Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

From Literature to Hypotheses: An AI Co-Scientist System for Biomarker-Guided Drug Combination Hypothesis Generation

Raneen Younis, Suvinava Basak, Lukas Chavez, Zahra Ahmadi

TL;DR

This work presents AI Co-Scientist (CoDHy), an interactive, human-in-the-loop system for biomarker-guided drug combination hypothesis generation in cancer research, and demonstrates its design, interaction workflow, and practical use cases.

Abstract

The rapid growth of biomedical literature and curated databases has made it increasingly difficult for researchers to systematically connect biomarker mechanisms to actionable drug combination hypotheses. We present AI Co-Scientist (CoDHy), an interactive, human-in-the-loop system for biomarker-guided drug combination hypothesis generation in cancer research. CoDHy integrates structured biomedical databases and unstructured literature evidence into a task-specific knowledge graph, which serves as the basis for graph-based reasoning and hypothesis construction. The system combines knowledge graph embeddings with agent-based reasoning to generate, validate, and rank candidate drug combinations, while explicitly grounding each hypothesis in retrievable evidence. Through a web-based interface, users can configure the scientific context, inspect intermediate results, and iteratively refine hypotheses, enabling transparent and researcher-steerable exploration rather than automated decision-making. We demonstrate CoDHy as a system for exploratory hypothesis generation and decision support in translational oncology, highlighting its design, interaction workflow, and practical use cases.

From Literature to Hypotheses: An AI Co-Scientist System for Biomarker-Guided Drug Combination Hypothesis Generation

TL;DR

This work presents AI Co-Scientist (CoDHy), an interactive, human-in-the-loop system for biomarker-guided drug combination hypothesis generation in cancer research, and demonstrates its design, interaction workflow, and practical use cases.

Abstract

The rapid growth of biomedical literature and curated databases has made it increasingly difficult for researchers to systematically connect biomarker mechanisms to actionable drug combination hypotheses. We present AI Co-Scientist (CoDHy), an interactive, human-in-the-loop system for biomarker-guided drug combination hypothesis generation in cancer research. CoDHy integrates structured biomedical databases and unstructured literature evidence into a task-specific knowledge graph, which serves as the basis for graph-based reasoning and hypothesis construction. The system combines knowledge graph embeddings with agent-based reasoning to generate, validate, and rank candidate drug combinations, while explicitly grounding each hypothesis in retrievable evidence. Through a web-based interface, users can configure the scientific context, inspect intermediate results, and iteratively refine hypotheses, enabling transparent and researcher-steerable exploration rather than automated decision-making. We demonstrate CoDHy as a system for exploratory hypothesis generation and decision support in translational oncology, highlighting its design, interaction workflow, and practical use cases.
Paper Structure (46 sections, 3 figures, 4 tables)

This paper contains 46 sections, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Design
  4. User Interface and Interaction Workflow
  5. Knowledge Graph Construction
  6. Graph Embedding Generation
  7. Hypothesis Generation Agent
  8. Hypothesis Validation Agent
  9. Hypothesis Ranking and Presentation
  10. System Deployment and Implementation (Demo)
  11. Backend service.
  12. Reproducibility and availability.
  13. Evaluation Results and Analysis
  14. Experimental Setup
  15. Scenarios.
  16. ...and 31 more sections

Figures (3)

  • Figure 1: Overview of the CoDHy system architecture. The user interacts through a web-based interface to specify the cancer context, biomarker, and literature scope. The backend constructs a task-specific knowledge graph from structured databases and PubMed literature, learns graph embeddings, generates biomarker-guided drug combination hypotheses, validates and ranks them using multi-agent reasoning, and returns ranked, evidence-grounded hypotheses to the user.
  • Figure 2: Web-based user interface of the CoDHy system. (a) The interface allows researchers to select the focus biomarker, cancer type, language model, and the scope of literature retrieval from PubMed, and to control the number of hypotheses generated. (b) Once the analysis is done, the generated hypotheses are shown based on their rank.
  • Figure 3: Example task-specific knowledge graph. Node colors denote entity types (e.g., gene, drug, disease, variant etc.), and edges correspond to integrated biomedical relations used by CoDHy.