Statistical analysis and biological interpretation of the flow cytometric heterogeneity observed in bacterial axenic cultures

Quantifying synthetic bacterial community composition with flow cytometry: efficacy in mock communities and challenges in co-cultures

Determination of bacterial community composition in synthetic communities is critical for understanding microbial systems. The community composition is typically determined through bacterial plating or through PCR-based methods, which can be labor-intensive, expensive, or prone to bias. Simultaneously, flow cytometry has been suggested as a cheap and fast alternative. However, since the technique captures the phenotypic state of bacterial cells, accurate determination of community composition could be affected when bacteria are co-cultured. We investigated the performance of flow cytometry for quantifying oral synthetic communities and compared it to the performance of strain specific qPCR and 16S rRNA gene amplicon sequencing. Therefore, axenic cultures, mock communities and co-cultures of oral bacteria were prepared. Random forest classifiers trained on flow cytometry data of axenic cultures were used to determine the composition of the synthetic communities, as well as strain specific qPCR and 16S rRNA gene amplicon sequencing. Flow cytometry was shown to have a lower average root mean squared error and outperformed the PCR-based methods in even mock communities (flow cytometry: 0.11 ± 0.04; qPCR: 0.26 ± 0.09; amplicon sequencing: 0.15 ± 0.01). When bacteria were co-cultured, neither flow cytometry, strain-specific qPCR, nor 16S rRNA gene amplicon sequencing resulted in similar community composition. Performance of flow cytometry was decreased compared with mock communities due to changing phenotypes. Finally, discrepancies between flow cytometry and strain-specific qPCR were found. These findings highlight the challenges ahead for quantifying community composition in co-cultures by flow cytometry.IMPORTANCEQuantification of bacterial composition in synthetic communities is crucial for understanding and steering microbial interactions. Traditional approaches like plating, strain-specific qPCR, and amplicon sequencing are often labor-intensive and expensive and limit high-throughput experiments. Recently, flow cytometry has been suggested as a swift and cheap alternative for quantifying communities and has been successfully demonstrated on simple bacterial mock communities. However, since flow cytometry measures the phenotypic state of cells, measurements can be affected by differing phenotypes. Especially, changing phenotypes resulting from co-culturing bacteria can have a profound effect on the applicability of the technique in this context. This research illustrates the feasibility and challenges of flow cytometry for the determination of community structure in synthetic mock communities and co-cultures.

Fast quantification of gut bacterial species in cocultures using flow cytometry and supervised classification

A bottleneck for microbial community experiments with many samples and/or replicates is the fast quantification of individual taxon abundances, which is commonly achieved through sequencing marker genes such as the 16S rRNA gene. Here, we propose a new approach for high-throughput and high-quality enumeration of human gut bacteria in a defined community, combining flow cytometry and supervised classification to identify and quantify species mixed in silico and in defined communities in vitro. We identified species in a 5-species in silico community with an F1 score of 71%. In addition, we demonstrate in vitro that our method performs equally well or better than 16S rRNA gene sequencing in two-species cocultures and agrees with 16S rRNA gene sequencing data on the most abundant species in a four-species community. We found that shape and size differences alone are insufficient to distinguish species, and that it is thus necessary to exploit the multivariate nature of flow cytometry data. Finally, we observed that variability of flow cytometry data across replicates differs between gut bacterial species. In conclusion, the performance of supervised classification of gut species in flow cytometry data is species-dependent, but is for some combinations accurate enough to serve as a faster alternative to 16S rRNA gene sequencing.

PhenoGMM: Gaussian Mixture Modeling of Cytometry Data Quantifies Changes in Microbial Community Structure

Microorganisms are vital components in various ecosystems on Earth. In order to investigate the microbial diversity, researchers have largely relied on the analysis of 16S rRNA gene sequences from DNA.

Microbial flow cytometry can rapidly characterize the status of microbial communities. Upon measurement, large amounts of quantitative single-cell data are generated, which need to be analyzed appropriately. Cytometric fingerprinting approaches are often used for this purpose. Traditional approaches either require a manual annotation of regions of interest, do not fully consider the multivariate characteristics of the data, or result in many community-describing variables. To address these shortcomings, we propose an automated model-based fingerprinting approach based on Gaussian mixture models, which we call PhenoGMM. The method successfully quantifies changes in microbial community structure based on flow cytometry data, which can be expressed in terms of cytometric diversity. We evaluate the performance of PhenoGMM using data sets from both synthetic and natural ecosystems and compare the method with a generic binning fingerprinting approach. PhenoGMM supports the rapid and quantitative screening of microbial community structure and dynamics.

