LP-HCLUS

LP-HCLUS predicts novel associations between non-coding RNAs (ncRNAs) and human diseases by extracting multi-type hierarchical clusters from heterogeneous biological networks.

Key Features:

• Multi-Type Hierarchical Clustering: Employs multi-type hierarchical clustering to analyze heterogeneous networks and form overlapping, hierarchically organized clusters of microRNAs, long non-coding RNAs (lncRNAs), target genes, and diseases.
• Heterogeneous Network Analysis: Processes complex networks composed of multiple object types and relationship types to capture diverse interactions and roles of ncRNAs at different levels of granularity.
• Predictive Capability: Extracts multi-type clusters from heterogeneous networks to predict potentially unknown ncRNA-disease associations and nominate candidate therapeutic targets.
• Robust Evaluation Metrics: Evaluates performance using quantitative metrics including True Positive Rate at k (TPR@k), Areas Under the TPR@k, Receiver Operating Characteristic (ROC) curves, Precision-Recall curves, and qualitative literature consultation.

Scientific Applications:

• ncRNA-disease association discovery: Predicts novel associations between microRNAs, lncRNAs, target genes, and human diseases to support investigation of disease mechanisms.
• Therapeutic target identification: Identifies candidate ncRNAs and target genes for downstream experimental validation and potential therapeutic development.
• Predictive studies of ncRNA function: Enables comparative and multi-granularity studies of ncRNA roles across heterogeneous biological data.

Methodology:

Applies multi-type hierarchical clustering to heterogeneous networks of microRNAs, long non-coding RNAs (lncRNAs), diseases, and genes to form overlapping hierarchical clusters and extract multi-type clusters for prediction, with evaluation via TPR@k, Areas Under the TPR@k, ROC and Precision-Recall curves, and literature consultation.