Fast Viral Diagnostics: FTIR-Based Identification,\nStrain-Typing, and Structural Characterization of SARS-CoV‑2

Severe\nacute respiratory syndrome coronavirus 2 (SARS-CoV-2) has\ntriggered an ongoing global pandemic, necessitating rapid and accurate\ndiagnostic tools to monitor emerging variants and preparedness for\nthe next outbreak. This study introduces a multidisciplinary approach\ncombining Fourier Transform Infrared (FTIR) microspectroscopy and\nMachine learning to comprehensively characterize and strain-type SARS-CoV-2\nvariants. FTIR analysis of pharyngeal swabs from different pandemic\nwaves revealed distinct vibrational profiles, particularly in nucleic\nacid and protein vibrations. The spectral wavenumber range between\n1150 and 1240 cm<sup>–1</sup> was identified as the classification\nmarker, distinguishing Healthy (noninfected) and infected samples.\nMachine learning algorithms, with neural networks exhibiting superior\nperformance, successfully classified SARS-CoV-2 variants with a remarkable\naccuracy of 98.6%. Neural networks were also able to identify and\ndifferentiate a small cohort infected with influenza A variants, H1N1\nand H3N2, from SARS-CoV-2-infected and Healthy samples. FTIR measurements\nfurther show distinct red shifts in vibrational energy and secondary\nstructural alterations in the spike proteins of more transmissible\nforms of SARS-CoV-2 variants, providing experimental validation of\nthe computational data. This integrated approach presents a promising\navenue for rapid and reliable SARS-CoV-2 variant identification, enhancing\nour understanding of viral evolution and aiding in diagnostic advancements,\nparticularly for an infectious disease with unknown etiology.