DSS

DSS performs differential analysis of count-based sequencing data to detect differentially expressed genes from RNA-seq and differentially methylated loci or regions from bisulfite sequencing (BS-seq).

Key Features:

• Supported data types: Works with count-based RNA-seq and bisulfite sequencing (BS-seq) data, including reduced representation bisulfite sequencing (RRBS).
• Dispersion shrinkage: Implements a dispersion shrinkage method that estimates dispersion parameters using Gamma-Poisson or Beta-Binomial distributions.
• Statistical model for methylation: Uses beta-binomial regression to model methylation counts for detection of differentially methylated loci and regions.
• Link function: Employs an "arcsine" link function within the beta-binomial regression framework.
• Complex experimental designs: Supports general and multiple-factor experimental conditions beyond simple two-group comparisons.
• Parameter estimation: Performs parameter estimation via a transformed-data approach using generalized least squares, avoiding iterative algorithms.

Scientific Applications:

• Differential expression analysis: Identification of differentially expressed genes (DEGs) from RNA-seq count data.
• Differential methylation analysis: Detection of differentially methylated loci (DML) and regions (DMR) from BS-seq and RRBS data.
• Multi-factor studies: Analysis of methylation and expression changes under general and multiple-factor experimental designs in epigenetics and transcriptomics research.

Methodology:

Estimates dispersion via Gamma-Poisson or Beta-Binomial models with dispersion shrinkage; fits beta-binomial regression using an "arcsine" link for methylation analysis; performs parameter estimation on transformed data using generalized least squares without iterative algorithms.