Accuracy of Biovolume Formulas for CMEIAS Computer-Assisted Microscopy and Body Size Analysis of Morphologically Diverse Microbial Populations and Communities

摘要

Cell biovolume is a commonly used metric of microbial abundance analyzed by computer-assisted microscopy, but the accuracies of most biovolume formulas have not been validated by ground truth data. We examined the accuracy of 17 biovolume formulas by comparing the computed volumes of 3D models representing 11 microbial morphotypes (cocci, spirals, curved rods, U-shaped rods, regular straight rods, unbranched filaments, ellipsoids, clubs, prosthecates, rudimentary branched rods, and branched filaments) to the volume displacement of the same objects as ground truth. As anticipated, formula accuracy was significantly influenced by the morphotype examined. A few formulas performed very accurately (> 95 %), especially those that adapted to the cell's shape, whereas others were consistently inaccurate or only accurate for one or two morphotypes. As an example of application, indices of morphological diversity in a freshwater biofilm assemblage were shown to be significantly different when microbial abundance among morphotype classes was measured as biovolume body mass rather than cell counts. Spatial analysis of biovolume body mass can also provide insights on the in situ ecophysiological attributes among individuals in microbial populations and communities, including their spatially autocorrelated allometric scaling interrelationships between body size, metabolic activity, resource apportionment and use, food web dynamics, and various cell-cell interactions affecting their growth and colonization behavior within spatially structured biofilm landscapes. This improved computing technology of biovolume algorithms with proven accuracy identifies which formula(s) should be used to compute microbial biovolumes in 2D images of morphologically diverse communities acquired by conventional phase-contrast light microscopy at single-cell resolution.