About circAtlas database
CircAtlas (circRNA atlas) is a database developed for circRNAs from 17 human, 13 macaque and 14 mouse normal tissue samples. For each tissue, three types of RNA-seq libraries were constructed: (i) RiboMinus/RNase R treatment for circRNA identification, (ii) Poly(A) enrichment for linear RNA identification and quantification, and (iii) RiboMinus treatment for accurate circRNA and mRNA expression comparison. These three libraries were sequenced with paired-end 250 bp, 150 bp , and 250 bp reads, respectively. The identified circRNAs and their expression patterns, genomic features, conservations and functional annotations have been integrated into circAtlas, where users can browse, search, retrieve, visualize and prioritize circRNAs and their related information.
Representative features of circAtlas are summarized as follows:
- Expression analysis: RNA-seq data with RiboMinus treatment were used to quantify the expression of circRNA and mRNA.
- Junction ratio: This parameter measures the relative expression of circRNA against its linear counterpart across the back-splice junction.
- Annotation: To predict functional circRNAs and their potential regulatory mechanisms, a co-expression network was constructed for each of the three species, where nodes represented genes (circRNAs or mRNAs) and edges represented a co-expression relationship between gene pairs. By employing a ‘guilt-by-association’ analysis, thousands of circRNAs were assigned functional annotation with GO terms.
- OO-type circRNAs: Orthologous genes were identified and the circRNAs on these orthologous genes were subsequently screened. The resulting circRNAs with the same BSJ sites between any two species were denoted as overlapped orthologous (OO-type) circRNAs.
- Full-length: The RNA-seq data sets were analyzed using the CIRI-full pipeline to detect alternative splicing events and to assemble into full-length circRNAs.