iORI-ENST

iORI-ENST identifies origins of replication (ORIs) in DNA sequences to support analysis of DNA replication mechanisms and regulation of gene expression.

Key Features:

• Feature Extraction: Integrates mono-nucleotide binary encoding with dinucleotide-based spatial autocorrelation to represent DNA sequence characteristics for ORI prediction.
• Elastic Net for Feature Selection: Applies the elastic net regularized regression (combined L1 and L2 penalties) to select an optimal subset of features.
• Stacking Learning Ensemble: Uses a stacking ensemble combining random forest, AdaBoost, gradient boosting decision tree, extra trees, and support vector machine (SVM) classifiers to predict ORIs and non-ORIs.

Scientific Applications:

• DNA replication research: Facilitates identification of ORIs to study DNA replication initiation and mechanisms.
• Gene regulation and disease studies: Supports investigation of regulatory mechanisms of gene expression and potential interventions for genic diseases.
• Benchmarking and validation: Provides validated performance on benchmark datasets S1 (91.41%), S2 (95.07%), and an independent dataset S3 (91.10%).

Methodology:

Features were extracted using mono-nucleotide binary encoding and dinucleotide-based spatial autocorrelation; elastic net was applied for feature selection; a stacking learning framework integrating random forest, AdaBoost, gradient boosting decision tree, extra trees, and SVM was trained and validated on datasets S1 (91.41%), S2 (95.07%), and independent S3 (91.10%).