TPpred-ATMV

TPpred-ATMV predicts multiple types of therapeutic peptides from sequence features to support therapeutic peptide discovery and drug development.

Key Features:

• Adaptive Multi-View Tensor Learning Framework: TPpred-ATMV employs an adaptive multi-view tensor learning framework that integrates diverse sequence features to predict multiple therapeutic peptide types.
• Class and Probability Information Construction: The method constructs class labels and probability information from various peptide sequence features.
• Latent Subspace Construction: TPpred-ATMV builds a latent subspace among multi-view features to capture relationships between different peptide properties.
• Auto-Weighted Multi-View Tensor Learning Model: An auto-weighted multi-view tensor learning model leverages high correlations within multi-view features to optimize predictions.
• Performance: Experimental results showed TPpred-ATMV performs on par with or better than other state-of-the-art methods in predicting therapeutic peptides.

Scientific Applications:

• Therapeutic peptide research and drug development: Supports identification of candidate therapeutic peptides across eight types by integrating multi-view sequence features.

Methodology:

Construct class and probability information from sequence features, build a latent subspace among multi-view features, and apply an auto-weighted multi-view tensor learning model that leverages high correlations within the data.