By Rob Dyer, UVA ChemSciComm

Bacteria and other germs remain a constant threat to our health and well-being, even as the science about infectious diseases continues to improve. Andreas Gahlmann and co-workers in the Department of Chemistry at the University of Virginia are taking strides to better understand how bacteria behave inside the human body by developing advanced imaging methods to observe critical host-pathogen interfaces in real-time. Tracking individual bacterial cells in complex environments has been a particular challenge, because bacteria are extremely small and high-resolution microscopy is generally not compatible with live cells.

In a recent publication, the Gahlmann lab shared a new approach called Bacterial Cell Morphometry 3D (BCM3D). BCM3D first utilizes a state-of-the-art fluorescence microscopy technique which is compatible with live cells, called light sheet microscopy. In a typical experiment, a linearly polarized light beam is stretched into a 2-D sheet using lenses and passed through a sample, while fluorescence images are taken from above. The microscopy data from this technique is then combined with deep learning convolutional neural networks (CNN), a computational method for data analysis which mimics how the brain learns and processes information. This integrated experimental-computational workflow enables researchers to visualize how multicellular bacterial communities operate in different environments.

The convolutional neural networks of BCM3D were shown to be able to classify individual bacterial cells according to their shapes. Prof. Gahlmann thinks that this technological breakthrough will open a window into the microscale interactions between bacterial populations in real time. “When living on and inside the human body, bacteria have to overcome a range of environmental challenges, such as attacks by the host’s immune cells or lengthy antibiotic drug treatments. In order to combat bacterial infections more efficiently, we need to better understand how different bacteria respond and adapt to these stressful situations.” 

A full article from the group and be found here:

Nature Communications 11, Article number: 6151 (2020)