Fiber optic sensors for molecular diagnostics of viral diseases

Fiber-optic interferometric biosensors generate high-dimensional spectra that encode biomolecular interactions at functionalized sensing layers. This work summarizes a unified machine-learning (ML) interpretation layer that converts such spectra into actionable diagnostic outputs across multiple sensing heads and biological matrices relevant to viral disease monitoring and infection-associated inflammation. We highlight (i) interferometric detection of SARS-CoV-2– specific IgG in wastewater, (ii) interferometric C-reactive protein (CRP) assessment in human urine with clinically motivated thresholding, and (iii) optical spectroscopy for CRP dynamics in complex matrices such as wastewater. Despite different biochemical functionalization strategies and target analytes, the analytical burden is structurally similar: spectra are nonlinear, noise-sensitive, and affected by matrix variability. A consistent pipeline sample-level splitting, preprocessing, engineered spectral features, and supervised models with robust metrics enables reliable discrimination and supports real-time and remote sensing scenarios. The manuscript emphasizes reproducible evaluation practice and identifies deployment risks (drift, temperature sensitivity, and proxy labels) relevant to clinical translation.