Instant AI-assisted test for viral infection

A non-DNA based test could identify viral infections in patients in minutes. When a clinician suspects a patient has a viral illness, the presence of specific virus types can be confirmed through a DNA sequencing test. However, the test takes several hours, even if a testing facility is available on site, and the test cannot discern whether the virus is viable. Noriyasu Hashida and colleagues designed a test that confirms the presence of live virus by pushing particles through a nanopore, one at a time, and measuring their electrical conductivity, which varies with size and surface charge as well as the unique molecular structure of the virus. Using an AI model, the device can use electrical conductivity waveforms to identify viruses and even distinguish viruses within the same family. The system was tested on herpesvirus that infect the eye, which can lead to blindness. Samples of aqueous humor—the watery fluid around the eye—were analyzed with both PCR and the AI nanopore method. The sensitivity of the AI nanopore method was 100%, and the specificity was 71.4%, improving to over 90% when conditions were adjusted. According to the authors, as the technology matures, it could help clinicians catch and treat viral ocular infections—and likely other types of viral infections as well.