Effects of scanning sensitivity and multiple scan algorithms on microarray data quality.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• DNA microarray data utility depends on scanner settings.
• Multiple scan approaches aim to reduce quantification error and saturation effects.
• Direct comparisons of these multi-scan methods are lacking.

Purpose of the Study:

• Compare individual scans (low, medium, high sensitivity) with multi-scan data combination methods.
• Evaluate efficacy of 2-scan and 3-scan approaches using a reference RNA standard dataset.
• Assess impact on data quality metrics like background signal, signal-to-noise ratio, and reproducibility.

Main Methods:

• Analysis of 40 technical replicates of a reference RNA standard.
• Comparison of individual scans at varying sensitivities.
• Evaluation of three distinct methods for combining data from multiple scans (2-scan and 3-scan).

Main Results:

• Low sensitivity scans showed the lowest background and highest signal-to-noise ratio.
• Most multi-scan methods increased probe intensity range but not dynamic range.
• One 3-scan method significantly improved reproducibility; a 2-scan version of the same method showed lower reproducibility due to model fit issues.

Conclusions:

• A suitable multiple scan approach can enhance data reproducibility in DNA microarrays.
• Model validation is crucial for accurate probe intensity estimation when using multi-scan methods.
• The choice of scan number and statistical model significantly impacts results.