New Features of CMEIAS Innovative Software for Computer-Assisted Microscopy of Microorganisms and their Ecology

Reddy, C., J. Liu, M. Wadekar, A. Prabhu, D. Trione, E. Marshall, J. Zurdo, F-I. Liu, J. Urbance, and F. B. Dazzo

Presented at the All Scientist Meeting (2002-10-04 )

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. We are addressing this challenge by developing a new generation of interactive software that can extract the full information content in digital images of actively growing microbial communities. Our system, called CMEIAS (Center for Microbial Ecology Image Analysis System) consists of plug-ins for free ImageTool software running in Win NT4.0, a stand-alone color recognition program, and various Excel macros for data management and analysis. CMEIAS v1.27 can automatically classify bacterial morphotypes at 0.1% frequency with 97% accuracy. A website for CMEIAS is under construction for free release of v1.27 program files along with an operator manual. An upgraded CMEIAS v3.0 is being developed to extract 5 ecologically relevant parameters of microbial communities in situ in digital images: (i) classify their morphological diversity at a complexity level equivalent to 98% of the genera described in Bergey’s Manual of Determinative Bacteriology, (ii) measure the abundance for both filamentous and non-filamentous microbes based on their cell density, biovolume and biomass carbon, (iii) detect their metabolic activity based on image color recognition, (iv) analyze their autecology and phylogeny based on image color recognition using molecular probes, and (v) analyze their spatial patterns of colonization and dispersion at single-cell resolution. CMEIAS data can also be exported into EcoStat and GS+ Geostatistics to compute numerous other indices and spatial autocorrelation models that further characterize microbial community structure. Here we report on four new developments for CMEIAS v3.0. First, we have introduced an Object Separation tool that automatically splits touching cells for object analysis and operates with an overall performance accuracy of 95.8%. Second, we have developed a Color Recognition program that segments foreground bacterial objects of interest in RGB color images. This program uses interactive sampling of up to 200 color pixels of the foreground microbes of interest, then applies a region-growing algorithm to find their cell boundary, and finally creates a new RGB image with these colored foreground microbes against a noise-free background. This new program will likely revolutionize the quantitative image analysis of bacteria differentiated by color in RGB images. Third, we introduce a Sampling Statistics macro that analyzes the appropriate sample size for community morphotype analysis by computing the tolerance envelope surrounding the existing data of morphotype diversity at a user-defined percentage confidence level. Fourth, we have evaluated 15 different formulas to compute microbial biovolumes from 2-dimensional projected images, and have identified / implemented the 8 most accurate formulas for all morphotypes recognized by CMEIAS. Our vision is for CMEIAS to become an accurate, robust and user-friendly software tool that can significantly enhance the analysis of whole community samples without cultivation, thereby creating many new applications for quantitative studies of in situ microbial ecology that were very difficult or almost impossible to do before (e.g., at spatial scales relevant to microbe-microbe interactions).

Back to meeting | Show |
Sign In