OriC-ENS

OriC-ENS predicts origins of replication (ORI) in Saccharomyces cerevisiae using sequence-based ensemble classification to identify DNA replication initiation sites.

Key Features:

• Feature extraction: Uses K-mer, K-gapped Mono-Di, and K-gapped Di-mono sequence features to capture composition and gapped pattern information.
• Classifier ensemble: Employs an ensemble of three Support Vector Machine (SVM) classifiers—one trained on K-mer features and two trained on K-gapped Mono-Di and K-gapped Di-mono features.
• Integration method: Combines SVM outputs via majority voting to produce final ORI predictions.
• Validation datasets: Evaluated on an S. cerevisiae dataset and independently validated on OriDB sequences including Schizosaccharomyces pombe.

Scientific Applications:

• ORI prediction in S. cerevisiae: Identification of origins of replication to support studies of DNA replication initiation and genomic replication architecture.
• Cross-species application: Validation on OriDB including Schizosaccharomyces pombe demonstrates applicability to multiple yeast species.
• Performance benchmarking: Achieved 91.62% accuracy, Matthews Correlation Coefficient (MCC) 0.83, Area Under Curve (AUC) 0.98, sensitivity 0.90, and specificity 0.92, outperforming other state-of-the-art methods on the tested dataset.

Methodology:

Extracts K-mer, K-gapped Mono-Di, and K-gapped Di-mono features; trains three SVM classifiers on these feature sets; integrates predictions by majority voting; implemented in Python.