visChem 0.1.2

A tool for chemical space visualization and clustering.

visChem v0.1.2

Steve Niu[^1]

visChem is a Python module designed for visualizing chemical space using UMAP and clustering techniques. This tool facilitates the transformation of chemical structures into a 2D space and visualizes clusters of similar compounds.

Installation

Install using pypi

pip install visChem==0.1.2

Install using conda

conda install -c stevexniu vischem=0.1.2

Note: Some packages like rdkit might require special installation instructions. Please refer to their official documentation for details.

Usage

1. See example.py for details.
2. Run python example.py.

The example.py script will execute the main function which:

• Load example 1000 SMILES from RDKit NCI/first_5K.smi.
• Converts SMILES to molecular representations.
• Generates molecular fingerprints.
• Computes Tanimoto similarity.
• Reduces dimensionality using UMAP.
• Performs clustering.
• Visualizes the chemical structures in the clustered chemical space below.

Refer to example.py for details and any potential modifications you might want to make for your specific use case.

[^1]: Safety Assessment, [e-mail]