CLADD

Research applicability

RAG-empowered agentic system for drug discovery: multiple LLM agents retrieve from biomedical knowledge bases, contextualize query molecules and integrate evidence without domain-specific fine-tuning; official code available.[16]

Getting started

Begin at github.com/Genentech/CLADD (official source code, checked live 2026-10-16); review the AAAI-26/arXiv paper for usage details.

Maintenance

Maintenance status not independently audited.

License

Repository licence field reported as NOASSERTION via GitHub API (2026-10-16); re-check licence before adoption; no independent license audit performed.

Workflow & human oversight

Resource-level classification is intentionally unspecified; instantiated workflows may differ. Humans set objectives and validate results.

Workflow level is not an evidence score; higher is not better evidence.

Approach

Collaborative LLM agents with retrieval augmentation over biomedical knowledge bases, addressing data heterogeneity, ambiguity and multi-source integration in biochemical RAG workflows; reported to outperform general-purpose and domain-specific LLMs and traditional deep learning baselines across drug-discovery tasks.[16]

Outcomes (as reported by the originating team)

Author-reported: outperforms general-purpose and domain-specific LLMs as well as traditional deep learning approaches across a variety of drug-discovery tasks.

Limitations

Author-reported evaluations; not independently reproduced. Repository availability verified (GitHub API); no software executed.

Reproducibility

Live source checks 2026-10-16: arXiv:2502.17506; AAAI-26 DOI 10.1609/aaai.v40i1.37020; repository API (Genentech/CLADD). No claimed outcome reproduced.

Adoption requirements

Python code from the official repository (github.com/Genentech/CLADD); LLM access and biomedical knowledge bases required; no independent setup was performed.