Skip to content

Tutorials

Welcome to the THEMAP tutorials! These step-by-step guides will help you master task hardness estimation and distance computation for molecular activity prediction.

Tutorial Overview

🚀 Beginner Tutorials

  1. Getting Started - Basic installation and first steps

Interactive Notebooks

All tutorials are available as interactive Jupyter notebooks that you can run locally:

# Install notebook dependencies
pip install -e ".[dev]"

# Launch Jupyter Lab
jupyter lab docs/tutorials/

# Download tutorial data (if not included)
from themap.utils import download_tutorial_data

download_tutorial_data("tutorials/data/")

Sample Datasets

  • ChEMBL Bioactivity Data: 10 training + 3 test tasks
  • Protein Sequences: Target protein sequences for each task
  • Molecular Embeddings: Pre-computed molecular features
  • Metadata: Assay descriptions and experimental conditions

Prerequisites

Python Knowledge

  • Basic Python programming
  • Familiarity with NumPy and Pandas
  • Optional: Jupyter notebook experience

Domain Knowledge

  • Basic understanding of molecular representations (SMILES, etc.)
  • Familiarity with machine learning concepts
  • Optional: Knowledge of protein sequences and drug discovery

Computational Resources

Most tutorials can run on: - CPU: Standard laptop/desktop (8GB+ RAM recommended) - GPU: Optional, speeds up OTDD computations - Storage: ~1GB for tutorial data and caches

Getting Help

Tutorial Support

If you encounter issues with tutorials:

  1. Check Prerequisites: Ensure all dependencies are installed
  2. Verify Data: Confirm tutorial data is properly downloaded
  3. Read Error Messages: THEMAP provides detailed error information
  4. Ask Questions: Open an issue on GitHub

Common Issues

# Installation issues
pip install -e ".[all]"  # Install all dependencies

# Memory issues
# Use smaller datasets or batch processing

# GPU issues
import torch
print(f"CUDA available: {torch.cuda.is_available()}")

Community

  • GitHub Discussions: Share experiences and ask questions
  • Issues: Report bugs or request features
  • Contributions: Submit improvements to tutorials

Contributing to Tutorials

We welcome contributions! To add or improve tutorials:

  1. Fork the repository
  2. Create tutorial content in Markdown and/or Jupyter format
  3. Test thoroughly with different environments
  4. Submit pull request with clear description

Tutorial Guidelines

  • Clear objectives: State what readers will learn
  • Step-by-step: Break complex tasks into manageable steps
  • Code examples: Include runnable code snippets
  • Error handling: Show how to handle common issues
  • Real data: Use realistic examples when possible

What's Next?

Ready to get started? Here are recommended next steps:

Happy learning! 🧪🔬📊