New prospects for routine microbiota analyses with flow cytometry
The group of NCCR Microbiomes’ Director Jan van der Meer has developed CellCognize, a machine-learning algorithm that assesses bacterial cell type diversity in samples run through a flow cytometer – in just 5 minutes.
In a study published on July 15, the group demonstrates that CellCognize can be trained with known bacterial standards to achieve a prediction accuracy of 80%. The algorithm quantifies absolute cell counts and detects temporal changes in an undetermined bacterial community exposed to chemicals.
Rapid flow cytometry data analysis is a useful complement to DNA-sequencing and other omics techniques, comparatively slower and more laborious. CellCognize enables fast routine analysis of microbiome diversity, with applications ranging from clinical and veterinary interventions to engineered microbial systems.The lead author of the study, Dr. Birge Özel Duygan, is now a grantee of the technology transfer office UNIL-CHUV. Her start-up will apply CellCognize to rapidly analyze human gut microbiota in clinical settings. She will be hosted in the lab of Gilbert Greub (NCCR Microbiomes) at the CHUV.