[17p-Z01-1~11] [Open Symposium] Applied Physics: An essential field in the effort against COVID-19

[17p-Z01-8] Identifying single-viruses via machine learning-enhanced nanopore sensing

〇Makusu TsuTsui¹, Takashi Washio¹, Tomoji Kawai¹ (1.Osaka Univ.)

Keywords:nanopore, machine learning, ionic current

Modern society is facing a challenge to prevent the spread of COVID-19 due in part to a lack of efficient methods for detecting viruses in individuals at an early stage of infection. Here we report on a nanosensor approach for fast screening of viral particles. The sensor consists of a nanoscale hole, which is called a nanopore, formed in a dielectric membrane. It allows a simple way of counting viral particles by measuring a temporal change in ionic current through the nanopore upon their translocation. Furthermore, we demonstrated that those viral particles can be discriminated from distinct difference in the ionic current signal waveforms by using machine learning to classify them in a high-dimensional feature space, thereby accomplished coronavirus discrimination to influenza typing at the single-virus level. The present technique can be used as a rapid and ultrasensitive sensor approach for diagnosing infectious diseases at an early stage of infection.