Related Experiment Video
Updated: Aug 11, 2026
13:00
A Rapid High-throughput Method for Mapping Ribonucleoproteins (RNPs) on Human pre-mRNA
Published on: December 2, 2009
Comparison of preprocessing procedures for oligo-nucleotide micro-arrays by parametric bootstrap simulation of
J Freudenberg1, H Boriss, D Hasenclever
1Interdisciplinary Center of Bioinformatics (IZBI), University of Leipzig, Germany.
Methods of Information in Medicine
|February 11, 2005
Summary
Simulating microarray data revealed that normalization methods can inflate false positive rates when many genes are similarly affected. Variance stabilizing normalization (VSN) performed best in this analysis.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Microarray experiments require robust preprocessing procedures due to limited calibration data.
- Simulation methods are crucial for assessing the performance of these procedures.
Purpose of the Study:
- To evaluate the robustness of various microarray preprocessing procedures.
- To analyze their performance against varying numbers of differentially expressed genes and up-regulation proportions.
Main Methods:
- Simulated oligo-nucleotide microarray raw data using a multivariate normal distribution.
- Incorporated chip effects and artificially spiked-in gene expression differences.
- Compared thirty preprocessing procedures from the BioConductor project.
Main Results:
- Background correction reduced bias but increased probe variance and mean squared error.
- Normalization generally reduced variance and increased sensitivity, with Variance Stabilizing Normalization (VSN) showing the best performance.
- Asymmetrical gene expression changes inflated false positive discovery rates.
Conclusions:
- A parametric bootstrap approach effectively simulated microarray data.
- Current normalization methods can increase false positive rates when numerous genes exhibit directional expression changes.
