AI Insight
ToolMol is an evolutionary agentic framework for de novo drug design that integrates a multi-objective genetic algorithm with a large language model acting as an iterative operator to refine candidate ligand populations. The system employs a curated toolbox of RDKit-backed functions that allow the LLM to make precise, chemically valid molecular modifications, overcoming the syntactic limitations that cause high rates of invalid outputs in prior LLM-based approaches. Evaluated against three protein targets, ToolMol achieves over 10% stronger predicted binding affinity than competing methods and improves Absolute Binding Free Energy scores by more than 35%, while producing drug-like and synthesizable ligand candidates.
Why it matters
This framework could accelerate early-stage drug discovery by reducing the time and cost associated with generating viable candidate molecules, with potential downstream benefits for diseases where targeted small-molecule therapies are needed. The tool-calling approach also offers a generalizable strategy for improving chemical reasoning in LLM-based scientific applications.
arXiv:2605.12784v2 Announce Type: replace-cross
Abstract: Advances in large language models (LLMs) have recently opened new and promising avenues for small-molecule drug discovery. Yet existing LLM-based approaches for molecular generation often suffer from high rates of invalid and low-quality ligand candidates, a result of the syntactic limitations of current models with regard to molecular strings. In this paper, we introduce $texttt{ToolMol}$, an evolutionary agentic framework for de novo drug design. $texttt{ToolMol}$ combines a multi-objective genetic algorithm with an agentic LLM operator that iteratively updates the ligand population. We build a comprehensive toolbox of RDKit-backed functions that allows our agentic operator to consisently make precise ligand modifications. $texttt{ToolMol}$ achieves state-of-the-art performance on multi-objective property optimization tasks, discovering drug-like and synthesizable ligands that have $>10%$ stronger predicted binding affinity compared to existing methods, evaluated on three protein targets. $texttt{ToolMol}$ ligands additionally achieve state-of-the-art results in gold-standard Absolute Binding Free Energy scores, gaining over existing methods by over $35%$. By studying chain-of-thought reasoning traces, we observe that tool-calling enables the model to more faithfully execute its planned modifications, efficiently exploiting the strong chemical prior knowledge in LLMs.
Source: ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery