AmIActive (AIA): A Large-scale QSAR Based Target Fishing and Polypharmacology Predictive Web Tool

Area of Science:

• Computational chemistry
• Cheminformatics
• Pharmacology

Background:

• Quantitative Structure-Activity Relationship (QSAR) models are crucial for predicting drug efficacy.
• Access to comprehensive and validated QSAR models facilitates drug discovery and development.
• Existing tools may lack broad coverage or user-friendly interfaces for biological activity prediction.

Purpose of the Study:

• To introduce AmIActive (AIA), a novel web-based QSAR tool.
• To provide a platform for biological activity prediction and target fishing.
• To support researchers in identifying potential drug candidates and their targets.

Main Methods:

• Development of 3,239 QSAR models using Random Forest classifiers.
• Training models with circular fingerprint descriptors and ChEMBL database activity data.
• Implementation of clearly defined applicability domains for model reliability.

Main Results:

• The AIA system covers 2,277 distinct targets, including proteins, complexes, families, cell lines, organisms, and tissues.
• Models are trained on extensive biological activity data, ensuring broad applicability.
• The tool provides predictive information, target details, validation metrics, and activity thresholds.

Conclusions:

• AmIActive (AIA) offers a user-friendly, free resource for QSAR-based biological activity prediction.
• The comprehensive model set and intuitive interface streamline target fishing and drug discovery efforts.
• AIA facilitates efficient exploration of chemical space for potential therapeutic agents.