How one professor identifies the organisms in the air you breathe

A person inhales about 10,000 liters of air every day, according to Professor Stephan Schuster, PhD, research director at the Singapore Centre on Environmental Life Sciences Engineering (SCELSE) at Nanyang Technological University. That daily intake of air can include a million microorganisms, bacteria, and fungi, some of which can be harmful to human health.¹

How do we know what’s in the air?

In 2014, Schuster’s lab received a significant, multiyear grant to develop a method they call “ultra-low biomass sequencing.” Using a military-grade air sampler about the size and shape of a soup can, they collect air samples at a flow rate of 300 liters per minute or less.

To sequence the samples, they need a fraction of a nanogram (<0.5ng) of DNA, where a typical researcher might aim for sequencing 100 nanograms. Schuster’s team can sequence less DNA than you would need for PCR.

They use an Illumina NextSeq 2000 Sequencing System. “There’s no other sequencing platform we’ve used that can sequence a 300 base pair read with a Q40 value,” says Schuster. The value, Q40, is a quality score indicating 99.99% accuracy.

Even though the samples are ultra-low biomass, the data is “very, very robust,” says Schuster, who has shown this via replication. They used a proprietary software pipeline that compares DNA reads to various sequence databases. Based on sequence homology, they were able to identify tens of thousands of organisms with high taxonomic accuracy.

The software also classifies the confidence of a species taxonomic identification with a red, yellow, or green indication. Schuster credits Illumina sequencing capabilities for being able to identify and classify with high certainty. “To make a claim like that is based on how accurate you can sequence and how accurate you can classify. If you tried to do this using a raw read generated by a long-read sequencing instrument, the error rate would be outside the species boundaries for any of these microbial organisms.”

By overlapping forward and reverse reads generated by the NextSeq 2000 300bp XLEAP-SBS chemistry, Schuster and team are demonstrating that metagenomics can now achieve coverage and confidence levels comparable to high coverage reference genomes of individual microbial taxa. This jump in sequencing quality of mixed DNA samples addresses long-standing challenges in the field (such as doubt that every microorganism had been captured, sequenced, and identified), making metagenomic findings more reliable and impactful at a very high samples throughput.

“When Illumina introduced the XLEAP-SBS chemistry, it was a drastic, drastic jump in the read quality,” says Schuster. “What you see is absolutely stunning.”

The coverage for complete genomes from air can be up to 99.8%, even when sequencing fungal taxa. Ascomycota and Basidiomycota fungi are especially important for agriculture, with most common genome sizes between 10 and 40 million base pairs. Still, these large genomes can get 40x coverage of the entire organism’s genome captured from an air sample.

Out of thin air
What does “clean air” look like? For a flagship study, Schuster’s team sampled air at a 3,500‑meter research station on the Jungfrau Joch, near the Aletsch Glacier in the Swiss Alps. The remote, high‑altitude location helped them characterize baseline air composition, providing a reference for what the atmosphere can look like with minimal influence from human settlements.

Using sequencing to monitor microbial air quality signals an exciting paradigm shift. It is now possible to detect human and environmental pathogens that may describe what is affecting or will affect human and environmental health.

SCELSE’s work was especially useful in 2020. “When the pandemic happened, the US military had few air samplers. We were the only ones in the world that had 40 units for immediate deployment,” says Schuster. “The head of my institute came to me and said, ‘You have waited for this pandemic all your life, haven’t you?’”

With his nearly $1 million fleet of samplers, the team went into enormous community isolation facilities where thousands of people were quarantined. Schuster remembers watching masked individuals pushing elevator buttons with toothpicks. But using the air sampling technology as well as doing surface swabs, his team was able to report to the Singapore government that positively detecting coronavirus in the air was ten times higher than from a surface (fomite).

More recently, the SCELSE team went into 140 homes in Singapore for different clinical cohorts—residents with asthma, bronchiectasis, COPD, and more. Their analysis showed that the microbiomes present in the homes varied by the patients’ conditions.

While a handful of labs may be performing similar research, they are not sequencing at the scale that Schuster is. Schuster has sampled and sequenced the air around the earth’s equator, and along another transect starting from Ireland, going through Europe to the Middle East, India, Southeast Asia, and all the way to New Zealand and Antarctica. His global survey of the air accounts for 1.6 terabytes of sequence data.

Schuster envisions using metagenomics in hospitals, airports, childcare centers, and food production facilities. He would be able to test for contagious viruses or bacteria that might infect product or people. But whereas an organism that is being cultured has a very slow growth rate, Schuster can report results within 24 hours using various means of nucleic acid detection.

Continuing to discover
Have you ever heard of Alloprevotella rava? Or Chelativorans alearense? Schuster hasn’t, either—as well as most, if not all, experts. (The latter is a bacterium, by the way, isolated in soil in places like southern France and South Korea, albeit in super low abundance.)

Schuster and team have created the largest metagenome data set from 10,000 air samples collected and processed using the standardized ultra-low biomass sequencing approach, and paired it with their AI-assisted metagenomic alignment software platform. The result is a searchable database of more than 22,000 air microbiome species. The platform annotates species using large language model algorithms and public data. It tells the user about the appearance, lifestyle, habitat, location, abilities, related research, and the medical relevance of every organism in a given sample. The software provides citations from PubMed articles and abstracts. Some lesser-known species have only one or two peer-reviewed papers, while more prominent taxa need an AI algorithm to identify the most relevant ones.

“The air is the most misunderstood global environment. Even the top experts in environmental microbiology do not know about the organisms that we find.” For example, one renowned environmental microbiologist could recognize just four of the 100 most abundant organisms they found in a given sample.    

Schuster and team recently received a very large government grant to study and characterize the environment using his ultra-low biomass sequencing techniques.

“Metagenomics is an analysis challenge. We have solved the challenge,” he says, smiling. “We can tell you, so to speak, whether or not a microorganism can fly in the atmosphere.”

¹Prussin, A. J. II & Marr, L. C. (2015). Sources of airborne microorganisms in the built environment. Microbiome, 3, 78.

M-GL-04985