Related Experiment Videos
Sources of nonlinearity in cDNA microarray expression measurements
L Ramdas1, K R Coombes, K Baggerly
1Department of Pathology, University of Texas M D Anderson Cancer Center, Houston, TX 77030, USA. wzhang@mdanderson.org
Genome Biology
|December 12, 2001
Summary
Microarray data analysis assumes linear signal-to-gene expression. This study found nonlinearities in both radioactive and fluorescent microarrays due to signal quenching and scanner transformation, impacting gene expression analysis.
Area of Science:
- Molecular Biology
- Genomics
- Biotechnology
Background:
- Microarray data analysis relies on the assumption of a linear relationship between quantified signal intensity and gene expression levels.
- This assumption is critical for accurate interpretation of gene expression data derived from microarrays.
Purpose of the Study:
- To experimentally validate the linearity assumption in microarray data analysis.
- To investigate potential sources of nonlinearity in both radioactive and fluorescent microarray platforms.
Main Methods:
- Examined the signal-to-expression relationship for radioactively labeled cDNA on nylon membranes.
- Investigated the signal-to-expression relationship for fluorescently labeled cDNA on glass slides.
- Utilized dilution series experiments to detect nonlinearities.
Main Results:
- Identified signal quenching due to excessive dye concentrations as a source of nonlinearity in fluorescent microarrays.
- Discovered a nonlinear data transformation by the scanner affecting radioactive microarrays.
- Noted that image-quantification software packages varied in their correction for scanner-induced nonlinearities.
Conclusions:
- Two sources of nonlinearity were identified, impacting both radioactive and fluorescent microarray analyses.
- The scanner-induced nonlinearity in radioactive microarrays is particularly problematic as it is not easily predictable.
- Simple dilution series are recommended as a quality control measure to detect these nonlinearities.