The Influence of the Global Gene Expression Shift on Downstream Analyses

Area of Science:

• Genomics and Bioinformatics
• Cancer Biology
• Molecular Biology

Background:

• Standard gene expression analysis assumes consistent total RNA abundance across cells.
• Master regulators (e.g., c-MYC) can induce global gene expression shifts, particularly in cancer cells.
• Such shifts challenge the validity of differential expression analysis and molecular tumor classification.

Purpose of the Study:

• To systematically evaluate the influence of global gene expression shifts on differential expression (DE) analysis.
• To assess the impact of global shifts on molecular classification of tumors using gene expression data.
• To compare the performance of existing analysis methods under global shift conditions.

Main Methods:

• Collected gene expression datasets with known global shift effects.
• Simulated global shift scenarios using diverse real gene expression data.
• Conducted comparative analyses of representative DE and classification methods.

Main Results:

• Differential expression analysis methods exhibit varying tolerance to global gene expression shifts.
• Some DE analysis methods are highly sensitive to global shifts, potentially leading to unreliable results.
• Molecular classification accuracy is generally robust to global shifts and can even be enhanced; however, gene selection for classification is significantly affected.

Conclusions:

• The assumption of stable total RNA abundance is critical and often violated in gene expression studies.
• Researchers must consider the potential impact of global gene expression shifts on data analysis reliability.
• Method selection for differential expression analysis should account for potential global shifts; classification approaches show resilience but require careful gene selection.