ABSSeq

ABSSeq models absolute expression differences in RNA-Seq data using a negative binomial distribution to identify differentially expressed genes with reduced false positives.

Key Features:

• Implementation: ABSSeq is implemented as an R package for RNA-Seq data analysis.
• Absolute expression modeling: Models absolute differences in gene expression across conditions using a negative binomial distribution that accounts for variation between genes and among samples.
• Enhanced detection power: Prioritizes the overall magnitude of expression differences to improve detection power and reduce false positives at very low and high expression levels.
• Type I error control: Demonstrates improved control of the Type I error rate relative to alternative differential expression methods.
• Shrinkage of fold change: Calculates shrinkage of fold change estimates to stabilize effect size estimates, improve gene ranking, and assist outlier detection.

Scientific Applications:

• Differential expression analysis: Identifying differentially expressed genes in RNA-Seq experiments for downstream biological inference.
• Biomarker discovery: Discovery of disease biomarkers from RNA-Seq differential expression results.
• Gene regulatory network elucidation: Supporting gene regulatory network studies by providing reliable differential expression signals.
• Downstream functional analysis: Serving as a basis for downstream biological function analysis such as pathway or gene set analysis.
• High-throughput transcriptomics: Applicable to high-throughput transcriptomics studies where precision in differential expression calls is critical.

Methodology:

Models absolute expression differences using a negative binomial distribution and applies shrinkage to fold change estimates while accounting for variation between genes and among samples.