Transcriptomic Data Mining to Identify Circular RNAs as Biomarkers in the Development of Grape Clusters

Circular RNAs (circRNAs) represent a novel class of RNA molecules with the potential toinfluence critical biological processes. These molecules exhibit tissue-specific expression patterns thatare depend on the cell type and developmental stage. Sistan wine grapes, characterized by seed lessnessand small berry size, were the focus of our investigation. In this study, we aimed to identify the indexcircRNAs that are active in cluster formation and investigated the effects of gibberellin treatment ontheir expression levels. Eight detection tools were used to predict the expressed circRNAs. Ten attributeweighting algorithms were used to identify the most important circRNAs from total reliable identifiedcircRNAs as features. Before performing the algorithms, data were normalized and changed into ۰-۱amounts. Three ranges greater than ۰.۵, ۰.۷۵ and ۰.۹۵ were defined as selection thresholds. In total, ChiSquared and Rule models had the most selected attributes (weigh>۰.۹) and PCA, SV, Relief andDeviation had the lowest one (weight<۰.۵). The number of ۱۷ circRNAs was selected considering thethreshold of ۰.۹۵ and at least six weighting algorithms as biomarkers. These circRNAs can beconsidered as biomarkers in the three stages of grape cluster formation treated with gibberellic acid ornon-treated plants. Among these, ۷ were related to treatment and others were controls. The onlycircRNA identified by eight weighting methods is ۱۱:۱۶۲۵۰۲۹۴-۱۶۲۵۰۵۳۲ from the group of controlplants, which had higher expression in the first and third stage of cluster formation. It corresponds togene ID VIT_۱۱s۰۱۰۳g۰۰۴۲۰ and is associated with the gene ontology (GO) term "regulatory ncRNAmediatedgene silencing" in the biological process (BP) category.

کلیدواژه ها:

circRNAs ، data mining ، gibberellin treatment ، grape ، weighting algorithms