JDC

JDC predicts essential proteins by integrating protein-protein interaction (PPI) networks with binarized gene expression profiles and combining degree centrality with Jaccard similarity to improve essential protein identification accuracy.

Key Features:

• Integration of Biological Data: Integrates PPI network data and gene expression profiles to predict protein essentiality while mitigating variability in expression data by comparing active and inactive states within PPI clusters.
• Dynamic Threshold Binarization: Applies a dynamic threshold to binarize gene expression into active/inactive states to filter noise and fluctuations.
• Centrality and Similarity Scoring: Computes a per-protein score by combining degree centrality with the Jaccard similarity index of expression-state overlap.
• Cluster-focused Analysis: Emphasizes similarity between active and inactive states within densely connected clusters of the PPI network.
• Benchmarking and Evaluation: Benchmarks performance across four organisms using ROC analysis, modular analysis, jackknife analysis, overlapping analysis, top analysis, and accuracy analysis.
• Comparative Performance: Compared against DC, IC, EC, SC, BC, CC, NC, PeC, and WDC and reported to outperform these methods, and also compared to NF-PIN and TS-PIN using the same input data.

Scientific Applications:

• Essential Protein Identification: Predicts essential proteins with improved reliability by integrating PPI and expression-state information.
• Protein Function and Interaction Analysis: Provides insights into protein function and interactions through combined network and expression-state similarity analysis.
• Drug Discovery: Informs target selection by identifying proteins likely to be essential in biological systems.
• Disease Modeling: Supports modeling of disease-relevant essential proteins by integrating dynamic expression states with PPI data.
• Functional Genomics: Aids functional genomics studies by prioritizing essential proteins based on network centrality and expression overlap.

Methodology:

Binarizes gene expression using a dynamic threshold to define active/inactive states, computes degree centrality on PPI networks and Jaccard similarity of expression-state overlap, combines these measures into a per-protein score, and evaluates performance via ROC analysis, modular analysis, jackknife analysis, overlapping analysis, top analysis, and accuracy analysis across four organisms while comparing to DC, IC, EC, SC, BC, CC, NC, PeC, WDC, NF-PIN, and TS-PIN.