MyoMiner

MyoMiner maps gene co-expression in normal and pathological striated muscle from high-throughput mRNA transcriptomics to enable analysis of muscle-specific gene interaction networks.

Key Features:

• Muscle-Specific Data Curation: Aggregates publicly available high-throughput mRNA expression data from 2376 mouse and 2228 human striated muscle samples categorized into 142 groups by species, tissue origin, age, gender, anatomic part, and experimental condition.
• Co-Expression Analysis: Provides per-category co-expression values and enables retrieval of correlated genes for a query gene with adjusted p-values and confidence intervals quantifying correlation strength and standardized scatterplots visualizing expression relationships for each gene pair.
• Network Interface: Constructs 2-shell correlation networks from the most highly correlated genes or from a user-defined gene list, with options to include linked genes from the database.
• Correlation Comparison Test: Statistically tests whether two correlation coefficients from different conditions are significantly different.
• Guilt-by-Association Prioritization: Prioritizes candidate genes based on co-expression patterns using a guilt-by-association approach.

Scientific Applications:

• Delineation of Muscle-Specific Networks: Identifies tissue-, cell-, and pathology-specific components of muscle protein interactions, cell signaling pathways, and gene regulation.
• Comparative Disease Analysis: Compares co-expression patterns between healthy and diseased tissues to reveal altered networks and potential disease mechanisms or therapeutic targets.
• Candidate Gene Prioritization: Ranks genes for functional follow-up in muscle-related diseases using co-expression-based guilt-by-association.

Methodology:

Acquisition and categorization of high-throughput mRNA transcriptomics data into 142 muscle-specific categories; calculation of co-expression values per category; computation of adjusted p-values and confidence intervals for gene pairs; generation of standardized scatterplots; construction of 2-shell correlation networks; and statistical testing of differences between correlation coefficients.