Evaluating Microarray Design Methods for Time Profiles
Author Information
Author(s): Ana C Fierro, Raphael Thuret, Kristof Engelen, Gilles Bernot, Kathleen Marchal, Nicolas Pollet
Primary Institution: CNRS UMR 8080, Laboratoire Développement et Evolution, Orsay, France
Hypothesis
How well do different linear models reconstruct expression ratios and time profiles from complex microarray designs?
Conclusion
Including a dye effect in certain methods makes results more robust against array failure.
Supporting Evidence
- The correlation between estimated ratios was high across different methods.
- Methods with a dye effect were more robust against array failure.
- Including random effects requires more parameters and is advised only with sufficient replicates.
Takeaway
This study looked at different ways to analyze gene expression data from experiments and found that some methods work better than others, especially when dealing with problems like bad data.
Methodology
The study compared five profile reconstruction methods using linear models on data from two-channel microarrays.
Potential Biases
Potential biases from array failures and normalization methods were noted.
Limitations
The study's findings may not apply to all types of microarray designs or conditions.
Participant Demographics
The study used data from a Xenopus tropicalis expression profiling experiment.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website