Microbiology

With multispectral infrared imaging, ADMIR offers a new instrument to characterize microbiological samples. Successfully showcased with a first application, microbial identification, our benchtop imager is now ready for all types of colonies and biofilms analysis.

Introducing BioMIR

Infrared imaging for microbiological identification
Microbiology

Label-free

Fast

Non-destructive

Technique and assets

Spectral imaging
Infrared spectroscopy is rapidly imposing itself as a next-generation tool for strain typing and bacterial analysis. It functions by probing absorbance biomolecules in the infrared wavelengths, which gives insights to the chemical composition of a sample. Here, we expand this approach in two dimensions, gathering both spectral and morphological data in a single instrument.

Tailored for microbiology
Our first prototype has been developed with microbiology in mind. Its lasers were carefuly selected to probe wavelengths relevant of spectral variations between microbial samples. This allows it to discriminate species – or even strains.

Microbiology
Infrared spectra of a few common microbiological strains (top), and example of spectral imaging on a colony of E. coli (bottom).
Microbiology
An example of microbial culture on a filtration membrane. Said membrane can be removed from culture medium for analysis, without altering colonies.

Label-free
Infrared absorbance is an intrinsic response that requiers no biomarker, nor selective or chromogenic culture medium to gather data. A single culture on a general medium is sufficient.

Non-destructive
Because water absorbs most of infrared light, protocols for infrared spectroscopy generally involves picking and drying colonies before analysis. Here, we use filtration membranes to solve this issue. These membranes can be used as growth support for colonies, and are transparent enough for our lasers, meaning they can also act as imaging support. Since colonies are tstill intact, they can be placed back into culture, or used for further analysis such as antibiotic susceptibility testing.

Fast
Within 5 to 10 minutes, all colonies of a given culture are analyzed and identified.

Artificial intelligence
Our analysis toolkit includes various deep learning algorithms trained on thousands of infrared fingerprints. On a proof-of-concept study, these algorithms correctly classified more than 3.300 colonies from waterborn bacteria with 97% of confidence.

Microbiology
Multispectral fingerprints of colonies formed by various bacterial strains.

Further reading

Multispectral infrared colony phenotyping for high-throughput microbiological control of waters

Joel Le Galudec, Mathieu Dupoy, Joris Baraillon, Boris Taurel, Laurent Duraffourg, Pierre R. Marcoux (2025). Under review.

To assess microbiological quality, water samples are filtered through porous membranes. These membranes are then placed on solid agar, which allows normal colonies growth. After 24h, membranes can be removed and analyzed using BioMIR. Within 5 to 10 minutes, all colonies are identified.

Read full article
White paper

In a few pages, get a full overview of the context of our work and the ambitions of ADMIR regarding assesment of microbiological water quality

microbiology_wp_french.pdf