Past, present, and future strategies for detecting and quantifying circular RNA variants

Abstract

Circular RNAs (circRNAs) are a family of covalently closed RNA transcripts ubiquitous across the eukaryotic kingdom. CircRNAs are generated by a class of alternative splicing called backsplicing, with the resultant circularization of a part of parental RNA producing the characteristic backsplice junction (BSJ). Because of the noncontiguous sequence of the BSJ with respect to the DNA genome, circRNAs remained hidden in plain sight through over a decade of RNA next-generation sequencing, yet over 3 million unique circRNA transcripts have been illuminated in the past decade alone. CircRNAs are expressed in a cell type-specific manner, are highly stable, with many examples of circRNAs being evolutionarily conserved and/or functional in specific contexts. However, circRNAs can be very lowly expressed and predictions of the circRNA context from BSJ-spanning reads alone can confound extrapolation of the exact sequence composition of the circRNA transcript. For these reasons, specific and ultrasensitive detection, combined with enrichment, bespoke bioinformatics pipelines and, more recently, long-read, highly processive sequencing is becoming critical for complete characterization of all circRNA variants. Concomitantly, the need for targeted detection and quantification of specific circRNAs has sparked numerous laboratory-based and commercial approaches to visualize circRNAs in cells and quantify them in biological samples, including biospecimens. This review focuses on advancements in the detection and quantification of circRNAs, with a particular focus on recent next-generation sequencing approaches to bolster detection of circRNA variants and accurately normalize between sequencing libraries.