DNA-Logic Trimodal Biofuel Cell Platform for AI-Assisted Dual Pathogen Detection.

A DNA-logic-gated, trimodal, self-powered biosensing platform is developed for the rapid, simultaneous, and on-site detection of two major sugarcane pathogens, Sporisorium scitamineum (smut) and Fusarium sacchari (pokkah boeng). Conventional diagnostic techniques such as qPCR are considered time-consuming, laboratory-dependent, and unsuitable for field deployment. To overcome these limitations, the proposed device integrates electrochemical, colorimetric, and photothermal signal readouts into a single portable system, powered by an enzymatic biofuel cell and operated via a smartphone interface. The sensing mechanism is orchestrated by target-specific DNA logic circuits. In the presence of smut DNA, glucose oxidase is released, generating an electrochemical current and a color shift in a methylene blue-based electrolyte. For pokkah boeng, an AuCo nanozyme is activated, catalyzing the oxidation of TMB to yield a blue color and a distinct photothermal response under NIR laser irradiation. Both pathways are triggered orthogonally without cross-interference. Multimodal signals are processed using a machine learning-assisted random forest regression model. The smut detection model achieves an R2 value of 0.982 with a learning rate of 0.024 and 63% training data allocation, while the pokkah boeng model attains perfect accuracy (R2 = 1.000) at a learning rate of 0.001 and a 90% training split. The platform demonstrates high sensitivity, with detection limits as low as 3.7 × 10-16 M for smut and 2.3 × 10-16 M for pokkah boeng, along with excellent repeatability and stability. Validation using field samples collected from infected sugarcane plants shows high consistency with standard qPCR assays. The system provides a powerful, low-cost, and user-friendly tool for early disease warning and represents a universal strategy for multiplex pathogen screening in smart agriculture frameworks.