TurboNorm

TurboNorm normalizes microarray intensity data using a P-spline scatterplot smoother to provide computationally efficient intensity-dependent normalization as an alternative to lowess/loess for one- and two-colour arrays and DNA methylation arrays.

Key Features:

• P-spline Scatterplot Smoother: Uses a P-spline scatterplot smoother that leverages all available data points for normalization.
• Avoidance of Random Subsets: Performs normalization on the entire dataset rather than on random subsets to improve stability of the normalization curve.
• Weighing of Observations: Supports weighting of individual observations during smoothing to modulate the influence of points on the fitted curve.
• Constant Weighted P-spline Normalization: Implements constant-weighted P-spline normalization for scenarios such as pre/post intervention comparisons and invariant-subset contexts.
• Compatibility with Array Data Types: Applicable to both one- and two-colour microarray data and to DNA methylation arrays.
• Computational Efficiency: Provides a faster and simpler alternative to lowess/loess for large datasets.
• R Package Implementation: Implemented as an R package.

Scientific Applications:

• DNA Methylation Array Normalization: Produces precise normalization curves for DNA methylation arrays to improve accuracy of downstream analyses.
• Large-scale Microarray Studies: Enables intensity-dependent normalization in large datasets where lowess/loess methods are computationally limiting.
• Pre/Post Intervention Comparisons: Supports normalization strategies for comparing groups before and after interventions, including use of invariant subsets with constant-weighted P-spline normalization.

Methodology:

Implements a P-spline scatterplot smoother applied to all data points, supports observation weighting and constant-weighted P-spline normalization for invariant-subset or group-comparison scenarios, and functions as a computationally efficient alternative to lowess/loess for one- and two-colour microarray and DNA methylation array data.