Antler-backed Singapore startup develops 60-second Covid-19 breath test
Singapore-based startup Breathonix announced that it has developed a breath test that can detect Covid-19 within a minute.

Breathonix CEO Zhunan Jia (left) and chief operating officer Fang Du / Photo credit: Breathonix
The company says that the test has achieved more than 90% accuracy in a clinical trial involving 180 patients. It also detected with a 95% specificity, which means it could identify those without the disease in 95 out of 100 cases.
Breathonix, which was spun off from the National University of Singapore, has a track record in detecting lung cancer and tuberculosis by analyzing volatile molecules in exhaled breath.
The company is backed by early-stage venture capital firm Antler, which in April invested a total of up to US$500,000 in startups that are tackling the Covid-19 outbreak.
Breathonix’s Covid-19 test detects volatile organic compounds (VOCs) in a person’s breath. An individual taking the test needs to blow into a disposable mouthpiece that catches the breath sample and feeds it into a spectrometer, a scientific instrument used to separate and measure components. The collected compounds are then analyzed by an in-built machine learning algorithm, and the result is generated in less than a minute, according to a statement.
VOCs are consistently produced by various biochemical reactions in human cells, Breathonix CEO Zhunan Jia explained. Different diseases cause specific changes to the compounds, resulting in detectable changes in a person’s breath profile.
Fang Du, the firm’s chief operating officer, added that the disposable mouthpiece has a one-way valve and a saliva trap. The latter prevents inhalation and any saliva from entering the machine, which he says “makes cross-contamination unlikely.”
Current Covid-19 detection methods include a swab test, which can be uncomfortable, and polymerase chain reaction tests, which could take a few hours.
Editing by Collin Furtado and Eileen C. Ang
(And yes, we’re serious about ethics and transparency. More information here.)




