mdqc

mdqc assesses microarray data quality by computing Mahalanobis distances on multivariate quality attributes to identify outlier arrays and support rigorous quality assessment.

Key Features:

• Multivariate Analysis: Calculates Mahalanobis distance for an array's quality attributes relative to other arrays, combining multiple QC parameters into a single multivariate metric.
• Outlier Detection: Flags arrays with unusually high Mahalanobis distances as potential low-quality outliers.
• Rich Information Extraction: Analyzes quality attributes jointly to capture interactions among different QC metrics.
• Computational Efficiency: Performs multivariate assessment with low computational cost once QC reports are available.
• Flexibility in Analysis: Computes distances on subsets of quality measures to refine detection and localize specific quality issues.

Scientific Applications:

• Microarray quality assessment: Detects and flags low-quality arrays in microarray experiments to improve dataset integrity.
• Bias prevention in downstream analysis: Identifies problematic arrays that could introduce artifacts or bias into differential expression and other downstream analyses.
• Experimental troubleshooting: Pinpoints subsets of QC measures associated with outlying arrays to inform investigation of experimental or technical issues.

Methodology:

Utilizes existing QC reports; computes Mahalanobis distance of each array's quality attributes relative to a reference set to identify outliers; optionally computes distances on subsets of quality measures to refine detection and elucidate specific quality issues.