Choosing Normalization Methods and Test Statistics for Microarray Data
Author Information
Author(s): Yang Xie, Kyeong S. Jeong, Wei Pan, Arkady Khodursky, Bradley P. Carlin
Primary Institution: University of Minnesota
Hypothesis
What normalization method should be used for two-channel microarray data?
Conclusion
The empirical Bayes method (B statistic) is recommended for analyzing microarray data as it performs better than traditional methods.
Supporting Evidence
- The empirical Bayes method outperformed traditional methods in terms of false discovery rates.
- Normalization methods were evaluated using descriptive plots and statistical tests.
- Print-tip group normalization was found to be the most effective method for reducing systematic biases.
Takeaway
This study helps scientists figure out the best way to analyze data from gene experiments, making sure they get accurate results.
Methodology
The study used real data from a DNA-protein binding microarray experiment and compared various normalization methods and statistical tests.
Potential Biases
Potential biases from systematic variations in microarray experiments were acknowledged.
Limitations
The study may not account for all possible biases in microarray experiments.
Participant Demographics
Data from wild-type Escherichia coli was used.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website