A hybrid deep learning framework for accurate N6,2'-O-Dimethyladenosine site prediction

N6,2'-O-dimethyladenosine (m⁶Am) is a critical RNA modification that plays a pivotal role in regulating RNA stability, translation efficiency, and gene expression. Accurate computational identification of m⁶Am sites is essential for advancing epitranscriptomic research and elucidating the molecular mechanisms underlying disease pathogenesis. However, existing methods are often constrained by limited feature representations and insufficient interpretability, hindering their predictive performance and generalisation across diverse transcriptomic contexts. To address these limitations, we propose a novel deep learning framework for accurate and interpretable m⁶Am site prediction. The proposed framework integrates a rich set of heterogeneous sequence-derived features, including Accumulated Nucleotide Frequency (ANF), Enhanced Nucleic Acid Composition (ENAC), Dinucleotide-based Auto-Cross Covariance (DACC), Trinucleotide-based Auto-Cross Covariance (TACC), and Pseudo Dinucleotide Composition (PseDNC), to comprehensively capture local sequence composition, positional nucleotide distribution, and higher-order inter-nucleotide dependencies within RNA sequences. To further enhance model efficiency and interpretability, SHapley Additive exPlanations (SHAP) are employed for principled feature importance analysis, systematically eliminating redundant attributes and retaining only the most discriminative features for downstream classification. The optimised feature set is subsequently fed into a deep neural network (DNN) for accurate m⁶Am site prediction. Evaluated using 10-fold cross-validation, the proposed model achieved 86.55% accuracy and 85.23% on an independent test dataset, confirming its robust generalisation capability. These results establish the proposed framework as an accurate, reliable, and interpretable computational tool for m⁶Am site identification, with broad applicability to epitranscriptomic analysis and the investigation of RNA modification functions in biological processes and disease mechanisms.