AI‑Enabled Discovery of High‑Value Therapeutic Targets in Obesity

1. Disease (Obesity) Landscape Creation

We generated a knowledge graph -to map target-disease associations integrating epidemiology, clinical subtypes, and unmet needs across obesity and associated comorbidities, along with current standards of care, clinical pipeline and key end points. The knowledge graph also integrated data from emerging biological pathways involved in energy homeostasis, appetite regulation, adipose tissue biology, inflammation, and metabolic control.

2. Target identification and assessment

Insights derived from the Obesity -specific knowledge graph were used to compile a comprehensive list of candidate targets supported by evidence from scientific literature, clinical databases, and other relevant data sources.

RxAgentAi enabled systematic analysis and rank ordering of these targets through a data-driven, multi-parametric assessment framework. Targets were evaluated across predefined criteria, including –

• Strength of biological rationale and human genetic evidence
• Disease relevance and mechanistic linkage
• Differentiation potential relative to existing and emerging therapies
• Druggability and modality suitability
• Availability of translational biomarkers and disease-modifying readouts
• Safety considerations based on historical attrition patterns

An expert-in-the-loop review process was integrated to validate AI-generated insights, refined underlying assumptions, and resolved uncertainty, ensuring scientific rigor and robustness of prioritization.

3. Target Prioritization framework

A structured target prioritization framework was applied to objectively evaluate and rank candidate targets using a weighted scoring approach based on scientific robustness, clinical feasibility, and strategic fit. Comparative analyses were performed against benchmarked competitor programs and existing pipeline assets to contextualize each target within the current therapeutic landscape and identify opportunities for differentiation. This systematic evaluation enabled robust, evidence-driven rank ordering of targets to support informed decision-making in early discovery.