Quantitative image analysis of microbial communities with BiofilmQ.
Hartmann R, Jeckel H, Jelli E, Singh PK, Vaidya S, Bayer M, Rode DKH, Vidakovic L, Díaz-Pascual F, Fong JCN, Dragoš A, Lamprecht O, Thöming JG, Netter N, Häussler S, Nadell CD, Sourjik V, Kovács ÁT, Yildiz FH, Drescher K. · 2021
Read Quantitative image analysis of microbial communities with BiofilmQ. by Hartmann R, Jeckel H, Jelli E, Singh PK, Vaidya S, Bayer M, Rode DKH, Vidakovic L, Díaz-Pascual F, Fong JCN, Dragoš A, Lamprecht O, Thöming JG, Netter N, Häussler S, Nadell CD, Sourjik V, Kovács ÁT, Yildiz FH, Drescher K. (2021) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc7840502.txt.
Introduction
Hartmann and co-authors present BiofilmQ, a tool for measuring microbial communities in images. Biofilms are a hugely common form of microbe life. Their members differ across space and over time. Microscopy is the main way to watch them. Numbers must be pulled from those images with care.
BiofilmQ turns microscope images into hard data. It splits images into cells or patches for measurement. Traits can then be tracked across the whole community. Growth patterns and local differences become numbers. The tool serves labs that study biofilm structure. It is built for shared, repeatable analysis. Such numbers help explain how biofilms work and resist treatment.
Catalog record
| CREATOR | Hartmann R, Jeckel H, Jelli E, Singh PK, Vaidya S, Bayer M, Rode DKH, Vidakovic L, Díaz-Pascual F, Fong JCN, Dragoš A, Lamprecht O, Thöming JG, Netter N, Häussler S, Nadell CD, Sourjik V, Kovács ÁT, Yildiz FH, Drescher K. |
|---|---|
| YEAR | 2021 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quantitative image analysis of microbial communities with BiofilmQ By Raimo Hartmann, Hannah Jeckel, Eric Jelli, Praveen K Singh, Sanika Vaidya, Miriam Bayer, Daniel K H Rode, Lucia Vidakovic, Francisco Díaz-Pascual, Jiunn C N Fong, Anna Dragoš, Olga Lamprecht Published in: Nature Microbiology DOI: 10.1038/s41564-020-00817-4 ABSTRACT Abstract Biofilms are microbial communities that represent a highly abundant form of microbial life on Earth. Inside biofilms, phenotypic and genotypic variations occur in three-dimensional space and time; microscopy and quantitative image analysis are therefore crucial for elucidating their functions. Here, we present BiofilmQ—a comprehensive image cytometry software tool for the automated and high-throughput quantification, analysis and visualization of numerous biofilm-internal and whole-biofilm properties in three-dimensional space and time. Subject terms: Biofilms, Microbial communities, Computational platforms and environments, Image processing MAIN Spatially structured microbial communities display spatial gradients of nutrients and other diffusible molecular compounds, as well as spatiotemporal variation in species composition and cellular differentiation1–4. For biofilm phenotyping, as well as for characterizing phenotypes of particular cells within biofilms, it is critical to be able to perform image-based quantitative measurements of fluorescent reporters and structural features for particular regions inside three-dimensional (3D) biofilms. Extracting the desirable information from 3D images relies on non-trivial automated image analysis. The most widely used tool for biofilm image analysis in the literature is COMSTAT5,6, which provided one of the first tools to objectively determine differences in biofilm morphology. COMSTAT and the alternative software tools for biofilm image analysis7–9 have been tremendously important for biofilm research by providing parameters for 3D phenotyping, yet they are not designed for analysing biofilm internal properties with spatial resolution, and they do not include the functionality to visualize data. Image analysis tools developed for microbial ecology10–12 have the ability to measure alternative parameters, including, morphology analysis10 and 3D correlation functions12. The design of biofilm research projects and the discovery of new biofilm behaviours are presently limited by the lack of modern cytometry software tools that can quantify a comprehensive set of spatially and temporally resolved structural parameters and fluorescent reporters inside 3D biofilms, and visualize the resulting data. Powerful and user-friendly tools that have recently enabled quantitative analyses for bacterial cell biology for two-dimensional (2D) images13–16. Inspired by these tools, we have integrated algorithms for image analysis of the internal properties of 3D microbial communities with data analysis and data visualization capabilities in the form of a software tool, BiofilmQ (https://drescherlab.org/data/biofilmQ), which provides a graphical user interface and requires no knowledge of programming. BiofilmQ is built on the basis of standard image analysis techniques, as well as new algorithms for image cytometry and object tracking, of which technical descriptions are provided in the online documentation. Extensive documentation and video tutorials guide users through each step in the image analysis, data