Development of an Automated Online Flow Cytometry Method to Quantify Cell Density and Fingerprint Bacterial Communities

Cell density is an important factor in all microbiome research, where interactions are of interest. It is also the most important parameter for the operation and control of most biotechnological processes. In the past, cell density determination was often performed offline and manually, resulting in a delay between sampling and immediate data processing, preventing quick action. While there are now some online methods for rapid and automated cell density determination, they are unable to distinguish between the different cell types in bacterial communities. To address this gap, an online automated flow cytometry procedure is proposed for real-time high-resolution analysis of bacterial communities. On the one hand, it allows for the online automated calculation of cell concentrations and, on the other, for the differentiation between different cell subsets of a bacterial community. To achieve this, the OC-300 automation device (onCyt Microbiology, Zürich, Switzerland) was coupled with the flow cytometer CytoFLEX (Beckman Coulter, Brea, USA). The OC-300 performs the automatic sampling, dilution, fixation and 4',6-diamidino-2-phenylindole (DAPI) staining of a bacterial sample before sending it to the CytoFLEX for measurement. It is demonstrated that this method can reproducibly measure both cell density and fingerprint-like patterns of bacterial communities, generating suitable data for powerful automated data analysis and interpretation pipelines. In particular, the automated, high-resolution partitioning of clustered data into cell subsets opens up the possibility of correlation analysis to identify the operational or abiotic/biotic causes of community disturbances or state changes, which can influence the interaction potential of organisms in microbiomes or even affect the performance of individual organisms.

Stagnation leads to short-term fluctuations in the effluent water quality of biofilters: A problem for greywater reuse?

A key characteristic of decentralized greywater treatment and reuse is high variability in both nutrient concentrations and flow. This variability in flow leads to stagnant water in the system and causes short-term fluctuations in the effluent water quality. Automated monitoring tools provide data to understand the mechanisms underlying the dynamics and to adapt control strategies accordingly. We investigated the fluctuations in a building-scale greywater treatment system comprising a membrane bioreactor followed by a biological activated carbon filter. Short-term dynamics in the effluent of the biological activated carbon filter were monitored with automated flow cytometry and turbidity, and the impact of these fluctuations on various hygiene-relevant parameters in the reuse water was evaluated. Continuous biofilm detachment into the stagnant water in the biological activated carbon filter led to temporarily increased turbidity and cell concentrations in the effluent after periods of stagnation. The fluctuations in cell concentrations were consistent with a model assuming higher detachment rates during flow than during times with stagnant water. For this system, total cell concentration and turbidity were strongly correlated. We also showed that the observed increase in cell concentration was not related to either an increase of organic carbon concentration or the concentration of two opportunistic pathogens, P. aeruginosa and L. pneumophila. Our findings demonstrate that turbidity measurements are sensitive to changes in the effluent water quality and can be used to monitor the fluctuations caused by intermittent flow. Intermittent flow did not lead to an increase in opportunistic pathogens, and this study provides no indications that stagnant water in biological activated carbon filters need be prevented.

The value of human data annotation for machine learning based anomaly detection in environmental systems

Anomaly detection is the process of identifying unexpected data samples in datasets. Automated anomaly detection is either performed using supervised machine learning models, which require a labelled dataset for their calibration, or unsupervised models, which do not require labels. While academic research has produced a vast array of tools and machine learning models for automated anomaly detection, the research community focused on environmental systems still lacks a comparative analysis that is simultaneously comprehensive, objective, and systematic. This knowledge gap is addressed for the first time in this study, where 15 different supervised and unsupervised anomaly detection models are evaluated on 5 different environmental datasets from engineered and natural aquatic systems. To this end, anomaly detection performance, labelling efforts, as well as the impact of model and algorithm tuning are taken into account. As a result, our analysis reveals the relative strengths and weaknesses of the different approaches in an objective manner without bias for any particular paradigm in machine learning. Most importantly, our results show that expert-based data annotation is extremely valuable for anomaly detection based on machine learning.

