Related Experiment Videos
Identifying pre-post chemotherapy differences in gene expression in breast tumours: a statistical method appropriate
E L Korn1, L M McShane, J F Troendle
1Biometric Research Branch, EPN-8128, National Cancer Institute, National Institutes of Health, Bethesda MD 20892, USA. korne@ctep.nci.nih.gov
British Journal of Cancer
|April 16, 2002
Summary
Cluster analysis may misinterpret gene expression data. New methods reveal 17 genes with increased expression after chemotherapy in breast tumors, findings missed by prior cluster analysis.
Area of Science:
- Bioinformatics
- Genomics
- Statistical analysis
Background:
- Microarray data analysis is crucial for understanding gene expression.
- Cluster analysis is a common but potentially unsuitable method for certain microarray data aims.
- Paired sample analysis is essential for detecting treatment-induced changes.
Purpose of the Study:
- To evaluate the suitability of cluster analysis for paired microarray data.
- To identify differentially expressed genes in breast tumors before and after chemotherapy using appropriate statistical methods.
- To highlight the limitations of cluster analysis in specific research contexts.
Main Methods:
- Reanalysis of existing breast tumor microarray data.
- Application of statistical methods that utilize paired differences.
- Identification of differentially expressed genes.
Main Results:
- Seventeen genes were identified as significantly upregulated after chemotherapy.
- These key findings were not reported in the original cluster analysis.
- The study demonstrates the superiority of paired difference analysis for this dataset.
Conclusions:
- Statistical methods utilizing paired differences are more appropriate for analyzing paired gene expression data.
- Cluster analysis may obscure important biological findings in such datasets.
- Accurate statistical approaches are vital for reliable gene expression analysis in cancer research.