A molecular descriptor dataset of 33,715 small molecules based on operator-based descriptor framework

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Traditional molecular descriptors struggle to capture complex molecular geometry.
• High-throughput screening and drug delivery optimization require advanced molecular characterization.

Purpose of the Study:

• To present a large-scale molecular descriptor dataset for computational drug discovery.
• To provide advanced descriptors encoding spatial features for enhanced predictive modeling.

Main Methods:

• A dataset of 33,715 small molecules was curated.
• An operator-based matrix encoding strategy captured spatial features like bond and dihedral angles.
• A PHP-based pipeline ensured standardized identifier mapping.

Main Results:

• A comprehensive dataset with 33,715 small molecules and advanced descriptors was generated.
• The dataset encodes high-dimensional structural data, including spatial features.
• The data is available in a 215.9 MB tab-separated text file.

Conclusions:

• This dataset enhances the predictive accuracy of inhibitor-target interaction models.
• It offers a robust framework for investigating molecular mechanisms in drug efficacy.
• The open-access resource supports computational drug discovery pipelines.