Machine Learning‐Assisted Polarity‐Switchable Photoelectrochemical Platform Based on B‐Eu 2 O 3 Nanospindle and Dual‐Amplified Rod‐Shaped DNA Walker

ABSTRACT This work first reports the synthesis of boron‐doped europium oxide (B‐Eu 2 O 3 ) nanospindle by carbonization of B‐doped Eu‐based metal organic framework (Eu‐MOF), as a high‐performance polarity‐switchable material with efficient charge transport. Based on B‐Eu 2 O 3 nanospindle and target‐triggered dual rod‐shaped DNA walker signal amplification, a novel polarity‐switchable photoelectrochemical (PEC) biosensing platform is constructed for retinol binding protein 4 (RBP4) detection, using machine learning for the effective classification of type 2 diabetes mellitus (T2DM). The fabricated PEC biosensor based on B‐Eu 2 O 3 nanospindle and dual‐amplified rod‐shaped DNA walker for RBP4 detection shows high selectivity and sensitivity with a detection limit of 0.49 fg/mL. This developed B‐Eu 2 O 3 nanospindle‐based PEC platform for RBP4 assay in serum effectually differentiates T2DM from healthy persons. Moreover, the machine learning model is applied to explore the potential pattern hidden in data of the established technology, and the accuracy of T2DM classification can acquire 100%. The integrated detection platform combining B‐Eu 2 O 3 nanospindle‐based PEC biosensor with machine learning offers a new avenue for biomarker monitoring.