In a laboratory at Osaka University, researchers have shaped a sliver of silicon nitride, thinner than a few hundred atoms stacked end to end, into a sieve through which single viruses can be felt rather than seen. Each passing particle briefly chokes the flow of ions and leaves a fingerprint in the current, and a trained algorithm learns to read those fingerprints the way a practiced ear learns voices. It is a modest step in a long human effort to know what invisible agents are moving through our bodies and communities, reported in the journal ACS Sensors in late 2020, as the world searched u
Nanopore Sensors Paired With AI Could Identify Single Virus Particles, Osaka Team Reports
Each dip reflects the physical properties of the particle
Mark
So what is actually new here? Virus tests already exist.
Mimi
The new part is that they detect single particles. A virus passes through a nanopore in a 50-nanometre silicon nitride layer, and it causes a brief dip in electrical current.
Mark
And the dip tells you which virus it is?
Mimi
The researchers say each dip reflects the particle's volume, surface charge, and shape. An AI classifier trained on known viruses reads those signals.
Luke
Worth noting that the AI claim is the researchers' own. They say it spots waveform differences humans can't see and classifies with high accuracy, but the report gives no accuracy figure.
Mark
Which viruses did they test?
Mimi
Coronavirus, adenovirus, influenza A, and influenza B.
Luke
That's four. Real clinical samples are messier, and the report doesn't say how the method handled those.
Mark
They also say it beats PCR and antibody tests.
Mimi
On speed and on not needing costly reagents, yes. Kawai said it will help develop a kit that outperforms conventional inspection.
Luke
That's a forward-looking statement, not a result. Faster and cheaper is the team's comparison, and no head-to-head data is given here.
Mark
And the strain idea?
Mimi
They believe the spiky proteins of coronaviruses may let strains be classified separately. Believe is the key word. It isn't shown yet.
Der Puls
- A silicon membrane barely 50 nanometres thick becomes the stage for a tense, precise test: one virus at a time, a tiny current dip signals that something has passed.
- The disruption is quiet but real, since conventional PCR and antibody screens demand costly reagents and time that a pandemic cannot easily spare.
- Researchers are navigating toward a classifier that hears differences in current waveforms invisible to human eyes, and they have already tested it on coronavirus, adenovirus, and influenza A and B.
- The method is landing as a promising laboratory demonstration, with its speed and cost claims still awaiting clinical validation and a commercial test kit.
In a laboratory at Osaka University, researchers have shaped a sliver of silicon nitride, thinner than a few hundred atoms stacked end to end, into a sieve through which single viruses can be felt rather than seen. Each passing particle briefly chokes the flow of ions and leaves a fingerprint in the current, and a trained algorithm learns to read those fingerprints the way a practiced ear learns voices. It is a modest step in a long human effort to know what invisible agents are moving through our bodies and communities, reported in the journal ACS Sensors in late 2020, as the world searched urgently for faster ways to name a threat. Whether this quiet method will leave the laboratory and reach the bedside remains an open question that only time and clinical trial can answer.
The device at the heart of this study is a silicon nitride membrane only 50 nanometres thick, pierced with tiny pores and stretched across a silicon wafer. Scientists at Osaka University in Japan believe it can identify viruses one particle at a time, and they reported their work in the journal ACS Sensors.
The method depends on electricity. A voltage across the solution pushes ions through the pores, and when a virus slips into one, it blocks some of those ions and the current briefly dips. The researchers argue that each dip carries a signature of volume, surface charge, and shape, enough to suggest what kind of virus has passed.
That is where artificial intelligence enters. The team trained a classification algorithm on known viruses so it could recognize new samples, detecting differences in waveforms that human eyes cannot see. The system was tested on coronavirus, adenovirus, and influenza A and B. Makusu Tsutsui, the senior author, described the result as highly accurate identification of multiple viral species.
The team suspects coronaviruses may suit the approach especially well, since their spiky proteins might even help distinguish strains. They also argue the method is much faster than PCR or antibody screens and needs no costly reagents, which they hope will lead to better diagnostics for emerging diseases.
For now, this is a laboratory demonstration. Its performance in clinical settings and its path to a usable test kit have yet to be shown.
The device at the center of this study is a layer of silicon nitride just 50 nanometres thick, suspended on a silicon wafer and pierced with tiny pores. A team of scientists, including researchers at Osaka University in Japan, says that layer can be used to identify individual virus particles one at a time. Their work was published in the journal ACS Sensors.
The mechanism rests on electrical current. When a voltage difference is applied to the solution on either side of the wafer, ions travel through the nanopores in a process called electrophoresis. The motion of those ions can be monitored through the current they generate. When a viral particle enters a pore, it blocks some of the ions from passing, and the current briefly dips.
The researchers say each dip carries information about the particle that caused it. The physical properties involved include volume, surface charge, and shape, and the scientists say this measure can be used to tell what kind of virus is passing through.
Reading those dips is where artificial intelligence comes in. The team built a classification algorithm and trained it on signals from known viruses, so it could then determine the identity of new samples. According to the scientists, the computer can discriminate differences in current waveforms that human eyes cannot pick out, which they say allows highly accurate classification.
Beyond coronavirus, the system was tested with adenovirus, influenza A, and influenza B. "By combining single-particle nanopore sensing with AI, we were able to achieve highly accurate identification of multiple viral species," said Makusu Tsutsui of Osaka University, the study's senior author.
The team thinks coronaviruses may be especially well suited to the approach. Their spiky outer proteins could even allow different strains to be classified separately, the researchers believe. Tomoji Kawai, another co-author, said the work will help with development of a virus test kit that outperforms conventional viral inspection methods.
The researchers also compared the method with other rapid tests, such as PCR and antibody-based screens. They say the new approach is much faster and does not require costly reagents. They believe it could lead to improved diagnostics for emerging viruses that cause infectious diseases like COVID-19 and influenza.
What the report describes is a laboratory demonstration, with the speed and cost advantages stated by the researchers themselves. Whether it performs in clinical settings, and whether it becomes a usable test kit, remains to be shown.
Bemerkenswerte Zitate
By combining single-particle nanopore sensing with AI, we were able to achieve highly accurate identification of multiple viral species.— Makusu Tsutsui, Osaka University, study senior author
This work will help with the development of a virus test kit that outperforms conventional viral inspection methods.— Tomoji Kawai, study co-author
Möchten Sie die ganze Geschichte?
Das Original lesen bei Devdiscourse ↗
Health & Medicine