DeepScreening

DeepScreening performs deep learning-based virtual screening and target-focused de novo library generation to identify chemical probes and drug candidates against biological targets.

Key Features:

• Deep learning algorithms: Implements advanced deep learning models for virtual screening and compound scoring.
• Model training: Trains predictive models using supplied datasets to prioritize compounds for biological targets.
• Data flexibility: Accepts public datasets or user-provided datasets for model development and validation.
• Virtual screening: Applies trained models to screen existing chemical libraries and prioritize potential drugs or probes against specified targets.
• De novo library generation: Constructs target-focused de novo libraries to expand searchable chemical space for screening.
• Integrated workflow: Integrates model training, de novo library generation, and virtual screening into a single computational pipeline.

Scientific Applications:

• Drug discovery: Prioritizes small-molecule drug candidates for downstream experimental validation.
• Chemical probe identification: Identifies candidate chemical probes for target validation in biological studies.
• Target-focused library design: Generates and screens target-focused compound libraries for chemical biology applications.
• High-throughput in silico screening: Enables large-scale virtual screening of chemical libraries to accelerate hit identification.

Methodology:

Train deep learning models on public or user-provided datasets, apply trained models to virtual screening of existing chemical libraries, and generate target-focused de novo libraries for subsequent screening.