Improving predictive performance of nucleotide biosensors via equivalent circuit modeling in transcriptomic analysis

Accurate consumer-oriented diagnostic devices for cancer screening at an early stage are limited despite the rapid evolution of consumer electronics technology, owing to a high error rate and lack of predictive accuracy. In our study, we propose an equivalent \circuit modeling methodology to improve the predictive accuracy of DNA/RNA-based impedimetric biosensors in the context of transcriptomics. Our methodology has two key objectives: one is to minimize errors to achieve analytical accuracy, and the other is to achieve a label-free biosensor to make it user-friendly. In our study, a biosensor was developed by immobilizing probe DNA on gold nanoparticle-modified screen-printed electrodes. Circuit parameters were estimated by simulation and curve-fitting techniques based on a conventional equivalent circuit known as the Randles circuit, resulting in an error rate of ∼6.5%. To accurately simulate multilayer structures consisting of gold nanoparticles, probe DNA, and target miRNA, our model was extended by adding resistive elements and a finite diffusion element, resulting in a significantly low error rate of ∼1.8% to ∼2.2%. However, this setting was not as appropriate for the case of magnetite and magnetite nanocomposite-based systems, in which errors were found to be around 5-6%. In the case of magnetic nanoparticle-based biosensors, a modified Randles circuit containing double-layer capacitance (Cdl), Cole-Cole (CC), and inductive elements (L) was found to have higher fitting accuracy, with errors as low as 1-2%. Moreover, the proposed biosensor has shown promising results in terms of ultra-low detection limits, i.e., 0.5 ag/mL, which makes this biosensor suitable for the detection of transcriptomic biomarkers such as miRNA. Therefore, this study has shown the potential of equivalent circuit modeling in improving the predictive accuracy of nucleotide-based biosensors, thereby promoting their use in scalable, label-free, and user-centric applications in the field of transcriptomics-based diagnostics and RNA-based interventions, including decentralized cancer screening in rural and urban areas.