Gene communities in co-expression networks across different tissues

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Tissue-specific gene expression data, such as from the Genotype-Tissue Expression (GTEx) Consortium, enables comparisons of gene co-expression patterns across different human tissues.
• Gene co-expression networks are valuable for identifying groups of genes with similar expression, potentially indicating shared biological functions, responses to stimuli, or regulatory mechanisms.

Approach:

• A multilayer network framework was employed, with each layer representing a tissue-specific gene co-expression network from four exocrine gland tissues.
• Novel methods for multilayer community detection were developed, specifically designed for correlation matrix input and incorporating a null model.
• The approach successfully identified both 'generalist' communities (genes co-expressed across multiple tissues) and 'specialist' communities (genes co-expressed within a single tissue).

Key Points:

• The analysis identified five generalist gene communities and two specialist gene communities.
• Gene co-expression communities were found to exhibit significant physical clustering on chromosomes 1 and 11, suggesting shared regulatory elements.
• Specific gene families, such as KRTAP3-1, KRTAP3-3, and KRTAP3-5, were implicated in shared regulation in skin and pancreas.
• CELA3A and CELA3B were found to share expression quantitative trait loci in the pancreas.

Conclusions:

• The developed multilayer community detection method effectively extracts biologically relevant gene communities from correlation matrix data.
• The findings highlight the utility of multilayer network analysis for understanding complex gene co-expression patterns across diverse tissues.
• The identification of tissue-specific and cross-tissue gene modules provides a foundation for further research into gene regulation and function.