Abstract

Bacteria, and microbes in general, are valuable tools in synthetic and systems biology research. Their genetic flexibility, quick replication, and metabolic engineering capabilities make them crucial for scientific exploration. In the realm of microbial systems analysis, microscopy techniques have proven instrumental, as they facilitate real-time monitoring of live microbial cells, providing essential insights into these dynamic systems.

Quantitative microbiology using microscopy holds transformative potential for synthetic and systems biology, yet poses significant challenges in data acquisition and processing. This seminar will explore the advancements made by the Smart Microscopy Lab in the Department of Engineering in overcoming these challenges through an interdisciplinary approach integrating microfluidics, machine learning, and advanced imaging techniques.

We will introduce the utility of the mother machine microfluidic system for capturing high-fidelity, high throughput, temporal data and discuss the limitations of traditional methods. We delve into the issues that arise in quantitative analyses due to diffraction-limited microscopy, and make the case for machine learning-based solutions. At the heart of the new methods being developed is SyMBac—Synthetic Micrographs of Bacteria—a novel tool that leverages synthetic training data for highly precise segmentation of experimental micrographs. Future directions such as adaptive and dynamic segmentation strategies using pseudo labels and timelapse synthetic data are highlighted. Finally, the seminar will elaborate on the broader capabilities of smart microscopy, including its role in uncovering and mitigating imaging artefacts, benchmarking new analytical models, and super-resolving data in both spatial and temporal dimensions.

The seminar will be held in the JDB Seminar Room, Department of Engineering, and online (zoom): https://newnham.zoom.us/j/92544958528?pwd=YS9PcGRnbXBOcStBdStNb3E0SHN1UT09