CircAST

CircAST reconstructs and quantifies full-length circular RNA (circRNA) transcripts, including alternatively spliced isoforms, from RNA-seq data.

Key Features:

• Full-length assembly: Assembles full-length circular RNA transcripts including alternatively spliced isoforms.
• Multiple splice graphs: Uses multiple splice graphs to model circRNA internal structure for accurate transcript assembly.
• Isoform-level read assignment: Assigns sequencing reads to specific isoforms to enable isoform-level quantification.
• Expression quantification accuracy: Estimates expression levels with reported correlation coefficients between theoretical and estimated values ranging from 0.85 to 0.99.
• Performance metrics: Simulation studies reported sensitivity of 85.63%–94.32% and precision of 81.96%–87.55% for full-sequence assembly.
• Low-abundance isoform detection: Demonstrates detection of low-abundance isoforms as evidenced by validation experiments.
• Experimental evaluation: Evaluated on an RNase R-treated mouse testis RNA-seq dataset, identifying 380 circRNAs with full-length sequences distinct from corresponding linear RNAs, with RT-PCR and Sanger sequencing validation confirming 32 of 37 randomly selected isoforms.
• Cross-sample diversity discovery: Applied to published experimental data to reveal substantial diversity of circular transcripts across samples.

Scientific Applications:

• CircRNA isoform reconstruction: Reconstruction and quantification of alternatively spliced circRNA isoforms from RNA-seq datasets.
• RNase R-enriched circRNA identification: Identification and characterization of circRNAs in RNase R-treated RNA-seq experiments.
• Low-abundance isoform analysis: Detection and validation of low-abundance circular RNA isoforms using RT-PCR and Sanger sequencing.
• Comparative circRNA profiling: Comparative analysis of circRNA isoform diversity across samples to investigate regulatory mechanisms and potential disease associations.

Methodology:

CircAST constructs multiple splice graphs from RNA-seq data, assembles full-length circRNA transcripts, assigns sequencing reads to specific isoforms, and estimates expression levels; performance was evaluated using simulation studies.