DeCoaD

DeCoaD quantifies similarities between genetic diseases by analyzing protein interaction networks and disease–protein associations to identify shared molecular profiles.

Key Features:

• Pair-wise Disease Similarity Scores: Computes similarity scores between pairs of genetic diseases by modeling a network of disease and protein nodes connected by protein–protein interactions and disease–protein associations.
• Random Walk Algorithm: Performs random walks initiated and terminated at specific disease nodes and records total visits to each protein node as node weights.
• Cosine Similarity Measure: Calculates cosine similarity between vectors whose elements are protein-visit weights associated with diseases.
• Disease Family Graphical Representation: Produces graphical representations of disease families and their relationships derived from similarity scores and clusters.
• Probabilistic Clustering Algorithm: Applies probabilistic clustering to group diseases based on similarity scores and assigns cluster membership probabilities to diseases.
• Enrichment Analysis: Performs enrichment analysis on disease-associated proteins or disease clusters to identify overrepresented biological functions and annotations.

Scientific Applications:

• Shared Mechanism Discovery: Identifies common molecular causes and shared pathogenic mechanisms among genetic diseases by comparing protein interaction profiles.
• Therapeutic Target and Pathway Identification: Suggests potential therapeutic targets and pathways by highlighting proteins commonly involved across similar diseases.
• Hypothesis Generation for Genetic and Proteomic Studies: Generates and prioritizes hypotheses for experimental validation in genetic and proteomic research based on disease similarity and enrichment results.

Methodology:

Models a network of disease and protein nodes connected by protein–protein interactions and disease–protein associations; runs random walks initiated and terminated at disease nodes and records protein visit counts as vector elements; computes cosine similarity between disease vectors; applies probabilistic clustering to similarity scores and performs enrichment analysis, grounded in network theory and statistical analysis.