Dazzo, F.B., G. Tang, G. Zhu, C. Gross, D. Nasr, C. Passmore, K. Kulek, E. Polone, A. Squartini, A. Prabhu, C. Reddy, R. Peretz, L. Gao, R. Bollempalli, D. Trione, E. Marshall, J. Wang, M. Li, D. McGarrell, S. Gantner, J. Liu, and Y. Yanni.
Presented at the All Scientist Poster Reception (2006-05-09 )
A major challenge in microbial ecology is to develop reliable methods of computer-assisted microscopy that can analyze digital images of complex microbial communities at single cell resolution, and compute useful ecological characteristics of their organization and structure in situ without cultivation. To address this challenge, our team of microbiologists, mathematicians and computer scientists have been developing image analysis software that can extract the full information content in digital images of actively growing microbial communities, thereby strengthening microscopy-based methods for understanding microbial ecology. Our system, called CMEIAS (Center for Microbial Ecology Image Analysis System) contains image analysis and object classification plug-ins for UTHSCSA ImageTool running in Win NT4 / 2K / XPpro; stand-alone programs for color image segmentation, creating image quadrats, and cluster analysis of object feature measurements; action sequences to edit images of microbial communities in Adobe Photoshop; Excel addins to process and statistically analyze image analysis data, plus user manual documents, example images and interactive training tutorials.Our first release version (v1.27) of CMEIAS analyzes images, reports object counts, sizes and shapes, plus performs a supervised classification of all major plus several minor microbial morphotypes (equivalent to 98% of the genera in Bergey’s Manual 9th ed.) in 14-dimensional space at frequencies of > 0.1% with 97% accuracy, and is available for free download at our CMEIAS website <http://cme.msu.edu/cmeias/>. A CMEIAS ver. 1.28 upgrade is scheduled for release in Fall ‘06 and will feature new user support files including a fully-featured Help file, updated user manual and expanded training tutorials that have undergone major revision based on feedback from an international team of beta-testers.The CMEIAS v3.0 upgrade is at various stages of development. Its main structure is designed to extract 5 major ecologically relevant types of information from microbial communities in digital images: (i) classification of morphological diversity, using v1.27 algorithms combined with up to 42 additional measurement attributes of cell size, shape, luminance, and spatial distribution; (ii) microbial abundance, using cell size, counts, dilution adjusted concentration, hyphal length, cumulative biovolume, biomass C, biosurface area of the entire community or of each individual morphotype within the community; (iii) metabolic/physiological activity using differential color staining, (iv) autecology/phylogeny using color recognition with fluorescent molecular probes, and (v) spatial distribution, using numerous plotless, plot-based and geostatistical analysis parameters. CMEIAS output data are typically exported into MS Excel, EcoStat and GS+ Geostatistics to compute various ecological statistics that further characterize microbial community structure.Recent developments for CMEIAS v3.0 include:Four image processing tools to help prepare images for analysis. A stand-alone CMEIAS Color Segmentation program interactively samples color pixels of the foreground microbes of interest, then finds each cell’s boundary, and finally creates a new RGB “object detected” image containing these colored microbes in a noise-free background (96.7% accuracy). Its primary use will be to facilitate segmentation of foreground microbes of interest within RGB digital images where color differentiation really counts, e.g., immunofluorescence, FISH, activity/viability stains, Gram stains, etc. CMEIAS Action Palette Sequences operate in Adobe Photoshop / Image Processing ToolKit to semi-automate the editing of foreground and background pixels in phase contrast, fluorescence and SEM grayscale images of microbial communities (97.1% accuracy). A CMEIAS Object Separation Plugin operating in ImageTool automatically splits touching microbial cells within thresholded images (95.8% accuracy). A stand-alone CMEIAS Quadrat Maker application is being built to facilitate the optimizing of grid dimensions that divide an image into smaller, constant size quadrats for spatial sampling of microbial density, and then prepares images of each quadrat for plot-based and geostatistical spatial distribution analyses.Five statistical computation tools to process and analyze CMEIAS data extracted from images. Three of these tools operate within MS Excel® and the other two are stand-alone