A transformer model that translates biology into medicine, at scale

Whole-cell computational modelling - understanding how a disease changes a cell and how a drug changes it back - has been described as the ultimate goal of systems biology and a grand challenge for the 21st century. The obstacle is not ambition but scale: the relationships between genes, compounds, diseases, and biological pathways are too numerous and too interconnected for conventional analytical approaches. A platform that could translate the abstract language of biology into drug-disease relationships computationally would compress discovery timelines and open new repositioning opportunities that manual research cannot find.

CellMo.AI is E-Group's platform for predictive cellular biology, combining current network analysis methods with the transformer architecture behind breakthroughs in large language models. The platform is built on an integrative knowledge graph assembled from 29 databases covering genes, compounds, diseases, biological pathways, and molecular interactions - 47,031 vertices with 11 labels and 2,250,197 edges with 24 labels. The CellMo transformer model (50M+ parameters) translates cellular effects of diseases into pathophysiological effects of drugs, enabling computational identification of repositioning candidates. The platform operates in three layers: a Data Layer (29-database knowledge graph), a Service Layer (CellMo transformer analytics for drug repositioning, disease modelling, and combination prediction), and a User Interaction Layer (open-source interfaces for research workflows without deep technical knowledge). CellMo.AI's drug repositioning capability directly powers E-Group's Animal Health Innovation Platform for antibiotic resistance research.

CellMo.AI is the computational engine underlying E-Group's drug repositioning programmes. Its connection to the Animal Health Innovation Platform demonstrates the transition from research capability to applied deployment. The platform positions E-Group at the frontier of AI-driven pharmaceutical research - applying the same deep learning architecture that transformed language processing to the challenge of biological intelligence.