Differentially correlated genes in co-expression networks control phenotype transitions

Area of Science:

• Systems biology
• Bioinformatics
• Genomics

Background:

• Co-expression networks are crucial for analyzing large biological datasets (e.g., transcriptomes, proteomes).
• Differential co-expression analysis identifies changes in gene interactions during biological state transitions.
• The regulatory role of differentially co-expressed genes remains understudied.

Purpose of the Study:

• Investigate differentially co-expressed genes in both simple (B lymphocyte deficiency) and complex (cervical cancer) biological systems.
• Determine the functional role of these genes in disease pathogenesis and regulation.

Main Methods:

• Reconstructed co-expression networks using Pearson correlation and local partial correlation methods for mouse and human datasets.
• Identified differentially correlated gene pairs and analyzed their network positions (Minimum Shortest Path, Bi-partite Betweenness Centrality).
• Validated findings using in vitro knockdown experiments for key genes in cervical cancer.

Main Results:

• In B cell deficiency, differentially co-expressed genes were enriched with immunoglobulin genes.
• In cervical cancer, these genes functioned as network bottlenecks, mediating information flow from drivers to peripheral genes.
• Knockdown of FGFR2 and CACYBP confirmed their regulatory roles in cancer cell growth.

Conclusions:

• Identifying differentially co-expressed genes is vital for uncovering regulatory genes driving phenotypic changes.
• This approach aids in understanding complex biological processes and disease mechanisms.