Proteome Profiling Outperforms Transcriptome Profiling for Coexpression Based Gene Function Prediction.

Area of Science:

• Genomics
• Proteomics
• Bioinformatics

Background:

• Gene coexpression analysis, often using mRNA data, infers gene function but has limitations.
• Advancements in mass spectrometry enable global protein expression profiling.

Purpose of the Study:

• To systematically investigate if proteome profiling outperforms transcriptome profiling for gene function prediction.
• To compare mRNA and protein coexpression networks in cancer.

Main Methods:

• Constructed and analyzed mRNA and protein coexpression networks using TCGA and CPTAC data for three cancer types.
• Compared network wiring, gene coexpression drivers, and module preservation between mRNA and protein networks.

Main Results:

• Protein coexpression networks were primarily driven by functional similarity, unlike mRNA networks which included chromosomal colocalization.
• Functionally coherent mRNA modules showed better preservation in protein networks.
• Proteomic data significantly strengthened gene-function links for Gene Ontology (GO) and KEGG pathways.
• Identified novel gene-function relationships, including ERBB2 in lipid biosynthesis and AEBP1 as an EMT marker.

Conclusions:

• Proteome profiling is superior to transcriptome profiling for coexpression-based gene function prediction.
• Proteomics should be integrated or preferred in gene function and human disease studies.