The burgeoning domain of brain disease research is increasingly recognizing the limitations of traditional genomic analysis methods, particularly in addressing the complex interplay of diverse data types inherent in the multi-omics landscape. Traditional approaches often fall short in integrating these varied data streams, resulting in a fragmented understanding of brain diseases. This fragmentation significantly impedes the development of effective diagnostic and therapeutic strategies, highlighting the urgent need for a more holistic and integrative methodology. This research introduces Design of Iterative Methods in Brain Disease Research (DIMBDR) to transcend these limitations by synergistically combining Canonical Correlation Analysis (CCA) with advanced computational techniques, including Recurrent Neural Networks (RNNs) with attention mechanisms, SHapley Additive exPlanations (SHAP) for interpretable machine learning, and transfer learning from pre-trained Convolutional Neural Networks (CNNs). CCA is employed for its prowess in integrating multi-omic data, facilitating a comprehensive view of the molecular landscape of brain diseases. RNNs, enhanced by attention mechanisms, excel in identifying subtle genomic patterns, while SHAP values offer transparency in feature importance and transfer learning leverages pre-existing models for refined disease prediction. The integration of these computational methods with clinical data offers a novel, multi-dimensional perspective on brain diseases. The proposed model was capable of achieving a superior accuracy rate of 94.2%, a higher precision of 93.6%, recall of 92.8% and F1 score of 93.2%. Its innovative approach not only enhances our understanding of complex brain diseases but also holds substantial promise for improving patient outcomes through more effective and personalized treatment strategies.