Related Experiment Videos
Bayesian estimation of fold-changes in the analysis of gene expression: the PFOLD algorithm
J Theilhaber1, S Bushnell, A Jackson
1Aventis Pharmaceuticals, Cambridge Genomics Center, 26 Landsdowne Street, Cambridge, MA 02139, USA. joachimtheilhaber@aventis.com
Summary
A new noise model and Bayesian method improve DNA microarray analysis for gene expression. The PFOLD program enhances detection of changing genes with greater sensitivity and accuracy than fold-change alone.
Area of Science:
- Genomics
- Bioinformatics
- Molecular Biology
Background:
- DNA microarrays are crucial for measuring gene expression.
- Accurate noise modeling is essential for reliable gene expression analysis.
- Current methods often rely solely on fold-change, potentially limiting sensitivity.
Purpose of the Study:
- To present a general noise model for DNA microarray measurements.
- To develop a Bayesian estimation scheme for gene expression ratios.
- To implement a computational tool (PFOLD) for enhanced gene expression analysis.
Main Methods:
- Developed a detailed noise model applicable to oligonucleotide and cDNA microarrays.
- Derived a Bayesian estimation scheme for expression ratios and fold-change.
- Implemented the scheme in the PFOLD software, incorporating confidence limits and P-values.
Main Results:
- The PFOLD program provides gene expression fold-change estimates with confidence limits and P-values.
- The noise model offers seamless estimation across a wide range of signal-to-noise ratios.
- Combined P-value and fold-change analysis significantly improves detection sensitivity for changing genes.
Conclusions:
- The developed Bayesian approach and PFOLD software offer a robust method for DNA microarray data analysis.
- This method enhances the detection of differentially expressed genes compared to traditional fold-change analysis.
- The noise model provides a unified framework for various microarray technologies.