saseR

saseR enables scalable aberrant expression and splicing analyses through a highly performant framework that processes BAM files, models count data with the negative binomial distribution, and computes p-values for detecting aberrant expression and splicing events.

Key Features:

• BAM Processing: BamtoAspliCounts processes BAM files to generate ASpli counts for downstream analysis.
• Count Conversion: convertASpli extracts gene, bin, or junction counts from ASpli SummarizedExperiment objects.
• Offset Calculation: calculateOffsets creates offset assays specifically tailored for aberrant expression or splicing analysis.
• Latent Factor Optimization: saseRfindEncodingDim estimates the optimal number of latent factors to include when estimating mean expression.
• Statistical Fitting: saseRfit performs parameter estimation of the negative binomial distribution and computes p-values for aberrant expression and splicing detection.

Scientific Applications:

• Aberrant Expression Analysis: Identification of genes with expression levels deviating from expected values in RNA-seq data.
• Aberrant Splicing Analysis: Detection of alternative splicing events that differ from normal splicing patterns in RNA-seq data.
• Differential Usage Analysis: Scalable analysis of differential exon or transcript usage enabled by the offset methodology.

Methodology:

saseR uses an offsets-based approach to unlock RNA-seq tools for fast scalable analyses, with parameter estimation based on the negative binomial distribution and latent factor estimation for mean expression modeling.