Reimagining virus diagnostics: novel approaches for virus detection supplemented by biophysical characterisation studies of viral envelope interactions

Thesis

The development of rapid, robust, and scalable virus diagnostics is a critical global health priority, as highlighted by the COVID-19 pandemic. This thesis introduces innovative diagnostic approaches that integrate biophysical and computational tech- niques to enhance virus detection and characterisation. Central to this work is the calcium-mediated labelling of viral envelopes with single-stranded DNA, enabling the rapid immobilisation and fluorescent tagging of viral particles. These methods are coupled with fluorescence microscopy and machine learning to achieve fast and accu- rate virus identification.

In addition to fluorescence-based diagnostics, this thesis explores a diffusion-based detection method, leveraging single-particle tracking and statistical modeling to clas- sify viral particles within seconds. This complementary approach offers ultra-fast detection capabilities, reducing diagnostic times to under one minute while maintain- ing high specificity and sensitivity.

Key contributions include the biophysical characterization of cation-mediated in- teractions, focusing on the effects of pH, ion concentration, and lipid composition on labeling efficiency. Supported lipid bilayers and virus-like particles were employed as biomimetic systems, elucidating the roles of membrane heterogeneity and surface pro- teins. The machine learning pipeline developed for this work distinguishes between closely related virus strains and accurately identifies SARS-CoV-2 in clinical samples, achieving high sensitivity and specificity.

By integrating these diagnostic methods, this thesis addresses limitations of es- tablished technologies and aligns with the ASSURED and REASSURED criteria for point-of-care diagnostics. These advancements not only contribute to pandemic pre- paredness but also deepen our understanding of virus-membrane interactions, with significant implications for the development of next-generation diagnostic and thera- peutic platforms.

Authors: Shiaelis, N

• DOI: 10.5287/ora-bmjr6ayde
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: diagnostics; neutron reflectometry; fluorescence microscopy; machine learning; virus
• Subjects: Biophysics