LVSmiRNA

LVSmiRNA implements Least-Variant Set (LVS) normalization for Agilent miRNA arrays to reduce technical variance and improve accuracy of miRNA expression measurements for downstream analyses.

Key Features:

• Least-Variant Set (LVS) normalization: Implements the Least-Variant Set approach adapted for Agilent miRNA arrays to improve normalization when standard assumptions are violated.
• Robust linear model probe-level fit: Applies a robust linear model fit at the probe level to account for variance differences between probes.
• Performance in spike-in studies: Demonstrated sensitivity and specificity comparable to ideal normalization methods in spike-in experiments.
• Comparative benchmarks: Outperformed no normalization, 75th percentile-shift, quantile, global median, VSN (Variance Stabilization Normalization), and lowess normalization in evaluations.
• Expression-summary measures: Evaluates four expression-summary measures with robust-model-derived summarization performing comparably to other methods on tissue datasets.
• Cross-tissue robustness: Exhibits superior operating characteristics across dissimilar and similar tissue expression data.

Scientific Applications:

• miRNA quantification: Provides normalized miRNA expression measurements for Agilent miRNA arrays to reduce technical variance.
• Downstream analyses: Supports differential expression studies and pathway analysis by providing a consistent baseline for miRNA data.

Methodology:

Implements Least-Variant Set (LVS) normalization using a robust probe-level linear model fit; evaluated with spike-in studies, four expression-summary measures on tissue datasets, and comparisons to 75th percentile-shift, quantile, global median, VSN, lowess, and no normalization.