SplicingCompass

SplicingCompass detects genes with differential alternative splicing between two conditions from RNA-seq exon read counts by computing geometric angles between exon-count vectors.

Key Features:

• Differential splicing detection: Predicts genes with differential splicing between two distinct conditions using RNA-seq exon read counts.
• Geometric angle analysis: Computes geometric angles between high-dimensional vectors representing exon read counts to detect splicing variation.
• Sensitivity to complex patterns: Detects complex and previously unknown splicing events.
• Validation: Performance validated using simulated experiments and complementary in silico analyses.
• Neuroblastoma application: Demonstrated ability to distinguish favourable and unfavourable clinical outcomes in neuroblastoma tumour data by predicting differential splicing events.
• Regulatory motif identification: Identifies regulatory splicing factor motifs within exons of predicted genes.
• Patient and tissue clustering: Enables clustering of patients and tissues based on splicing-derived information.
• Supporting evidence outputs: Provides normalized read coverage plots and reports reads spanning exon-exon junctions as evidence of splicing diversity.
• Short-read sequencing compatibility: Designed for analysis of short-read RNA-seq data.
• Implementation: Implemented as an R package.

Scientific Applications:

• Differential splicing analysis: Identification of genes with significant alternative splicing between two conditions using RNA-seq exon counts.
• Cancer outcome stratification: Differentiating favourable and unfavourable clinical outcomes in neuroblastoma based on splicing patterns.
• Regulatory mechanism discovery: Linking splicing changes to regulatory splicing factor motifs within exons.
• Patient and tissue stratification: Clustering patients and tissues based on splicing information to explore biological diversity and disease mechanisms.
• Large-scale alternative splicing studies: Scalable analysis of alternative splicing from short-read sequencing data.

Methodology:

Computes geometric angles between high-dimensional exon read-count vectors derived from RNA-seq; identifies regulatory splicing factor motifs within exons; performs patient and tissue clustering based on splicing-derived features; generates normalized read coverage plots and extracts reads spanning exon-exon junctions as evidence.