Automated Online Flow Cytometry Advances Microalgal Ecosystem Management as in situ, High-Temporal Resolution Monitoring Tool

Microalgae are emerging as a next-generation biotechnological production system in the pharmaceutical, biofuel, and food domain. The economization of microalgal biorefineries remains a main target, where culture contamination and prokaryotic upsurge are main bottlenecks to impair culture stability, reproducibility, and consequently productivity. Automated online flow cytometry (FCM) is gaining momentum as bioprocess optimization tool, as it allows for spatial and temporal landscaping, real-time investigations of rapid microbial processes, and the assessment of intrinsic cell features. So far, automated online FCM has not been applied to microalgal ecosystems but poses a powerful technology for improving the feasibility of microalgal feedstock production through in situ, real-time, high-temporal resolution monitoring. The study lays the foundations for an application of automated online FCM implying far-reaching applications to impel and facilitate the implementation of innovations targeting at microalgal bioprocesses optimization. It shows that emissions collected on the FL1/FL3 fluorescent channels, harnessing nucleic acid staining and chlorophyll autofluorescence, enable a simultaneous assessment (quantitative and diversity-related) of prokaryotes and industrially relevant phototrophic Chlorella vulgaris in mixed ecosystems of different complexity over a broad concentration range (2.2–1,002.4 cells ⋅μL^–1). Automated online FCM combined with data analysis relying on phenotypic fingerprinting poses a powerful tool for quantitative and diversity-related population dynamics monitoring. Quantitative data assessment showed that prokaryotic growth phases in engineered and natural ecosystems were characterized by different growth speeds and distinct peaks. Diversity-related population monitoring based on phenotypic fingerprinting indicated that prokaryotic upsurge in mixed cultures was governed by the dominance of single prokaryotic species. Automated online FCM is a powerful tool for microalgal bioprocess optimization owing to its adaptability to myriad phenotypic assays and its compatibility with various cultivation systems. This allows advancing bioprocesses associated with both microalgal biomass and compound production. Hence, automated online FCM poses a viable tool with applications across multiple domains within the biobased sector relying on single cell–based value chains.

Online analysis: Deeper insights into water quality dynamics in spring water

We have studied the dynamics of water quality in three karst springs taking advantage of new technological developments that enable high-resolution measurements of bacterial load (total cell concentration: TCC) as well as online measurements of abiotic parameters. We developed a novel data analysis approach, using self-organizing maps and non-linear projection methods, to approximate the TCC dynamics using the multivariate data sets of abiotic parameter time-series, thus providing a method that could be implemented in an online water quality management system for water suppliers. The (TCC) data, obtained over several months, provided a good basis to study the microbiological dynamics in detail. Alongside the TCC measurements, online abiotic parameter time-series, including spring discharge, turbidity, spectral absorption coefficient at 254nm (SAC254) and electrical conductivity, were obtained. High-density sampling over an extended period of time, i.e. every 45min for 3months, allowed a detailed analysis of the dynamics in karst spring water quality. Substantial increases in both the TCC and the abiotic parameters followed precipitation events in the catchment area. Differences between the parameter fluctuations were only apparent when analyzed at a high temporal scale. Spring discharge was always the first to react to precipitation events in the catchment area. Lag times between the onset of precipitation and a change in discharge varied between 0.2 and 6.7h, depending on the spring and event. TCC mostly reacted second or approximately concurrent with turbidity and SAC254, whereby the fastest observed reaction in the TCC time series occurred after 2.3h. The methodological approach described here enables a better understanding of bacterial dynamics in karst springs, which can be used to estimate risks and management options to avoid contamination of the drinking water.

Evaluating Monitoring Strategies to Detect Precipitation-Induced Microbial Contamination Events in Karstic Springs Used for Drinking Water

Monitoring of microbial drinking water quality is a key component for ensuring safety and understanding risk, but conventional monitoring strategies are typically based on low sampling frequencies (e.g., quarterly or monthly). This is of concern because many drinking water sources, such as karstic springs are often subject to changes in bacterial concentrations on much shorter time scales (e.g., hours to days), for example after precipitation events. Microbial contamination events are crucial from a risk assessment perspective and should therefore be targeted by monitoring strategies to establish both the frequency of their occurrence and the magnitude of bacterial peak concentrations. In this study we used monitoring data from two specific karstic springs. We assessed the performance of conventional monitoring based on historical records and tested a number of alternative strategies based on a high-resolution data set of bacterial concentrations in spring water collected with online flow cytometry (FCM). We quantified the effect of increasing sampling frequency and found that for the specific case studied, at least bi-weekly sampling would be needed to detect precipitation events with a probability of >90%. We then proposed an optimized monitoring strategy with three targeted samples per event, triggered by precipitation measurements. This approach is more effective and efficient than simply increasing overall sampling frequency. It would enable the water utility to (1) analyze any relevant event and (2) limit median underestimation of peak concentrations to approximately 10%. We conclude with a generalized perspective on sampling optimization and argue that the assessment of short-term dynamics causing microbial peak loads initially requires increased sampling/analysis efforts, but can be optimized subsequently to account for limited resources. This offers water utilities and public health authorities systematic ways to evaluate and optimize their current monitoring strategies.

