Related Experiment Videos
Correcting log ratios for signal saturation in cDNA microarrays
Lori E Dodd1, Edward L Korn, Lisa M McShane
1Biometric Research Branch, DCTD, National Cancer Institute, Bethesda, MD 20892-7434, USA. doddl@mail.nih.gov
Bioinformatics (Oxford, England)
|May 18, 2004
Summary
This study introduces a novel method to correct pixel saturation in microarray data, improving gene expression estimation accuracy. The approach uses pixel-level data and a censored regression model for reliable results.
Area of Science:
- Bioinformatics
- Genomics
- Microarray Technology
Background:
- Pixel saturation in microarrays leads to biased gene expression estimates.
- Saturation is influenced by photomultiplier tube (PMT) voltage and RNA abundance.
- Current methods of discarding or using biased saturated spot data are suboptimal.
Purpose of the Study:
- To develop and evaluate a method for correcting pixel saturation in microarray data.
- To improve the accuracy of gene expression estimators affected by saturation.
- To address the limitations of current saturation handling techniques.
Main Methods:
- A novel method utilizing pixel-level data for saturation correction.
- Application of a censored regression model.
- Evaluation through analysis of multiple microarray datasets.
Main Results:
- The proposed method effectively corrects for pixel saturation.
- Evaluations on several arrays demonstrate good performance.
- Simulation studies confirm the method's robustness under model violations.
Conclusions:
- The developed method offers a significant improvement for analyzing saturated microarray data.
- Accurate gene expression quantification is achievable even with saturated pixels.
- This approach provides a more reliable alternative to discarding or using biased data.