An Efficient Algorithm for Sensitively Detecting Circular RNA from RNA-seq Data.

Area of Science:

• Bioinformatics
• Genomics
• Molecular Biology

Background:

• Circular RNAs (circRNAs) are a significant class of non-coding RNAs.
• Existing computational methods for circRNA detection from RNA-sequencing (RNA-seq) data exhibit variability in sensitivity and precision.
• There is a need for improved algorithms to accurately identify circRNAs, particularly those with low expression levels.

Purpose of the Study:

• To introduce CIRCPlus, a novel and efficient *de novo* algorithm for circRNA detection.
• To enhance the sensitivity of circRNA detection while preserving acceptable precision.
• To address limitations in existing circRNA detection methods.

Main Methods:

• CIRCPlus identifies circRNA candidates by analyzing back-spliced junction reads.
• It compares local similar sequences of spanning junction read pairs.
• The algorithm leverages information from unbalanced spanning reads for improved detection.

Main Results:

• CIRCPlus achieved up to 90% sensitivity in common simulation settings.
• The algorithm demonstrated balanced sensitivity and reliability on real RNA-seq datasets.
• Performance was validated against existing *de novo* methods using simulation and real data.

Conclusions:

• CIRCPlus offers an improved approach for *de novo* circRNA detection from RNA-seq data.
• The algorithm is particularly effective for detecting circRNAs with low expression levels.
• CIRCPlus provides a sensitive and reliable tool for circRNA identification in bioinformatics research.