CircMiner

CircMiner detects circular RNAs (circRNAs) from RNA-Seq data by filtering linear reads and identifying back-splice junctions to enable accurate circRNA discovery for studies of their roles in disease and as biomarkers.

Key Features:

• Stand-alone method: Implements an independent pipeline for circRNA detection without reliance on external mapping workflows.
• Linear-read filtering: Efficiently identifies and filters out linear RNA sequencing reads to reduce false positives.
• Rapid pseudo-alignment: Employs a pseudo-alignment technique to discern reads originating from transcripts, genes, or the genome.
• Back-splice junction resolution: Processes remaining reads to pinpoint back-splice junctions with single-nucleotide resolution.
• Chimeric-read targeting: Addresses detection of chimeric reads necessary for locating back-splice junctions.
• Performance validation: Validated on simulated datasets derived from known back-splice junctions and on RNase R treated cell line datasets, showing superior accuracy and speed compared to existing tools.
• Implementation: Implemented in C++.

Scientific Applications:

• CircRNA discovery: Identification and precise mapping of circRNAs from short-read RNA-Seq data across eukaryotic samples.
• Biomarker and disease studies: Detection of candidate circRNAs associated with diseases such as cancer for biomarker and functional investigations.
• Experimental comparison and validation: Comparative analysis of RNase R treated versus untreated samples to identify high-confidence circRNAs and benchmarking with simulated datasets.

Methodology:

Filter linear reads via rapid pseudo-alignment to identify reads originating from transcripts/genes/genome, then process the remaining (chimeric) reads to pinpoint back-splice junctions at single-nucleotide resolution; implemented in C++.