Improving Microarray Data Analysis with Quality Weights
Author Information
Author(s): Matthew E Ritchie, Dileepa Diyagama, Jody Neilson, Ryan van Laar, Alexander Dobrovic, Andrew Holloway, Gordon K Smyth
Primary Institution: The Walter and Eliza Hall Institute of Medical Research
Hypothesis
Can a graduated approach to array quality improve the analysis of microarray data?
Conclusion
The proposed method allows poorer quality arrays to be included in analyses, enhancing the detection of differentially expressed genes.
Supporting Evidence
- The method successfully assigns lower weights to less reproducible arrays.
- Down-weighting observations from suspect arrays increases the power to detect differential expression.
- In smaller experiments, this approach outperforms the usual method of filtering the data.
Takeaway
This study shows that we can use less reliable data from microarrays without throwing it away, which helps us find important gene information.
Methodology
The study used a heteroscedastic linear model to assign weights to microarrays based on their quality, improving the analysis of gene expression data.
Potential Biases
Potential biases may arise from the reliance on the quality of the arrays and the assumptions made in the statistical model.
Limitations
The method is restricted to experiments with some level of replication and may not be suitable for all types of microarray data.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website