RUV-III

RUV-III removes unwanted variation from Nanostring nCounter gene expression data by leveraging technical replicates and control genes to improve normalization for downstream analyses such as differential expression and biomarker discovery.

Key Features:

• Normalization Methodology: Introduces a normalization approach that leverages technical replicates and suitable control genes to remove unwanted variation in Nanostring nCounter data, addressing limitations of nSolver analysis software methods that use negative control probes for background correction, positive controls for within-sample normalization, and reference genes for cross-sample normalization.
• Technical Replicates Utilization: Incorporates technical replicates directly into the normalization model to account for variability introduced during sample processing and improve measurement reliability.
• Pseudo-replicates Approach: Provides an alternative pseudo-replicate strategy to approximate technical replication when true technical replicates are unavailable.
• Dataset Validation: Demonstrated effectiveness and robustness across four distinct datasets.

Scientific Applications:

• Nanostring nCounter experiments: Improves normalization and reduces unwanted technical variation in Nanostring nCounter gene expression studies.
• Differential expression analysis: Reduces technical noise prior to statistical testing, enabling more reliable differential expression results.
• Biomarker discovery: Enhances consistency of transcript quantification across samples to support biomarker identification.
• Oncology, immunology, and developmental biology: Provides refined normalization for studies in these fields that require precise transcript-level measurements.

Methodology:

Models and removes unwanted variation from nCounter count data using technical replicates and selected control genes, and constructs pseudo-replicates when technical replicates are absent, contrasting with conventional nSolver steps of negative control probe background correction, positive-control within-sample normalization, and reference-gene cross-sample normalization.