EMM-DTI: ENHANCED MAMBA-BASED MODEL FOR DRUG-TARGET INTERACTION PREDICTION

Drug-target interaction (DTI) prediction is a cornerstone of drug development, enabling accurate predictions that accelerate the process and reduce costs. However, traditional methods in this field are often costly and time-intensive. In recent years, deep learning models, such as Transformers and Graph Neural Networks (GNNs), have driven significant advancements in DTI prediction. Despite these advances, many models fail to account for the structural features of DTIs, resulting in limited predictive power. To address this limitation, we propose EMM-DTI, a novel prediction model based on Mamba technology, which excels at capturing features from long sequence data. The experimental results demonstrate that its performance is superior to existing advanced methods, with reduced training costs and shorter training times.