adme-prediction

Predict ADME properties from SMILES strings using Morgan fingerprints and Random Forest models.

52|11|Updated Mar 10, 2026

Search Tags:#drug-discovery #fingerprints #smiles #rdkit #adme #admet #random-forest

Ready to run

npx skills add https://github.com/InternScience/ChemClaw --skill adme-prediction

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: adme-prediction
Source: https://github.com/InternScience/ChemClaw/tree/main/skills/adme-prediction
Command: npx skills add https://github.com/InternScience/ChemClaw --skill adme-prediction

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires rdkit, scikit-learn, numpy, pandas, tdc, and includes scripts (resource) components.

What problem does it solve?

Predict ADME properties of small molecules from SMILES strings to accelerate early-stage drug discovery by rapidly evaluating pharmacokinetic traits.

Core Features & Use Cases

Morgan fingerprint-based predictions using Random Forest models for six key ADME properties: Caco-2 permeability, PAMPA, HIA, Pgp inhibition, Bioavailability, and Lipophilicity.
CLI and Python API support for single and batch predictions, enabling integration into existing pipelines and notebooks.
Real-world scenario: screen a library of SMILES to filter candidates with favorable permeability and oral bioavailability before synthesis.

Quick Start

Use the adme-prediction skill to predict ADME properties for a given SMILES string.

Frequently Asked Questions about adme-prediction

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

FAQPage Schema

How do I predict ADME properties from SMILES strings?▼

You can predict ADME properties by inputting SMILES strings into Python or CLI interfaces. The tool converts them into Morgan fingerprints and uses Random Forest models to evaluate Caco-2, PAMPA, HIA, Pgp, bioavailability, and lipophilicity.

Can I run batch ADME predictions for a library of molecules?▼

Yes, batch ADME predictions are supported via both CLI and Python API. You can process a library of SMILES strings to filter drug discovery candidates with favorable permeability and oral bioavailability before synthesis.

What machine learning algorithm does this ADME prediction tool use?▼

This ADME prediction tool uses Random Forest machine learning models. It transforms molecular structures into Morgan fingerprints using RDKit and scikit-learn to evaluate pharmacokinetic traits across six key properties.

Do I need RDKit and scikit-learn to calculate molecular fingerprints for ADME?▼

Yes, RDKit and scikit-learn are required dependencies to calculate Morgan fingerprints for ADME prediction. The implementation also relies on numpy, pandas, and tdc to load models and output JSON or human-readable tables.

What is the best way to screen small molecules for oral bioavailability and permeability?▼

The best way to screen small molecules is using Random Forest models on Morgan fingerprints. This approach rapidly evaluates pharmacokinetic traits like Caco-2 permeability, PAMPA, HIA, and oral bioavailability from SMILES strings.

How does Morgan fingerprint-based prediction work for drug discovery?▼

Morgan fingerprint-based prediction works by converting SMILES strings into numerical molecular representations. Random Forest models then map these fingerprints to predict six key ADME properties, accelerating early-stage drug discovery evaluation.