Physicochemical Hydrodynamics and Machine Learning Characterization of Isothermal Nucleic Acid Amplification Nucleation Site Analysis

Abstract

Nucleic acid diagnostics have advanced since PCR's first use in sickle cell and HIV diagnosis in the 1980s. PCR remains the gold standard for detecting SARS-CoV-2 and monitoring HIV viral load but is limited by resource-intensive requirements, making it impractical for low-resource or home settings. My research aims to adapt PCR’s strengths using isothermal nucleic acid amplification for rapid, low-cost diagnostics to support global health.I first present a novel assay combining RPA-based amplification with lateral flow detection, offering PCR-level sensitivity with LFA-like ease. It meets WHO SARS-CoV-2 detection standards, demonstrates high specificity, variant resilience, and uses a simple lysis method suitable for minimal devices. For HIV viral load monitoring, I developed a buffer-modified recombinase polymerase amplification assay on microfluidic chips using amplification nucleation site analysis (ANSA), where nucleation site counts correlate with nucleic acid input, enabling precise, mobile phone-compatible measurements. Finally, I describe a machine learning approach using a ResNet-18 model to analyze temporal ANSA data and predict DNA concentrations. Two models classify DNA by clinical groups or log-fold changes. This work supports robust, POC-suitable HIV diagnostics and establishes a platform for broader quantitative nucleic acid testing across global health settings.