frma

frma preprocesses Affymetrix microarray data to enable reliable single-array analysis by using frozen parameter estimates and probe-level weighting that downweights probes with high between-batch residual variance to mitigate batch effects.

Key Features:

• Single-array preprocessing: Performs preprocessing on individual arrays while preserving the benefits of multiarray approaches for probe-level summarization.
• Batch effect mitigation: Reduces batch effects by downweighting probes with significant between-batch residual variance in the probe-level model.
• Extension to Affymetrix Human Exon and Gene 1.0 ST: Adapts the fRMA probe-level model for Affymetrix Human Exon and Gene 1.0 ST platforms with compatibility for oligo ExonFeatureSet and GeneFeatureSet objects.
• Customizable parameter estimation (frmaTools): Enables generation of custom frozen parameter vectors from user datasets when standard parameter databases are not available.
• R/Bioconductor implementation: Provided as R/Bioconductor packages for integration into R-based analysis workflows.

Scientific Applications:

• Single-array analyses: Enables analysis of individual microarrays when multiarray preprocessing is impractical or infeasible.
• Gene expression profiling: Supports gene-level expression analysis on Affymetrix platforms.
• Exon-level analyses: Supports exon-level analysis on Affymetrix Human Exon arrays.
• Cross-batch and cross-study comparisons: Facilitates comparisons across batches or studies by mitigating batch effects through probe weighting.

Methodology:

Generates frozen parameter estimates from a database of arrays; applies a probe-level model that downweights probes with high between-batch residual variance; provides adaptations of the probe-level model for Affymetrix Human Exon and Gene 1.0 ST with oligo ExonFeatureSet/GeneFeatureSet compatibility; and allows creation of custom frozen parameter vectors via the frmaTools package; implemented in R/Bioconductor.