IDSSIM

IDSSIM calculates functional similarity between long non-coding RNAs (lncRNAs) by integrating improved disease semantic similarity with lncRNA functional data to enhance lncRNA–disease association prediction.

Key Features:

• Improved disease semantic similarity: Incorporates an information content contribution factor into semantic value calculations for diseases.
• Disease DAG and specificity handling: Accounts for hierarchical structures in disease directed acyclic graphs (DAGs) and the specificities of individual diseases.
• Integration with lncRNA data: Combines disease semantic similarities with lncRNA functional similarity data for downstream analysis.
• Semantic matrix refinement: Refines semantic similarity matrices used for lncRNA functional similarity calculations.
• Association prediction framework: Applies the refined similarity matrices within the WKNKN association prediction framework.
• Comparative evaluation: Evaluated against LNCSIM1, LNCSIM2, and ILNCSIM using disease semantic similarity matrices, lncRNA functional similarity matrices, and human lncRNA–disease association data.
• Datasets and metrics: Uses lncRNADisease and MNDR databases for association data and assesses performance with Receiver Operating Characteristic (ROC) curves and Area Under Curve (AUC) values.

Scientific Applications:

• lncRNA function prediction: Predicts functions of long non-coding RNAs based on integrated similarity measures.
• lncRNA–disease association identification: Identifies potential lncRNA–disease associations for hypothesis generation.
• Pre-screening candidates for experiments: Provides candidate lncRNAs for biological validation supported by database and literature corroboration.
• Disease-specific case studies: Applied to case studies including breast cancer and adenocarcinoma to validate predicted associations.

Methodology:

IDSSIM augments disease semantic value calculations with an information content contribution factor within disease DAGs, integrates the resulting disease semantic similarities with lncRNA functional similarity matrices, refines those semantic matrices, and applies them in the WKNKN association prediction framework; evaluation used ROC curves and AUC against LNCSIM1, LNCSIM2, and ILNCSIM with data from lncRNADisease and MNDR.