analysis and data visualization workflow. Here we describe the concept, capabilities and limitations of BiofilmQ and demonstrate its usefulness for the quantitative characterization of microbial communities. BiofilmQ is designed for analysing fluorescence images of a wide variety of spatially structured microbial communities and growth geometries, including microscopic, mesoscopic, macroscopic colonies and biofilms on surfaces, and free-floating aggregates as well as communities in the context of eukaryotic hosts. Microbial communities can be analysed irrespective of the size, growth geometry, morphology, species or the number of fluorescence channels (Fig. 1a). The only requirement for BiofilmQ is that the software must be able to identify the biovolume of the biofilm using one fluorescence channel or using an imported segmentation. Biofilm biovolume detection is an example of semantic segmentation in image analysis17, and different segmentation algorithms have received considerable attention in the biofilm literature18–24, as the segmentation quality can have a large impact on the analysis results. To perform accurate biofilm segmentation for a wide variety of image types and signal levels (Fig. 1a), BiofilmQ includes the following three different segmentation options: (1) automatic segmentation using classical algorithms, such as Otsu25, Ridler–Calvard26, robust background or maximum correlation thresholding27; (2) semi-manual thresholding supported by immediate visual feedback; and (3) import of presegmented images into BiofilmQ. If users choose to import presegmented images, we recommend general-purpose segmentation tools (such as ilasik28) or convolutional neural networks (such as U-Net29), which can be trained for particular image types, fluorescence levels or biofilm morphologies to give very high segmentation accuracy. After the segmentation of the biofilm biovolume, we recommend visual inspection of the segmentation accuracy, which is displayed by BiofilmQ. The automated and semi-manual segmentation options that are provided yielded a good segmentation accuracy for the different types of biofilm images that are shown in Figs. 1 and 2. Thus, the main focus of BiofilmQ is the cytometry, data analysis and data visualization after the segmentation step. [Figure: Quantification of microbial community properties with spatial resolution in BiofilmQ.a, Examples of different biofilm image categories that can be analysed using BiofilmQ: E. coli macrocolony, V. cholerae meso- and microcolonies, Bacillus subtilis pellicle and floating aggregates, V. cholerae biofilm (yellow) on mouse intestinal villi (blue). Many file formats are supported, based on the Bio-Formats toolbox36. Scale bars, 1 mm (top left), 30 µm (middle and bottom left, and top right), 40 µm (middle right), 100 µm (bottom right). b, The BiofilmQ image processing pipeline for a 3D biofilm image. The raw fluorescence image is filtered and thresholded to obtain a binary representation of the biofilm. These 3D binary data are then dissected into cubes of a user-defined size, and the cubes are then used to quantify the biofilm properties. Alternatively, binary images of the biofilm or single cells can be imported. Here, each cube in the biofilm is coloured according to the local biovolume density, which is one of the cube properties that can be extracted. c,d, Many parameters can be quantified for each cube (c) and for the whole biofilm (d). e, A B. subtilis microcolony (left) on an agar pad consisting of two strains (a strain constitutively expressing superfolder green fluorescent protein (sfGFP) and a strain constitutively expressing mKate2) was analysed using cube-based image cytometry, single-cell image cytometry and flow cytometry. The image cytometry results are qualitatively similar, and the flow cytometry results also show two additional cell populations (non-fluorescent cells and cells with fluorescence in both channels). The data shown are one example out of n = 3 experiments, which all showed the same qualitative results. Scale bar, 10 µm. f, For an E. coli macrocolony (left) on agar consisting of two strains (a strain constitutively expressing sfGFP and a strain constitutively expressing mRuby3), we compared the results from cube-based image cytometry and flow cytometry. The data shown are one example out of n = 3 experiments, which all showed the same qualitative results. Scale bar, 500 µm. a.u., arbitrary units.] [Figure: Applications of spatial image cytometry for characterizing biofilm biology.a, By analysing biofilm spatiotemporal development of whole V. cholerae microcolonies in flow chambers, the transcription of key matrix biosynthesis genes was correlated with matrix localization and the cell density structure inside biofilms. Top left, the space–time kymographs of reporters for rbmA, rbmC and bap1 transcription were normalized to the signal of constitutively expressed sfGFP. Representative of n = 3 independent biofilms for each reporter. Top right, for biofilms grown to a particular timepoint (15 h), the correlation of the spatial