DRIMC

DRIMC applies Bayesian inductive matrix completion to predict novel drug-disease associations for drug repositioning by integrating drug and disease similarity matrices.

Key Features:

• Bayesian Inductive Matrix Completion: Implements Bayesian inductive matrix completion to model association probabilities and handle incomplete data.
• Integration of Multiple Data Sources: Integrates four drug data sources into a drug similarity matrix and two disease data sources into a disease similarity matrix.
• Latent Space Mapping: Describes each drug and disease by similarity values relative to nearest neighbors and projects these features into a shared latent space.
• Confidence Weighting: Assigns higher confidence to manually verified known drug-disease associations than to unknown pairs during modeling.

Scientific Applications:

• Drug Repositioning: Predicts new indications for existing drugs by identifying likely drug-disease associations.
• Benchmark Evaluation: Has been evaluated on three benchmark datasets and compared with six state-of-the-art approaches.

Methodology:

Embed drug and disease data into similarity matrices; describe entity features via proximity to nearest neighbors and project them into a shared latent space; apply Bayesian inductive matrix completion to model association probabilities with confidence weighting for known associations.