Graphene/TiO₂ Nanocomposite based Electrochemical Biosensor Enhanced by Support Vector Machine Classification Model to Detect Different DENV Serotype IgG

Sequential infections by different DENV serotypes could lead to life-threatening complications. Thus, it is crucial to detect serotype specific DENV infection at early onset which current rapid test kits are not able to do so. This work presents a nano-enabled electrochemical biosensor for detection of serotype specific DENV infection. The proposed biosensor incorporated Artificial Intelligence (AI) as post processing tool for enhanced biosensor performance. The key element of the biosensor is the use of consensus DENV envelope protein domain-III (cEDIII) peptide probe that has relatively high specific affinity towards different DENV antibodies (IgG) serotypes. Screen-printed carbon electrode (SPCE) was modified with graphene/titanium dioxide (G/TiO2) nanocomposite for improved electron charge transfer. The immunosensor exhibited high sensitivity towards two different DENV IgG serotype, namely DENV1 IgG and DENV2 IgG with limit of detection (LOD) of 14.05 ng/mL and 10.56 ng/mL respectively. The linear working range for DENV1 IgG and DENV2 IgG were 562.5-9000 ng/mL and 31.25-500ng/mL, respectively. For optimal post processing tool, machine learning algorithm was used to improve the proposed biosensors' performance, where support vector machine attained highest accuracy of 94% in differentiation between different DENV IgG targets detection based on the principal components extracted from EIS data.