Improving Microarray Data Analysis by Correcting Non-Specific Binding
Author Information
Author(s): Eugene F Schuster, Eric Blanc, Linda Partridge, Janet M Thornton
Primary Institution: European Bioinformatics Institute
Hypothesis
Can statistical models improve the estimation of non-specific binding in microarray experiments?
Conclusion
Using combined statistical methods improves the detection of differential expression in microarray data.
Supporting Evidence
- The MAS5 PM-MM model was found to be inadequate for estimating non-specific binding.
- The GC robust multi-array average method outperformed others in detecting differential expression.
- The study emphasizes the importance of correcting for non-specific binding to improve data analysis.
Takeaway
This study shows that better methods can help scientists understand which genes are really active by correcting for background noise in experiments.
Methodology
The study used various statistical models to analyze microarray data and compared their effectiveness in estimating non-specific binding.
Potential Biases
There is a risk of bias in P value distributions for null probesets, leading to inaccurate false discovery rates.
Limitations
The study highlights gaps in understanding specific binding signals and the need for more datasets with known transcript concentrations.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website