Supersplat

Supersplat identifies splice junctions from RNA sequencing (RNA-seq) reads by comparing reads to genomic references to support genome annotation and transcript structure discovery.

Key Features:

• Empirical splice-junction discovery: Identifies splice junctions directly from RNA-seq data without relying on existing splice annotations and reports intron/exon boundaries.
• Processing throughput: Processes RNA-seq datasets at a reported rate of approximately 11.4 million reads per hour.
• High-throughput scaling: Scales to large RNA-seq datasets to support large-scale genomic studies.
• Benchmarking and validation: Benchmarked using Illumina RNA-seq reads mapped against the Arabidopsis thaliana genome.
• De novo annotation utility: Applied for de novo annotation use cases, including analysis of Brachypodium distachyon.
• Implementation: Implemented in C++ and optimized for performance.

Scientific Applications:

• Genome annotation: Delineates intron/exon boundaries and splice junctions to improve structural genome annotations.
• Transcriptome characterization: Detects splice junctions from RNA-seq data for both model organisms and novel species, including Arabidopsis thaliana and Brachypodium distachyon.
• De novo gene model annotation: Supports generation of gene models in species lacking comprehensive reference annotations.

Methodology:

Maps RNA-seq reads to genomic reference sequences and empirically identifies splice junctions from the mapped reads; implemented in C++ with a reported throughput of ~11.4 million reads per hour and validated using Illumina RNA-seq mapped to the Arabidopsis thaliana genome.