Despite the promising performance of nanomaterial-based
electrochemical
detection techniques, product-level point-of-care sensor systems are
still evolving. Unique approaches toward nanoscale tunability, enriched
measurements, and data analysis are therefore imperative. The recent
outbreak of coronavirus disease provides a suitable test bench toward
this pursuit, attracting several innovative efforts from the sensor
research community. Accordingly, this work reports an ultrasensitive
and selective molecular biosensor for RT-PCR-confirmed hospital samples
of SARS-CoV-2 built on ZnO nanoflowers. The unique experimental and
analytical methodologies drastically improve the device efficiency,
relying on three conserved genomic sequences (E, N, and ORF1ab genes).
It involves multiple techniques like cyclic voltammetry, electrochemical
impedance spectroscopy, and differential pulse voltammetry where normalized
responses of these techniques are used to create 1D, 2D, and 3D data
points representing each sample for clustering both by heuristic (2D
separation plane) and ML approach (logistic regression) for accurate
diagnosis. Enhancing data features also enlarges the Euclidean distances
(ED) of data points, which raises the separation efficiency of the
device to 93.75%. As the 3D space also allows the data (with nearly
identical EDs) to have different spatial orientations for efficient
clustering, samples were dichotomized into positive and negative with
100% efficiency, sensitivity, and specificity. Furthermore, owing
to the ultrasensitive nanostructured transducer, the viral load was
successfully quantified in the spiked control samples with a commendable
detection sensitivity and limits of detection as low as 14.137 fM
(E-gene). The device’s fast response time (<45 min) justifies
its potential on-field applicability.