Nanopores and AI used to classify respiratory viruses
Researchers at Osaka University have developed a label-free method for identifying respiratory viruses, based on dips in electrical current as they pass through silicon nanopores. The method could form the basis of a rapid new Covid-19 test.
This new system for identification of common respirator pathogens – responsible for conditions such as Covid-19 and influenza – is based on a machine-learning algorithm, trained on the changes in current across silicon nanopores.
The technique uses silicon nanopores sensitive enough to detect a single virion when coupled with the algorithm.
A 50nm silicon nitride layer is suspended on a silicon wafer, and tiny nanopores are added, jut 300nm in diameter. When a current is applied to the solution on either side of the wafer, ions travel through the nanopores (a process known as electrophoresis).
The motion of these particles can be monitored by the current they generate. When a viral particle enters a pore, it blocks some of the ions from passing through, leading to a dip in current. Each dip reflects the physical characteristics of the particle, such as volume, surface charge, and shape; this allows for distinction between viruses.
The natural variation in the physical characteristics of virus particles had obstructed implementation of this approach. However, coupling the nanopores with machine learning based on signals from known respiratory viruses allowed the team to build a working tool for distinguishing between viruses.
“By combining single-particle nanopore sensing with artificial intelligence, we were able to achieve highly accurate identification of multiple viral species,” said Professor Makusu Tsutsui, senior author of the ACS Sensors study.
The system was tested with the SARS-CoV-2 virus and similar pathogens: respiratory syncytial virus, adenovirus, influenza A, and influenza B. It was capable of discriminating between electrical current waveforms, which are impossible for humans to do. It performed with very high accuracy, and is faster than other rapid viral tests such as PCR and antibody-based screening, and also does not require expensive reagents.
“This work will help with the development of a virus test kit that outperforms conventional viral inspection methods,” said Professor Tomoji Kawai.
According to the researchers, coronaviruses are particularly well suited to this tool due to their spiked outer proteins, which may allow different strains to be classified separately.
