A network diffusion approach to inferring sample-specific function reveals functional changes associated with breast...

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Gene function is often inferred from its interactions within biological networks.
• Altered network contexts, particularly in diseases like cancer, may change gene functions.
• Previous studies have explored network changes but not their functional impact.

Purpose of the Study:

• To investigate network-induced functional changes in genes specific to breast cancer.
• To characterize how altered gene networks impact gene function in cancer.
• To explore novel methods for tumor characterization beyond standard functional annotation.

Main Methods:

• Utilized transcriptomic data from The Cancer Genome Atlas (TCGA) for 1047 breast tumors and 110 healthy tissues.
• Derived sample-specific protein-interaction networks for each sample.
• Applied a diffusion strategy to assign sample-specific gene functions.

Main Results:

• Identified significant gains or losses of genes associated with specific functions in cancer samples.
• These functional shifts were primarily driven by changes in network topology, not just differential gene expression.
• Functional changes predicted by network topology were supported by mutation and copy number data.

Conclusions:

• Network topology changes significantly alter gene functions in breast cancer.
• This diffusion-based functional assignment offers a novel tumor characterization complementary to existing methods.
• The approach effectively predicts patient survival and identifies histopathological subtypes.