Kitchenware

Utility tools to load and process biomolecular data. This package provides a foundation for deep learning applications in structural biology, offering efficient tools for handling molecular structures and their analysis.

Features

Data Processing

• Load and save molecular structures (PDB/CIF formats)
• Structure encoding and processing
• Biomolecular data type handling

Geometric Analysis

• Distance matrix computation
• Connectivity extraction
• Neighborhood analysis
• Structure superposition
• Contact detection

Structure Analysis

• Secondary structure analysis
• RMSD calculations
• lDDT scoring
• Angle and dihedral measurements

Utilities

• Standard amino acid encodings
• Template handling
• Chain and residue operations
• Density map operations

Installation

pip install kitchenware

Quick Start

import kitchenware as kw

# Load a structure (PDB or CIF format)
structure = kw.load_structure("input_structure.pdb")

# Encode structure and convert to a collection of tensors (PyTorch)
data = kw.encode_structure(structure)

# Convert back to a structure
structure = kw.data_to_structure(data)

# Write PDB files
kw.save_pdb(data, "output_structure.pdb")

Dependencies

• gemmi
• mrcfile
• numpy
• pandas
• torch
• rdkit
• scipy
• scikit-learn

For a list of dependencies see pyproject.toml

Applications

This toolkit serves as the foundation for several deep learning methods in structural biology:

• PeSTo: parameter-free geometric deep learning for accurate prediction of protein binding interfaces
• CARBonAra: Context-aware geometric deep learning for protein sequence design

Contributing

We welcome contributions! Please feel free to submit a Pull Request.

License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Reference

This toolkit was developed as the foundation for both PeSTo and CARBonAra. If you use this toolkit in your research, please cite one of the following publications.

PeSTo

Krapp, L.F., Abriata, L.A., Cortés Rodriguez, F. et al. PeSTo: parameter-free geometric deep learning for accurate prediction of protein binding interfaces. Nat Commun 14, 2175 (2023). https://doi.org/10.1038/s41467-023-37701-8

CARBonAra

Krapp, L.F., Meireles, F.A., Abriata, L.A. et al. Context-aware geometric deep learning for protein sequence design. Nat Commun 15, 6273 (2024). https://doi.org/10.1038/s41467-024-50571-y