Ultrasensitive MicroRNA Detection Combining Reduced Graphene Oxide Electrolyte-Gated Transistors and Machine Learning

MicroRNAs (miRNAs) are promising biomarkers for disease diagnosis, but conventional detection methods such as reverse transcription polymerase chain reaction (RT-PCR) require complex instrumentation and reagents, limiting their suitability for portable diagnostics. Here, we report an ultrasensitive and selective biosensor that integrates DNA-functionalized reduced graphene oxide (rGO), electrolyte-gated transistors (EGTs), and machine learning (ML) for miRNA detection. The platform targets the miR-34 family (miR-34a, miR-34b, and miR-34c), which is associated with cancer and neurological disorders. The biosensor discriminates perfectly matched from mismatched sequences over a wide dynamic range (0.1-1000 amol L^-1) with an ultralow limit of detection of 0.098 amol L^-1. ML enables multidimensional analysis of EGT transfer curves and extraction of physically meaningful features from high-dimensional data. This approach advances point-of-care technologies for highly sensitive and selective miRNA detection, with strong potential for portable molecular diagnostics.