Detecting differentially expressed circular RNAs from multiple quantification methods using a generalized linear mixed model

• DOI: 10.1016/j.csbj.2022.05.026
• Publication Date: 2022-05-20T17:08:44Z
• Authors: Alessia Buratin; Chiara Romualdi; Stefania Bortoluzzi; Enrico Gaffo
• Keywords: Robustness; Mixed model; Expression (computer science)

Finding differentially expressed circular RNAs (circRNAs) is instrumental to understanding the molecular basis of phenotypic variation between conditions linked circRNA-involving mechanisms. To date, several methods have been developed identify circRNAs, and combining multiple tools becoming an established approach improve detection rate robustness results in circRNA studies. However, when using a consensus strategy, it unclear how expression estimates should be considered integrated into downstream analysis, such as differential assessment. This work presents novel solution test quantifications algorithms simultaneously. Our analyzes tools' abundance count data within single framework by leveraging generalized linear mixed models (GLMM), which account for sample correlation structure quantification tools. We compared GLMM with three widely used models, showing its higher sensitivity detecting efficiently ranking significant circRNAs. strategy first consider combined methods, we propose powerful model analysis.