Tools and methods for high-throughput single-cell imaging with the mother machine

• DOI: 10.7554/elife.88463
• Publication Date: 2023-07-26T14:28:06Z
• Authors: Ryan Thiermann; Michael Sandler; Gursharan Ahir; John T Sauls; Jeremy Schroeder; Steven Brown; Guillaume Le Treut; Fangwei Si; Dongyang Li; Jue D Wang; Suckjoon Jun
• Keywords: QH301-705.5; infectious disease; Image Processing; Science; Culture; microfluidics; Mothers; Genetics and Molecular Biology; Bioengineering; Article; Computer-Assisted; image analysis; Image Processing, Computer-Assisted; Humans; Biology (General); Microbiology and Infectious Disease; Microscopy; General Immunology and Microbiology; General Neuroscience; microbiology; Q; E. coli; R; General Medicine; Research Personnel; 004; mother machine; bacterial physiology; Networking and Information Technology R&D (NITRD); General Biochemistry; Medicine; Female; Generic health relevance

Despite much progress, image processing remains a significant bottleneck for high-throughput analysis of microscopy data. One popular platform single-cell time-lapse imaging is the mother machine, which enables long-term tracking microbial cells under precisely controlled growth conditions. While several machine pipelines have been developed in past years, adoption by non-expert audience challenge. To fill this gap, we implemented our own software, MM3, as plugin multidimensional viewer napari. napari-MM3 complete and modular pipeline data, takes advantage high-level interactivity Here, give an overview test it against well-designed widely used pipelines, including BACMMAN DeLTA. Researchers often analyze data with custom scripts using varied methods, but quantitative comparison output different has lacking. end, show that key physiological parameter correlations distributions are robust to choice method. However, also find small changes thresholding parameters can systematically alter extracted from experiments. Moreover, explicitly deep learning-based segmentation, 'what you put what get' (WYPIWYG) - is, pixel-level variation training cell segmentation propagate model bias spatial temporal measurements. Finally, while primary purpose work introduce software over last decade lab, provide information those who want implement machine-based methods their research.