A comprehensive chemical analysis tool that transforms a single SMILES string into a unified pharmacology and ADMET profile.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install smiles-profiling
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install smiles-profiling using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is SMILES Profiling?
The SMILES Profiling skill is a specialized tool designed for researchers and developers working with chemical data within the Openclaw Skills framework. It automates the complex task of cross-referencing multiple bioinformatics and cheminformatics databases to build a complete picture of a molecule's properties. By inputting a single SMILES string, the agent aggregates critical data from five major industry sources, ensuring a high-fidelity overview of chemical identity and biological potential.
This skill is particularly valuable for accelerating the early stages of drug discovery and lead optimization. It streamlines workflows by handling the heavy lifting of data retrieval from SwissTargetPrediction, PubChem, ADMETlab 3.0, ChEMBL, and PK-Smart. By utilizing Openclaw Skills, users can bridge the gap between simple chemical notation and actionable pharmacological insights without manual database navigation.
SMILES Profiling Use Cases
- Generating instant pharmacology summaries for newly identified chemical hits.
- Automating ADMET and pharmacokinetic safety screening during lead optimization phases.
- Cross-referencing experimental compounds against known drugs in ChEMBL and PubChem using Openclaw Skills.
- Identifying potential off-target effects and protein binding probabilities via SwissTargetPrediction.
How SMILES Profiling Works
- The user provides a SMILES string as the primary input for the profiling process.
- The agent cleans and normalizes the input, stripping common salts or counterions to isolate the active moiety.
- SwissTargetPrediction is queried to predict the top protein targets and their associated probabilities.
- PubChem provides the chemical identity, including CID, IUPAC name, and essential physicochemical properties.
- ADMETlab 3.0 evaluates the structure for toxicity flags and overall drug-likeness metrics.
- ChEMBL searches for historical similarity hits, clinical phases, and known mechanisms of action.
- PK-Smart calculates pharmacokinetic estimates such as clearance, half-life, and volume of distribution.
- The agent merges these disparate data points into a single, concise Markdown report for the user.
SMILES Profiling Setup
To integrate SMILES Profiling into your Openclaw Skills workflow, ensure your environment is configured to execute Python scripts. Use the following commands to run a profile:
# Run the full profiling report for a specific SMILES string
python scripts/run_smiles_profile.py "CC(=O)OC1=CC=CC=C1C(=O)O"
# Run benchmark or regression tests using a TSV input file
python scripts/run_smiles_smoke.py --cases <path_to_tsv>
SMILES Profiling Data Schema & Taxonomy
The skill organizes chemical data into a structured taxonomy for clear interpretation:
| Data Source | Key Metadata Captured |
|---|---|
| SwissTargetPrediction | Top 5 targets, species, and probability scores |
| PubChem | CID, Synonyms, MW, XLogP, TPSA, and HBD/HBA |
| ADMETlab 3.0 | ADMET flags, toxicity sections, and drug-likeness |
| ChEMBL | Similarity hits (threshold 70-90), max phase, and target names |
| PK-Smart | Human/Animal estimates for VDss, CL, t½, fu, and MRT |
All data is processed through Openclaw Skills to ensure a uniform output format regardless of the source database's original structure.
SMILES Profiling Advanced Features
- Intelligent Similarity Fallback: Automatically lowers ChEMBL similarity thresholds (90% to 70%) to find relevant analogs when exact matches are unavailable.
- Graceful Degradation: The workflow continues profiling even if one or more external web tools are unresponsive.
- Batch Processing: Support for regression runs and benchmarks using TSV-based case files.
- Automated Normalization: Built-in handling for salts and mixtures to ensure data accuracy.
- Evidence Prioritization: Logic-driven reporting that favors direct database evidence over predictive models within the Openclaw Skills ecosystem.
SKILL.md
Loading
Related Openclaw Skills
Wyndham Hotel Group Guide
A comprehensive digital guide for navigating the Wyndham Hotels & Resorts ecosystem, from luxury brands to loyalty rewards.
system_cmd
A secure and robust Windows system command execution skill featuring four-level security and full multi-language output support.
Science Agent Compare
A specialized tool for comparing AI science agents side-by-side while resolving same-name project collisions within the ecosystem.
Science Agent Recommend
A comprehensive discovery engine that matches researchers with specialized AI agents and Openclaw Skills across 10 scientific categories.
Vrbo Vacation Rental Guide
A comprehensive integration guide for navigating the Vrbo global vacation rental marketplace to book whole-property accommodations.
Trivago Hotel Comparison & Search
A comprehensive hotel meta-search integration for comparing prices across 300+ booking sites and 190 countries.
Featured*
