Development of a spot reliability evaluation score for DNA microarrays

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• DNA microarrays are widely used for gene expression profiling.
• Assessing the reliability of microarray data compared to other methods like quantitative real-time PCR (QRT-PCR) is crucial for accurate biological interpretation.
• Existing methods may not adequately quantify the concordance between microarray and QRT-PCR expression levels.

Purpose of the Study:

• To develop a novel reliability index for DNA microarray data.
• To quantify the probability that calibrated gene expression levels from DNA microarrays are within a twofold difference compared to QRT-PCR.
• To establish a statistically robust method for evaluating spot-level reliability in gene expression profiling.

Main Methods:

• Developed the Spot Reliability Evaluation Score (SRED) based on two key parameters: measurement reproducibility and relative expression value.
• Selected these parameters from nine candidate metrics for optimal performance.
• Applied the SRED score to a large dataset of approximately 1,500,000 expression profile points.

Main Results:

• The SRED score provides a probability measure for the concordance of gene expression levels between DNA microarrays and QRT-PCR.
• The developed index effectively evaluates the reliability of individual spots on a DNA microarray.
• Demonstrated the application of SRED on a substantial dataset within the RIKEN Expression Array Database.

Conclusions:

• The SRED score offers a valuable tool for assessing the reliability of DNA microarray gene expression data.
• This index enhances confidence in microarray results by providing a quantitative measure of agreement with QRT-PCR.
• SRED facilitates more accurate data interpretation and downstream analysis in genomics research.