IntEREst

IntEREst estimates intron retention and exon-exon junction levels from RNA-seq data to analyze pre-mRNA splicing efficiency and aberrant splicing associated with disease.

Key Features:

• Intron Retention Analysis: Enables detailed examination of intron retention levels across transcripts, supporting analysis of both annotated and non-annotated introns beyond RefSeq.
• Data Input and Processing: Accepts binary sequence alignment/map (.bam) files and performs genome-wide estimation of intron retention and exon-exon junction levels.
• Comparative Analysis: Facilitates intra-sample and inter-sample comparisons to identify significant differences in intron retention across conditions or sample groups.
• Subset Comparison: Allows definition and comparison of specific intron subsets, such as U12-type versus U2-type introns.
• Visualization Tools: Provides plotting functions to visualize the distribution of intron retention levels.
• Statistical Methods: Incorporates statistical methods adapted from DESeq2, edgeR, and DEXSeq to identify significantly more or less retained introns.
• Parallel Processing: Supports sequential execution on a single core or parallel execution across multiple cores.

Scientific Applications:

• U12- and U2-type intron retention studies: Applied to human and plant RNA-seq datasets to investigate retention differences between U12- and U2-type introns.
• U12-dependent spliceosome defect analysis (ZRSR2 mutations): Used to study samples with defects in the U12-dependent spliceosome due to mutations in the ZRSR2 gene component.
• Retention pattern characterization: Revealed that U12-type introns are retained at higher levels than U2-type introns in control samples, with retention exacerbated in samples carrying mutated ZRSR2.

Methodology:

Performs genome-wide estimation of intron retention and exon-exon junction levels from .bam RNA-seq files; statistical framework adapted from DESeq2, edgeR, and DEXSeq; supports sequential or parallel execution across multiple cores; comparative analyses with IRFinder have shown considerable overlap.