i6mA-stack

i6mA-stack predicts DNA N6-methyladenine (6mA) sites in Rosaceae genomes, specifically Rosa chinensis and Fragaria vesca, to support analysis of epigenetic 6mA distributions.

Key Features:

• Machine Learning-Based Approach: Employs a two-layer stacking ensemble of machine learning algorithms for 6mA site prediction.
• Feature Optimization: Uses recursive feature elimination with cross-validation (RFECV) to determine an optimal number of features (ONF) from five DNA sequence encoding schemes: Binary Encoding (BE), Ring-Function-Hydrogen-Chemical Properties (RFHC), Electron-Ion-Interaction Pseudo Potentials of Nucleotides (EIIP), Dinucleotide Physicochemical Properties (DPCP), and Trinucleotide Physicochemical Properties (TPCP).
• Benchmarking: Demonstrated superior performance versus peer tools in predicting 6mA sites.

Scientific Applications:

• 6mA site mapping in Rosaceae: Identification of genomic 6mA sites in Rosa chinensis and Fragaria vesca for studies of methylation distribution.
• Epigenetic function analysis: Support for investigating cellular processes influenced by 6mA, including aspects of epigenetic inheritance.
• Cross-domain research: Applicability to studies of 6mA roles in both eukaryotic and prokaryotic organisms.

Methodology:

Train a two-layer stacking ensemble using an ONF subset of features selected by RFECV derived from BE, RFHC, EIIP, DPCP, and TPCP encoding schemes.