molclaw-mol-opt-physchem

What problem does it solve?

This Skill streamlines iterative optimization of small-molecule physicochemical properties (LogP, QED, solubility) by combining computed property analysis, optional RL-driven candidate generation, and LLM-guided structural edits while preserving stereochemistry and synthetic feasibility.

Core Features & Use Cases

• Computed Property Analysis: Integrates with molclaw-admet to generate precise ADMET and physicochemical metrics used as the basis for edits.
• RL-Assisted Candidate Generation: Supports numeric-target optimization via reinvent_similarity_optimization to explore nearby chemotypes when precise property ranges are specified.
• LLM-Guided Design + Validation: Provides stepwise medicinal-chemistry rationale, preserves stereochemical markers, and routes generated SMILES through molclaw-smiles-valid-check before re-evaluation.
• Use Case: Improve aqueous solubility of a lead compound while maintaining core pharmacophore and synthetic accessibility.

Quick Start

Optimize the provided SMILES to decrease LogP into the 1–3 range and increase QED above 0.7 while preserving all stereochemical markers.