About CLaSP
CLaSP (Contrastive Learning-Guided Latent Scoring Platform) is an advanced druggability prediction platform developed by the Laboratory of Molecular Modeling and Drug Design (LMMD) at East China University of Science and Technology.
Related Publication
Platform Overview
CLaSP calculates drug-like chemical space and provides scores using the CLVAE method based on 22 distinct physicochemical properties, facilitating drug development.
The platform leverages advanced deep learning algorithms and molecular modeling techniques to accelerate the drug discovery process and improve the success rate of pharmaceutical research.
Model Architecture
Our model architecture employs a sophisticated multi-layered approach that integrates various computational methodologies. The system processes both structural and chemical information to generate accurate predictions.
This algorithm integrates ADMET properties, physicochemical characteristics, and additional synthesizability metrics to provide a comprehensive drug-likeness evaluation.
Model Performance
CLaSP demonstrates exceptional performance across multiple evaluation metrics. Our comprehensive validation studies show superior accuracy, precision, and recall compared to GIN, GCN, GAT, ChemBERTa, and DNN methods.
The model has been rigorously tested on diverse datasets and consistently delivers reliable predictions for various therapeutic targets and compound libraries.