IMPORTANCE Microorganisms are vital components in various ecosystems on Earth. In order to investigate the microbial diversity, researchers have largely relied on the analysis of 16S rRNA gene sequences from DNA. Flow cytometry has been proposed as an alternative technology to characterize microbial community diversity and dynamics. The technology enables a fast measurement of optical properties of individual cells. So-called fingerprinting techniques are needed in order to describe microbial community diversity and dynamics based on flow cytometry data. In this work, we propose a more advanced fingerprinting strategy based on Gaussian mixture models. We evaluated our workflow on data sets from both synthetic and natural ecosystems, illustrating its general applicability for the analysis of microbial flow cytometry data. PhenoGMM supports a rapid and quantitative analysis of microbial community structure using flow cytometry.

Flow Cytometric Single-Cell Identification of Populations in Synthetic Bacterial Communities

Bacterial cells can be characterized in terms of their cell properties using flow cytometry. Flow cytometry is able to deliver multiparametric measurements of up to 50,000 cells per second. However, there has not yet been a thorough survey concerning the identification of the population to which bacterial single cells belong based on flow cytometry data. This paper not only aims to assess the quality of flow cytometry data when measuring bacterial populations, but also suggests an alternative approach for analyzing synthetic microbial communities. We created so-called in silico communities, which allow us to explore the possibilities of bacterial flow cytometry data using supervised machine learning techniques. We can identify single cells with an accuracy >90% for more than half of the communities consisting out of two bacterial populations. In order to assess to what extent an in silico community is representative for its synthetic counterpart, we created so-called abundance gradients, a combination of synthetic (i.e., in vitro) communities containing two bacterial populations in varying abundances. By showing that we are able to retrieve an abundance gradient using a combination of in silico communities and supervised machine learning techniques, we argue that in silico communities form a viable representation for synthetic bacterial communities, opening up new opportunities for the analysis of synthetic communities and bacterial flow cytometry data in general.

Differences in stationary-phase cells of a commercial Saccharomyces cerevisiae wine yeast grown in aerobic and microaerophilic batch cultures assessed by electric particle analysis, light diffraction and flow cytometry

We applied electric particle analysis, light diffraction and flow cytometry to obtain information on the morphological changes during the stationary phase of Saccharomyces cerevisiae. The reported analyses of S. cerevisiae populations were obtained under two different conditions, aerobic and microaerophilic, at 27°C. The samples analysed were taken at between 20 and 50 h from the beginning of culture. To assist in the interpretation of the observed distributions a complexity index was used. The aerobically grown culture reached significantly greater cell density. Under these conditions, the cell density experienced a much lower reduction (3%) compared with the microaerophilic conditions (30%). Under aerobic conditions, the mean cell size determined by both electric particle analysis and light diffraction was lower and remained similar throughout the experiment. Under microaerophilic conditions, the mean cell size determined by electric particle analysis decreased slightly as the culture progressed through the stationary phase. Forward and side scatter distributions revealed two cell subpopulations under both growth conditions. However, in the aerobic growing culture the two subpopulations were more separated and hence easier to distinguish. The distributions obtained with the three experimental techniques were analysed using the complexity index. This analysis suggested that a complexity index is a good descriptor of the changes that take place in a yeast population in the stationary phase, and that it aids in the discussion and understanding of the implications of these distributions obtained by these experimental techniques.

Skew-laplace and cell-size distribution in microbial axenic cultures: statistical assessment and biological interpretation

We report a skew-Laplace statistical analysis of both flow cytometry scatters and cell size from microbial strains primarily grown in batch cultures, others in chemostat cultures and bacterial aquatic populations. Cytometry scatters best fit the skew-Laplace distribution while cell size as assessed by an electronic particle analyzer exhibited a moderate fitting. Unlike the cultures, the aquatic bacterial communities clearly do not fit to a skew-Laplace distribution. Due to its versatile nature, the skew-Laplace distribution approach offers an easy, efficient, and powerful tool for distribution of frequency analysis in tandem with the flow cytometric cell sorting.

