DSGseq

DSGseq identifies differentially spliced genes between two groups of RNA-sequencing (RNA-seq) samples by modeling exon-level read counts to detect changes in exon usage.

Key Features:

• Exon-Based Approach: Models sequencing reads at the exon level to avoid reliance on isoform composition or prior isoform annotations.
• Negative Binomial (NB) Modeling: Uses the negative binomial distribution to model exon read count variability across samples.
• NB-Statistic: Computes an NB-statistic that compares read counts across exons between sample groups to identify differential splicing.
• Detection of Novel Events: Highlights exons most likely to be differentially spliced, enabling discovery of previously unknown alternative splicing events.
• Validation on Data: Demonstrated on simulated data and on real human RNA-seq datasets from kidney and liver samples.

Scientific Applications:

• Gene regulation studies: Enables exon-level analysis of alternative splicing to investigate regulatory mechanisms from RNA-seq data.
• Disease mechanism investigation: Identifies differential exon usage that may serve as biomarkers or contribute to pathogenesis.
• Developmental biology: Detects splicing changes across conditions or developmental stages at the exon level.
• Comparative transcriptome analysis: Facilitates comparison of alternative splicing patterns between sample groups using next-generation RNA-seq.

Methodology:

Models exon read counts using the negative binomial distribution and computes an NB-statistic to compare exon-level read counts across sample groups to detect differentially spliced genes.