bit-collision-free-MF 0.9.0

Generate molecular fingerprints with guaranteed collision‑free bits.

bit_collision_free_MF

A Python package for generating molecular fingerprints without bit collisions.

Description

bit_collision_free_MF generates Morgan fingerprints while eliminating bit collisions, which can significantly improve the accuracy and reliability of molecular fingerprints in cheminformatics applications. The package automatically determines the optimal fingerprint length to ensure that each structural feature maps to a unique bit in the fingerprint.

Installation

Requirements

• Python 3.9 or higher
• numpy
• pandas
• rdkit

Simple Installation

pip install -U bit_collision_free_MF

This will automatically install all dependencies, including RDKit.

Manual Installation

# Install dependencies
pip install numpy pandas rdkit

# Install the package
pip install -U bit_collision_free_MF  

For development installation:

# Clone the repository
git clone https://github.com/Shifa-Zhong/bit_collision_free_MF.git
cd bit_collision_free_MF

# Install in development mode
pip install -e .

Features

• Automatically determines the optimal fingerprint length to avoid bit collisions
• Supports custom fingerprint radius
• Option to remove zero-value columns
• Easy CSV export with customizable headers
• Seamless integration with pandas and NumPy

Usage

Basic Usage

from bit_collision_free_MF import generate_fingerprints, save_fingerprints
import pandas as pd

# Load your data
data = pd.read_csv('your_molecules.csv')

# Generate fingerprints
fingerprints, fp_generator = generate_fingerprints(
    data, 
    smiles_column='smiles',
    radius=1,
    remove_zero_columns=True
)

# Save fingerprints to CSV
save_fingerprints(
    fingerprints,
    fp_generator,
    output_path='path/to/output.csv',
    include_header=True
)

Using the CollisionFreeMorganFP Class Directly

from bit_collision_free_MF import CollisionFreeMorganFP
import pandas as pd

# Load your data
data = pd.read_csv('your_molecules.csv')
smiles_list = data['smiles'].tolist()

# Create and fit the fingerprint generator
fp_generator = CollisionFreeMorganFP(radius=1)
fp_generator.fit(smiles_list)

# Generate fingerprints
fingerprints = fp_generator.transform(smiles_list, remove_zero_columns=True)

# Get feature names
feature_names = fp_generator.get_feature_names()

# Create a DataFrame with the fingerprints
result_df = pd.DataFrame(fingerprints, columns=feature_names)

# Save to CSV
result_df.to_csv('fingerprints.csv', index=False)

License

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

Contact

For academic inquiries or collaboration, please contact:

• Shifa Zhong ([e-mail])
• Jibai Li ([e-mail])