Pooja Arora, ThoughtWorks Inc, HCLS Advisory & Strategy lead
Ganesh Kumar Sathiyamoorthi, Thoughtworks Inc, Principal Data Engineer

Clinical trials face a high failure rate, with 90% of drugs failing to progress due to challenges in translating preclinical results into clinical outcomes. Sydney Brenner’s reverse translation approach advocates for using clinical trials and unexpected patient responses to drive new hypotheses. However, foundational issues like data silos and semantic disparities hinder hypothesis generation and collaboration. To address these challenges, we propose an ecosystem of “AI scientists” in self-driving labs that enable bedside-to-benchtop research. These AI agents integrate AI models, semantics, and biomedical tools with experimental platforms to organize experimental data and drive meaningful reasoning. Our case study from a pharmaceutical merger highlights the use of standardized data vocabularies, knowledge graphs, and ontologies to improve drug discovery. Key components include ontology-enriched data products, knowledge graph construction, entity and relationship onboarding, and automated annotation recommendations. We enhance researcher experience by integrating semantic layers with large language models, facilitating intuitive searches and hypothesis generation. The semantic agent, integrated with a data mesh platform, unifies siloed data and enables collaborative discovery through specialized AI-driven agents. This framework fosters a dynamic, patient-centric ecosystem where clinical insights continuously inform preclinical research, driving innovation.