🧪 Curated Compound Analysis

A comprehensive pipeline for compound data curation and target prediction analysis using PubChem API and SwissTargetPrediction.

📋 Overview

This project provides a two-stage pipeline for compound analysis:

1. Data Curation Stage: Extract CID and SMILES from compound names using PubChem API
2. Target Prediction Stage: Predict molecular targets using SwissTargetPrediction web interface

🚀 Features

• Automated PubChem Integration: Fetch compound IDs and SMILES strings from compound names
• Robust Name Handling: Supports Greek letters (α, β, γ) and name variations
• Quality Control: Validates and cleans data before analysis
• Comprehensive Logging: Detailed logs for troubleshooting and monitoring
• Error Handling: Retry mechanisms and fallback options
• Semi-Automated SwissTarget: Browser automation for target prediction

📁 Project Structure

curated_compound_analysis/
├── notebooks/
│   └── curated_compound_swiss_target.ipynb
├── scripts/
│   └── swisstarget.py
├── data/
│   └── compound_sample_data.csv
├── requirements.txt
├── README.md
└── docs/
    ├── installation.md
    ├── usage.md
    └── troubleshooting.md

🛠️ Installation

Prerequisites

• Python 3.7 or higher
• Google Chrome browser (for SwissTarget automation)
• ChromeDriver (see ChromeDriver Setup)

Step 1: Clone Repository

git clone https://github.com/yourusername/curated_compound_analysis.git
cd curated_compound_analysis

Step 2: Install Python Dependencies

pip install -r requirements.txt

Step 3: ChromeDriver Setup

Option A: Automatic Installation (Recommended)

The script will automatically download ChromeDriver using webdriver-manager.

Option B: Manual Installation

1. Check your Chrome version: chrome://version/
2. Download matching ChromeDriver from ChromeDriver Downloads
3. Extract and place chromedriver.exe in the project folder

Option C: System PATH Installation

1. Download ChromeDriver
2. Add to system PATH
3. Verify: chromedriver --version

📊 Usage

Stage 1: Data Curation (Jupyter Notebook)

1. Prepare Input Data

   • Create CSV file with compound names in 'Name' column
   • Example: compound_data.csv

2. Run Curation Notebook

   jupyter notebook notebooks/curated_compound_swiss_target.ipynb

3. Execute All Cells

   • Cell 1: Setup and configuration
   • Cell 2: Input data validation
   • Cell 3: PubChem API functions
   • Cell 4: Data processing and CID/SMILES extraction
   • Cell 5: Quality control and cleaning
   • Cell 6: Download results (optional)

4. Output Files

   • compounds_with_cid_smiles.csv: All processing results
   • data_compounds_final_full.csv: Clean dataset for SwissTarget
   • curation_log.txt: Detailed processing log

Stage 2: SwissTarget Prediction (Python Script)

1. Prepare Input

   • Use the clean dataset from Stage 1
   • Rename to data_compounds_final_full.csv

2. Run SwissTarget Script

   python scripts/swisstarget.py

3. Manual Interaction Required

   • Script opens Chrome browser for each compound
   • Download results manually (CSV/Excel/PDF)
   • Take screenshots as needed
   • Close browser to continue to next compound

4. Output Organization

   • Results saved in timestamped folders
   • Individual folders for each compound
   • Processing logs included

📋 Input Data Format

Your input CSV should have this structure:

"Name","Formula","Calc. MW" => must have this.

"Name","Formula","Annot. DeltaMass [ppm]","Calc. MW","RT [min]","Area (Max.)"
"Rhynchophylline","C22 H28 N2 O4","-3.08","384.20372","6.81","23288814985.171"
"Mitragynine","C23 H30 N2 O4","-3.35","398.21922","8.635","14364893452.408"
"deacetylvindoline","C23 H30 N2 O5","-2.8","414.21431","7.443","5066565186.2213"
"L-α-PALMITIN","C19 H38 O4","-3.15","330.27597","14.341","925221458.83439"
"1-Stearoylglycerol","C21 H42 O4","-3.24","358.30715","15.378","693539239.29014"

Additional columns are preserved in the output.

🔧 Configuration

PubChem API Settings

• Request delay: 0.5 seconds (respects rate limits)
• Timeout: 20 seconds per request
• Retry mechanism: Name variations for better success rates

SwissTarget Settings

• Browser timeout: 60 seconds for page load
• Results timeout: 240 seconds for predictions
• Retry attempts: 3 per compound
• Target organism: Homo sapiens (default)

📈 Success Rates

Typical success rates based on compound types:

• Well-known compounds: 85-95%
• Natural products: 70-85%
• Synthetic compounds: 60-80%
• Rare/proprietary compounds: 40-60%

🐛 Troubleshooting

Common Issues

1. ChromeDriver Version Mismatch

   Solution: Update ChromeDriver to match your Chrome version
   

2. PubChem API Timeout

   Solution: Check internet connection, compound names for typos
   

3. SwissTarget Page Load Issues

   Solution: Retry with stable internet, check if website is accessible
   

4. Missing Dependencies

   pip install --upgrade -r requirements.txt

Debug Information

Enable detailed logging by checking the generated log files:

• curation_log.txt: PubChem processing logs
• process_log.txt: SwissTarget automation logs

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• PubChem for compound data API
• SwissTargetPrediction for target prediction service
• Selenium for web automation
• ChromeDriver for browser automation

📞 Support

If you encounter issues or have questions:

1. Check the troubleshooting guide
2. Search existing issues
3. Create a new issue with detailed information

📊 Citation

If you use this pipeline in your research, please cite:

@software{curated_compound_analysis,
  title={Compound Analysis Pipeline: Automated PubChem and SwissTarget Integration},
  author={Nandatama, Engki},
  year={2024},
  url={https://github.com/yourusername/curated_compound_analysis}
}

---

⭐ Star this repository if you find it helpful! ⭐