MccDTI

MccDTI predicts drug-target interactions to support drug repositioning and discovery by integrating heterogeneous drug and target data using multiview network embedding and a matrix completion scheme.

Key Features:

• Multiview Network Embedding: Integrates heterogeneous information from multiple sources related to drugs and targets, capturing consistent and complementary information across different network views.
• High-Quality Low-Dimensional Representations: Learns compact embeddings of drugs and targets that preserve essential relationships within and between various data sources.
• Matrix Completion Scheme: Applies matrix completion on learned representations to predict and fill missing drug-target interaction entries.

Scientific Applications:

• Drug-target interaction prediction: Prioritizes potential DTIs for experimental follow-up.
• Drug repositioning and discovery: Supports identification of novel therapeutic opportunities and candidate repurposing hypotheses.
• Benchmarking: Demonstrated superior performance compared to four state-of-the-art methods on two datasets.
• Literature verification: Provides predictions that were validated through literature verification.

Methodology:

Integrates heterogeneous drug and target data via multiview network embedding to learn low-dimensional representations and applies a matrix completion scheme on those representations to predict drug-target interactions.