Scientists have learned how to sort respiratory viruses using machine learning

Scientists from Osaka University have unveiled a new system using silicon nanopores. They are residual

sensitive to detect even a single viral particle using a machine learning algorithm.

To a silicon nitride layer only 50 nm thick,scientists added tiny nanopores. When a different voltage is applied to the solution on both sides of the wafer, ions pass through the nanopores during electrophoresis.

The movement of ions can be tracked by the generatedthem current. When a viral particle enters the nanopore, it blocks the passage of some ions, resulting in a short-term current dip. Each tilt angle reflects the physical properties of the particle, such as its volume, surface charge, and shape. These metrics can be used to determine the type of virus.

Detection of single viral particles using solid-state nanopores. Credit: Osaka University.

Natural change in physical propertiesviral particles previously hindered this approach. But using machine learning, the team created a classification algorithm to identify new samples. “By combining single-particle sensing of nanopores with artificial intelligence, we were able to achieve highly accurate identification of several types of viruses,” explains senior author of the study, Makusu Tsutsui.

The team believes that coronaviruses are particularly well suited for this method. Their spiny outer proteins even make it possible to classify different strains separately.

Compared to other rapid viral tests, such as PCR or antibody-based screening, the new method is much faster and does not require expensive reagents.

Read also

The Doomsday glacier turned out to be more dangerous than scientists thought. We tell the main thing

Scientists discover two new mammals in Australia

Voluntary death. We tell how the euthanasia procedure works around the world