validation.

The USD 110 million figure bundles upfront and success-based milestone payments without disclosing the respective amounts separately. No milestone gate triggers, royalty rates, territorial rights, or equity investment were disclosed in the announcement.

The agreement extends a relationship between the two companies established in December 2024, when GSK paid Relation USD 45 million upfront — including a USD 15 million equity investment — across two multi-program strategic collaborations targeting fibrotic diseases and osteoarthritis. Those agreements made Relation eligible for success-based collaboration payments of up to USD 63 million, plus preclinical, development, commercial, and sales milestone payments averaging USD 200 million per target across both deals, along with tiered royalties on net sales. Under that structure, Relation led observational studies to generate functional disease datasets, while GSK retained worldwide development and commercialization rights to any resulting targets. The July 2026 deal is described as an expansion of that existing collaboration, extending the data generation and foundation model training work beyond the original disease indications. GSK's stated rationale in December 2024 was to gain "deep and precise insights into human causal biology" through Relation's platform, a framing consistent with the expanded scope now announced.

The AllSci BriefSystematic R&D and deal news. Daily.

Relation has established a pattern of platform-based pharma collaborations in rapid succession. In December 2025, the company signed a strategic collaboration with Novartis covering atopic diseases, with Novartis committing USD 55 million upfront and up to USD 1.7 billion in total milestones to pair Relation's AI-powered discovery platform with Novartis's immuno-dermatology expertise. In January 2026, Relation entered a separate research collaboration with Deerfield Management, under which nominated targets can be advanced via a jointly owned NewCo, with both parties entitled to royalties on future product sales. The GSK expansion is Relation's third major pharma partnership in under two years, each structured around the same core value proposition: proprietary human cellular data generation combined with AI-driven target identification.

GSK has pursued a parallel strategy of licensing AI cellular foundation models from multiple platform companies. In January 2026, it licensed Noetik's OCTO virtual cell foundation models for oncology research in a USD 50 million upfront deal, gaining access to multimodal spatial training datasets to support cancer drug discovery — a structurally analogous transaction to the MORGAN model arrangement with Relation. GSK has also signed data-driven target discovery collaborations with Muna Therapeutics in Alzheimer's disease, Engitix in liver fibrosis regression, and the JAX-NYSCF Collaborative in neurodegeneration, each involving human-derived biological datasets and computational analysis as the primary deliverable rather than a named clinical asset.


Access the AllSci platform to explore the science behind the news.


Spot something wrong? Report an issue with this article