applications. The CMEIAS Action Performance Macro computes performance accuracy for newly developed action sequences by analyzing object analysis and morphotype classification data extracted from the corresponding edited images of microbes. The CMEIAS Data Preparation Macro compiles and concatenates CMEIAS object analysis and classification data, computes descriptive statistics, and automatically prepares the input tables of data for further analysis. The CMEIAS Data Analysis Addin performs frequency distribution analyses with user-selected bin parameters on object analysis data, analyzes object classification datasets built from multiple images to determine if the sample size is sufficient to estimate morphotype diversity, computes and plots a wide variety of diversity indices and similarity coefficients to evaluate and compare community structures, performs statistical inference tests, and analyzes plot-less, plot-based, and georeferenced data for spatial distribution analysis. Two stand-alone statistical analysis applications have been developed to perform cluster analysis on CMEIAS object analysis data in order to build the size border files that define the diversity of Operational Morphological Units recognized by the CMEIAS-3 classifier. The CMEIAS Upper Size Border tool performs a Monte Carlo cluster analysis to optimize the upper class limits for the user-defined size border file that is tailored for diversity analysis of the specific community being examined. The flexible design built into the user-defined size border file and these cluster analysis applications expands the ability of the CMEIAS-3 OMU classifier to recognize an almost unlimited range of diversity optimized for the community under investigation. The CMEIAS Hi-Low Size Border tool performs cluster analyses on population data extracted by CMEIAS and from Bergey’s Manual database to build the general default size border file used for a wide variety of communities (not custom-tailored for any specific community).Additional notables in CMEIAS v.3.0: Eight out of 18 different Microbial Biovolume formulas have been identified as most accurate and implemented to measure this cell abundance parameter for all microbial morphotypes recognized by CMEIAS. A CMEIAS Cluster Index has been introduced to measure the proximity of aggregated cells surrounding each individual microbe as a Z variate for geostatistical spatial distribution analyses of colonization behavior. Several new CMEIAS measurement features of object analysis (Empirical Distribution, % Microbial Cover, Spatial Randomness, 1st and 2nd Nearest Neighbor Distances) have been added to analyze spatial distribution patterns of microbial colonization in situ. Mean Radius, Maximum Radius, and Aerial Porosity measurement feature attributes have been added to analyze microbial biofilm structure/development/function. A CMEIAS Script has been built to report the area of the AOI polygon for cumulative object analysis. Also a feature has been added to the CMEIAS Cumulative Object Analyzer that can assign the Cartesian coordinates for quadrat image posting centers that are weighted by the local density of objects within it or randomized for spatial analysis.Recent collaborative projects using CMEIAS include in situ analysis of the autecological biogeography of candidate biofertilizer inoculant strains of rhizobia that promote rice growth, microbial community analyses of corn leaf surfaces (normal vs. genetically engineered BT-corn), distribution and abundance of active communities within soil aggregates, seasonal succession of epilithic microbial biofilm communities in streambed rocks of Japanese freshwater streams, diversity of bacterioplankton in pristine Italian Alpine lakes, measurements of the in situ calling distances and spatial gradients for quorum sensing cell-to-cell communication during microbial colonization of roots, and the perturbations in microbial community structure associated with bacterial vaginosis.In summary, CMEIAS-based applications have potential for filling major gaps in our understanding of microbial ecology by providing accurate, robust and user-friendly computing tools that can extract ecologically important, quantitative information from digital images in situ at spatial scales relevant to the microbes themselves. It adds a powerful new dimension to examining microbial communities, and is especially valuable when combined with molecular-based and other methods of polyphasic analysis. These tools are admirably suited for microbial community analyses currently under investigation in this LTER program focused on row agriculture. Announcements of progress in CMEIAS development and new releases are provided at our website.
Back to meeting | Show |