RNCE

RNCE integrates multi-modal pharmacogenomic and multi-omics drug data using reciprocal nearest neighbor relationships and contextual encoding to detect drug communities and infer drug targets and mechanisms of action (MoA).

Key Features:

• Network Integration with Reciprocal Neighbors: Leverages reciprocal nearest neighbor relationships to preserve connectivity and context within drug networks for more accurate drug similarity and community structure representation.
• Contextual Information Encoding: Encodes contextual information into network representations to capture complex interactions and dependencies among drugs.
• Tailor-made Clustering Algorithm: Applies a specialized clustering algorithm designed for drug community detection based on RNCE-integrated network features.
• Handling Sparse and Imbalanced Community Sizes: Accounts for sparse and imbalanced community size structures in drug networks to improve detection of both small and large communities.
• Multi-modal / Multi-omics Integration: Integrates multi-modal and multi-omics drug information from pharmacogenomic datasets to support comprehensive network analysis.
• Mining of Drug Targets and MoA: Facilitates identification of potential drug targets and mechanisms of action (MoA) from integrated pharmacogenomic data.

Scientific Applications:

• Drug Target and MoA Identification: Identifies potential drug targets and mechanisms of action (MoA) from integrated pharmacogenomic and multi-omics data.
• Anticancer Drug Discovery: Supports discovery and characterization of novel anticancer drugs by revealing drug communities and associated MoA.
• Network Similarity and Community Detection Across Databases: Improves network similarity assessments and community detection tasks across drug databases, demonstrated on two drug databases.
• Multi-modal/Multi-omics Integrative Studies in Pharmacogenomics: Enhances integrative analyses in pharmacogenomics to uncover nuanced drug interactions and dependencies.

Methodology:

RNCE employs spectral clustering in conjunction with reciprocal nearest neighbor relationships and contextual encoding.