Pre-processing Agilent microarray data
Author Information
Author(s): Zahurak Marianna, Parmigiani Giovanni, Yu Wayne, Scharpf Robert B, Berman David, Schaeffer Edward, Shabbeer Shabana, Cope Leslie
Primary Institution: Johns Hopkins University School of Medicine
Hypothesis
The goal of this study is to quantify some of the sources of error that affect measurement of expression using Agilent arrays and to compare Agilent's Feature Extraction software with pre-processing methods that have become the standard for normalization of cDNA arrays.
Conclusion
Simple loess normalization without background subtraction resulted in low variance fold changes that more reliably ranked gene expression than the other methods.
Supporting Evidence
- Simple loess normalization without background subtraction produced the lowest variability.
- ROC analysis showed that differentially expressed genes are most reliably detected when background is not subtracted.
- Loess normalization and no background subtraction yielded an AUC of 99.7%.
Takeaway
This study looked at how to best process data from Agilent microarrays to get accurate results. It found that a simple method without background subtraction works best.
Methodology
The study compared Agilent's Feature Extraction software with simple loess normalization methods, both with and without background subtraction.
Potential Biases
Gene specific dye effects persisted after normalization, affecting nearly half the genes.
Limitations
The study did not fully realize the larger goal of defining best study design and pre-processing practices for Agilent arrays.
Statistical Information
Confidence Interval
99%CI : 39%, 47%
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website