Importance of data structure in comparing two dimension reduction methods for classification of microarray gene expression data
2007

Comparing Methods for Analyzing Gene Expression Data

Sample size: 58 publication Evidence: moderate

Author Information

Author(s): Caroline Truntzer, Catherine Mercier, Jacques Estève, Christian Gautier, Pascal Roy

Primary Institution: CNRS, UMR 5558 – Equipe Biostatistique Santé, Villeurbanne, France

Hypothesis

The study evaluates how gene expression variance structure influences the performance of different classification methods.

Conclusion

Discriminant Analysis outperformed Between-Group Analysis because it allows for the dataset structure.

Supporting Evidence

  • Discriminant Analysis is recommended for datasets with complex structures.
  • Simulated datasets were used to assess the performance of analysis methods.
  • The study highlights the importance of understanding dataset structure before analysis.

Takeaway

This study looks at how different ways to analyze gene data can help doctors tell apart different types of cancer based on gene information.

Methodology

The study compared Between-Group Analysis and Discriminant Analysis using simulated and public datasets.

Limitations

The study warns against using datasets with simple structures for method comparison.

Participant Demographics

58 patients with Diffuse Large B-Cell Lymphoma, divided into 32 'cured' and 26 'fatal/refractory'.

Digital Object Identifier (DOI)

10.1186/1471-2105-8-90

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication