skills.rest
InternScience
InternScienceOFFICIAL
Explore all skills by this author →
GitHub

molclaw-rgroup-optimization

Optimize R-group decorations on a fixed scaffold using staged_learning RL.

28|2|Updated Mar 31, 2026
Search Tags:#reinforcement-learning#drug-discovery#libinvent#r-group#reinvent4#molecule-optimization#scaffold-constrained
One-click install
npx skills add https://github.com/InternScience/MolClaw --skill molclaw-rgroup-optimization
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: molclaw-rgroup-optimization
Source: https://github.com/InternScience/MolClaw/tree/main/skills/L1_tools/molclaw-rgroup-optimization
Command: npx skills add https://github.com/InternScience/MolClaw --skill molclaw-rgroup-optimization

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill finds R-group decorations for a fixed scaffold and optimizes generated molecules toward multi-objective property targets (e.g., QED, molecular weight, LogP, TPSA) using reinforcement learning, eliminating manual trial-and-error and post-filtering inefficiencies.

Core Features & Use Cases

  • RL-guided staged_learning combines LibInvent scaffold-constrained generation with REINVENT4's staged_learning optimization to produce property-aware decorated molecules.
  • Multi-objective scoring and constraints supports weighted QED, MW, LogP, TPSA scoring, optional Tanimoto similarity to a reference, and R-group-specific constraints such as R-group MW and ring-count limits.
  • Practical outputs and requirements expects a .smi scaffold file with [*:N] attachment points, validates file existence, and saves optimized molecules to a CSV with an uppercase SMILES column for downstream analysis; suitable for scaffold-constrained lead optimization and hit refinement.

Quick Start

Run libinvent_rgroup_optimization on a scaffolds.smi file to optimize R-groups for QED and property constraints using staged_learning with your chosen weights and max_steps.

Frequently Asked Questions about molclaw-rgroup-optimization

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I optimize R-groups for a fixed molecular scaffold using reinforcement learning?▼

Scaffold-constrained R-group optimization generates new decorations for fixed molecular frameworks. It enforces property constraints such as molecular weight and LogP limits while using reinforcement learning to maximize multi-objective scores like QED across the generated molecules.

How do I set up scaffold-constrained lead optimization with REINVENT4 and LibInvent?▼

Provide a .smi scaffold file with [*:N] attachment points, then configure multi-objective scoring weights, R-group constraints, and the max_steps parameter. This inputs the required data to run staged_learning reinforcement learning for optimizing molecular decorations.

Can I enforce specific molecular weight and ring-count limits during R-group decoration?▼

Yes, you can enforce R-group-specific molecular weight and ring-count limits during optimization. The workflow applies these constraints alongside weighted multi-objective scoring for properties like QED, LogP, and TPSA to filter generated molecule decorations.

What's the best way to guide scaffold-constrained molecule generation toward a reference compound?▼

Apply optional Tanimoto similarity guidance to direct reinforcement learning optimization toward a reference compound. This ensures generated R-group decorations maintain structural similarity to your target molecule while maximizing multi-objective property scores.

Do I need a specific file format for scaffold attachment points in drug discovery optimization?▼

Yes, you need a .smi file containing your scaffold with [*:N] attachment point notation. The workflow validates this file existence before running staged_learning reinforcement learning to optimize R-group decorations for your drug-like molecules.

Why use staged_learning reinforcement learning instead of manual trial-and-error for lead optimization?▼

Staged_learning reinforcement learning eliminates manual trial-and-error and post-filtering inefficiencies in lead optimization. It directly generates property-aware R-group decorations satisfying multi-objective constraints, saving significant time compared to manual molecule filtering approaches.

Related Skills

Explore More Education & Research Skills→
InternScienceInternScience
28

molclaw-rgroup-sampling

Decorate molecular scaffolds with R-groups at [*:N] attachment points using LibInvent.

Official
Intermediate
InternScienceInternScience
28

molclaw-equiscore-docking

Rank candidate molecules by docking score against a protein target.

Official
Advanced
ldzhhxxldzhhxx
1

fine-tuning-with-trl

Fine-tune language models with TRL using SFT, DPO, PPO, or GRPO.

Community
Advanced
pradyumnasagarpradyumnasagar
6

scaffold-analysis

Analyze chemical libraries with Bemis-Murcko scaffolds, R-group decomposition, and MMP analysis.

Community
Intermediate