Quadratic form: a robust metric for quantitative comparison of flow cytometric histograms

Comparison of fluorescence distributions is a fundamental part of the analysis of flow cytometric data. This approach is applied to detect differences between control and test sample and thus analyze a biological response. Comparison of standard test samples over time provides an estimate of instrument stability for quality control. However, application of statistical methods of distribution comparison in flow cytometry is difficult owing to instrument noise and the complex shape of intensity distributions. We applied quadratic form (QF) as a mathematical metric for comparison of flow cytometry histograms. QF operates on histograms as vectors and calculates the total distance in an interbin manner using a matrix of distances between single histogram bins. Euclidean interbin distance and histograms normalized to unity were used. Critical values corresponding to 95% significance level were calculated using Monte-Carlo simulation and single-maximum Gaussian distributions populated with several numbers of events. The QF statistic was then validated for non-Gaussian single-maximum distributions and multiple-maxima distributions. We determined that the critical values for Gaussian distributions depended on standard deviations and number of events in the compared histograms. A simple empirical function was constructed to characterize this dependence. Furthermore, it was verified that critical values (corresponding to 95% significance) for non-Gaussian histograms were similar to values for the Gaussian histograms characterized by the same standard deviation. We applied the QF statistic to estimate the differences between histograms of DNA content (ploidy) in cells of old and young leaf tissue of Brassica campestris. Furthermore, we quantified differences in fluorescence intensity in immunostaining of human lymphocytes. Quadratic form (QF) provides a true (mathematical) metric for estimation of distance between flow cytometry histograms of arbitrary shape. QF can be applied as a statistical test for estimation of significance of the distance measure. The respective critical values depend only on the number of events and standard deviations of compared histograms and are not affected by distribution shape. Therefore, applications of QF do not require assumptions concerning distribution shape and can be easily implemented in practice. This notion was confirmed using empirical distributions of DNA content in plant tissue and distributions of immunofluorescence in human cells.

Analysis and IbM simulation of the stages in bacterial lag phase: basis for an updated definition

The lag phase is the initial phase of a culture that precedes exponential growth and occurs when the conditions of the culture medium differ from the pre-inoculation conditions. It is usually defined by means of cell density because the number of individuals remains approximately constant or slowly increases, and it is quantified with the lag parameter lambda. The lag phase has been studied through mathematical modelling and by means of specific experiments. In recent years, Individual-based Modelling (IbM) has provided helpful insights into lag phase studies. In this paper, the definition of lag phase is thoroughly examined. Evolution of the total biomass and the total number of bacteria during lag phase is tackled separately. The lag phase lasts until the culture reaches a maximum growth rate both in biomass and cell density. Once in the exponential phase, both rates are constant over time and equal to each other. Both evolutions are split into an initial phase and a transition phase, according to their growth rates. A population-level mathematical model is presented to describe the transitional phase in cell density. INDividual DIScrete SIMulation (INDISIM) is used to check the outcomes of this analysis. Simulations allow the separate study of the evolution of cell density and total biomass in a batch culture, they provide a depiction of different observed cases in lag evolution at the individual-cell level, and are used to test the population-level model. The results show that the geometrical lag parameter lambda is not appropriate as a universal definition for the lag phase. Moreover, the lag phase cannot be characterized by a single parameter. For the studied cases, the lag phases of both the total biomass and the population are required to fully characterize the evolution of bacterial cultures. The results presented prove once more that the lag phase is a complex process that requires a more complete definition. This will be possible only after the phenomena governing the population dynamics at an individual level of description, and occurring during the lag and exponential growth phases, are well understood.

Rapid G+ count and subpopulation assessment of the intestinal bacteria in Apodemus sylvaticus and Mus musculus by flow cytometry

We report a novel application of calcein-acetomethyl ester in flow cytometry for rapid estimation of the number of G+-bacteria in rodent feces (Apodemus sylvaticus and Mus sp.f. muridae). We also use the combined application of flow cytometry and Syto-13 or Sypro Orange staining to count rapidly the total bacterial population and to describe bacterial subpopulations in the intestine.

Skew-Laplace distribution in Gram-negative bacterial axenic cultures: new insights into intrinsic cellular heterogeneity

The application of flow cytometry and skew-Laplace statistical analysis to assess cellular heterogeneity in Gram-negative axenic cultures is reported. In particular, fit to the log-skew-Laplace distribution for cellular side scatter or ‘granulosity’ is reported, and a number of theoretical and applied issues are considered in relation to the biological significance of this fit.