Graph Attention Network and Graph Convolutional Network for Classification of Dengue Virus Variants

Dengue fever is a serious global health concern, particularly in warm regions where it causes significant morbidity and mortality. Dengue symptoms range from mild fever to severe hemorrhagic complications and are caused by four serotypes (DENV-1, DENV-2, DENV-3, and DENV-4), with several infections increasing the likelihood of severe results. The proper classification of Dengue virus variations is critical for vaccine development, public health policies, and treatment optimization.This article provides a new approach for categorizing Dengue virus variations that employs Graph Convolutional Networks (GCNs) and Graph Attention Networks (GATs). The graph built using k-mers and k-nearest neighbors, processed with the GAT model, achieved a 99.75% classification accuracy, outperforming existing methods.