What format do the molecules need to be in?

The skill primarily works with SMILES strings and RDKit molecule objects, often leveraging the Datamol library for seamless conversion and handling.

What medicinal chemistry rules can this skill apply?

It supports Lipinski's Rule of Five, Veber's rules, Oprea, CNS-likeness, Lead-like filters, and many more standard pharmaceutical guidelines.

Does it detect toxic or reactive groups?

Yes, it includes comprehensive structural alert filters like PAINS (Pan-Assay Interference Compounds), NIBR filters, and Lilly Demerits to identify problematic functional groups.

Can I use this for high-throughput screening?

Yes, the skill supports parallelized processing using multiple CPU cores, making it efficient for triaging libraries with thousands of molecules.

Can I define custom filtering criteria?

Yes, you can use the Medchem Query Language or custom constraints to define specific ranges for molecular weight, LogP, TPSA, and other chemical properties.

Scientific Medicinal Chemistry

Name: Scientific Medicinal Chemistry
Author: Zehong-Wang

byZehong-Wang

0•

Data Science & ML

Applies medicinal chemistry rules and structural filters to prioritize and triage molecular compound libraries for drug discovery.

Medchem is a specialized scientific skill for Claude that integrates the medchem Python library to automate molecular filtering and prioritization in drug discovery workflows. It allows researchers to apply well-established drug-likeness rules such as Lipinski's Rule of Five and Veber's rules, detect problematic structural alerts like PAINS or Lilly Demerits, and calculate molecular complexity metrics. By combining standard medicinal chemistry principles with programmable constraints and a specialized query language, this skill streamlines the triage process for large-scale compound libraries during lead optimization and virtual screening.

Key Features

01Calculate molecular complexity metrics to estimate synthetic accessibility

02Detect specific chemical functional groups and reactive patterns

03Filter molecules using NIBR, PAINS, and Lilly Demerits structural alerts

040 GitHub stars

05Apply standard drug-likeness rules including Lipinski, Veber, and CNS-likeness

06Execute complex filtering workflows via the Medchem Query Language

Use Cases

01Lead optimization by assessing compound quality and medicinal properties

02Identification and removal of reactive or problematic functional groups from datasets

03Initial triage and cleaning of large-scale virtual compound libraries

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zehong-wang/kosmos medchem

For use in Claude.ai and ChatGPT

Download Skill