StackCirRNAPred

StackCirRNAPred classifies long circular RNAs (circRNAs) from other long non-coding RNAs (lncRNAs) to support analyses of post-transcriptional gene regulation, including microRNA sponge activity and disease-related roles.

Key Features:

• Feature Extraction: Uses nucleic acid composition, sequence spatial features, physicochemical properties, and the presence of Alu and tandem repeats to characterize sequences.
• Stacking Strategy: Integrates multiple classifiers through a stacking ensemble that incorporates Random Forest (RF), LightGBM, and XGBoost.
• Performance Metrics: Primary evaluation reports precision 0.843, accuracy 0.833, F1 score 0.831, recall 0.819, and MCC 0.666, with mouse dataset validation showing precision 0.837, accuracy 0.839, F1 0.839, recall 0.841, and MCC 0.677.

Scientific Applications:

• CircRNA identification: Distinguishes long circRNAs from other lncRNA species for downstream transcriptome analyses.
• Gene regulation studies: Supports investigation of post-transcriptional regulation and microRNA sponge functions of circRNAs.
• Disease and therapeutic research: Enables analysis of circRNA involvement in disease mechanisms and prioritization of potential therapeutic targets.

Methodology:

Extracts features (nucleic acid composition, sequence spatial features, physicochemical properties, Alu and tandem repeats) and combines Random Forest, LightGBM, and XGBoost via a stacking ensemble.