RoDiCE: robust differential protein co-expression analysis for cancer complexome.

Area of Science:

• Proteomics
• Cancer Biology
• Bioinformatics

Background:

• The full spectrum of abnormalities in cancer-associated protein complexes is largely unknown.
• Protein complex dysfunction in cancer can be studied by comparing co-expression structures between tumor and healthy cells.
• Mass spectrometry-based proteomics data is often noisy due to biological variants, leading to inaccurate co-expression estimates.

Purpose of the Study:

• To develop a robust algorithm for identifying protein complex aberrations in cancer using differential protein co-expression testing.
• To improve the accuracy of identifying cancer-specific protein dysfunction in noisy proteomic data.
• To provide insights into higher-order differential co-expression structures beyond traditional linear correlations.

Main Methods:

• Proposed a novel algorithm based on copula for differential protein co-expression testing.
• Applied the algorithm to large-scale proteomic data from renal cancer.
• Compared the proposed copula-based method with conventional linear correlation-based approaches.

Main Results:

• The copula-based method demonstrated improved identification accuracy with noisy data compared to linear correlation methods.
• Successfully identified important protein complexes, regulatory signaling pathways, and potential drug targets in renal cancer.
• The approach revealed higher-order differential co-expression structures, offering deeper insights than traditional methods.

Conclusions:

• The developed algorithm provides a robust approach for identifying protein complex aberrations in cancer.
• The method enhances the accuracy of analyzing noisy proteomic data for cancer research.
• This technique can aid in discovering novel cancer-specific pathways and therapeutic targets.