Laboratory-Scale Simulation and Real-Time Tracking of a Microbial Contamination Event and Subsequent Shock-Chlorination in Drinking Water

Rapid contamination of drinking water in distribution and storage systems can occur due to pressure drop, backflow, cross-connections, accidents, and bio-terrorism. Small volumes of a concentrated contaminant (e.g., wastewater) can contaminate large volumes of water in a very short time with potentially severe negative health impacts. The technical limitations of conventional, cultivation-based microbial detection methods neither allow for timely detection of such contaminations, nor for the real-time monitoring of subsequent emergency remediation measures (e.g., shock-chlorination). Here we applied a newly developed continuous, ultra high-frequency flow cytometry approach to track a rapid pollution event and subsequent disinfection of drinking water in an 80-min laboratory scale simulation. We quantified total (TCC) and intact (ICC) cell concentrations as well as flow cytometric fingerprints in parallel in real-time with two different staining methods. The ingress of wastewater was detectable almost immediately (i.e., after 0.6% volume change), significantly changing TCC, ICC, and the flow cytometric fingerprint. Shock chlorination was rapid and detected in real time, causing membrane damage in the vast majority of bacteria (i.e., drop of ICC from more than 380 cells μl^-1 to less than 30 cells μl^-1 within 4 min). Both of these effects as well as the final wash-in of fresh tap water followed calculated predictions well. Detailed and highly quantitative tracking of microbial dynamics at very short time scales and for different characteristics (e.g., concentration, membrane integrity) is feasible. This opens up multiple possibilities for targeted investigation of a myriad of bacterial short-term dynamics (e.g., disinfection, growth, detachment, operational changes) both in laboratory-scale research and full-scale system investigations in practice.

Flow cytometric bacterial cell counts challenge conventional heterotrophic plate counts for routine microbiological drinking water monitoring

Drinking water utilities and researchers continue to rely on the century-old heterotrophic plate counts (HPC) method for routine assessment of general microbiological water quality. Bacterial cell counting with flow cytometry (FCM) is one of a number of alternative methods that challenge this status quo and provide an opportunity for improved water quality monitoring. After more than a decade of application in drinking water research, FCM methodology is optimised and established for routine application, supported by a considerable amount of data from multiple full-scale studies. Bacterial cell concentrations obtained by FCM enable quantification of the entire bacterial community instead of the minute fraction of cultivable bacteria detected with HPC (typically < 1% of all bacteria). FCM measurements are reproducible with relative standard deviations below 3% and can be available within 15 min of samples arriving in the laboratory. High throughput sample processing and complete automation are feasible and FCM analysis is arguably less expensive than HPC when measuring more than 15 water samples per day, depending on the laboratory and selected staining procedure(s). Moreover, many studies have shown FCM total (TCC) and intact (ICC) cell concentrations to be reliable and robust process variables, responsive to changes in the bacterial abundance and relevant for characterising and monitoring drinking water treatment and distribution systems. The purpose of this critical review is to initiate a constructive discussion on whether FCM could replace HPC in routine water quality monitoring. We argue that FCM provides a faster, more descriptive and more representative quantification of bacterial abundance in drinking water.

Short-term microbial dynamics in a drinking water plant treating groundwater with occasional high microbial loads

Short-term fluctuations in bacterial concentrations in drinking water systems, occurring on time scales of hours-to-weeks, are essentially unexplored due to a lack of microbial monitoring tools that allow high frequency measurements. Here, we applied fully automated online flow cytometry to measure the total cell concentrations (TCC) in both raw water (karstic groundwater) and treated water (flocculation - ultrafiltration (UF) - ozonation - granular active carbon (GAC) filtration) during a period of 70 days at high temporal resolution (n > 4000 for both water types). We detected and characterized in considerable detail aperiodic fluctuations in the raw water following regional precipitation, with TCC increasing up to 50-fold from a dry weather baseline of approximately 120 cells μl^-1 to an event peak of > 5000 cells μl^-1. Moreover, we observed the buffering of the treatment plant against these fluctuations, but in addition we recorded a completely unexpected periodic fluctuation of TCC in the treated water after GAC filtration. We concluded that the latter was the result of fluctuating water abstraction from the treatment plant reservoir by two connected water utilities, which resulted in variations in water throughput in the plant. This in turn influenced bacterial detachment and dilution in the GAC filter. This study provides strong evidence of multiple different microbial dynamics occurring in a drinking water treatment system. Given numerous possible sources of natural and operational fluctuations in raw water and drinking water treatment plants, such microbial fluctuations should be expected in many systems. The high-frequency monitoring approach presented herein can improve the understanding and eventual mitigation of such fluctuations.

Online flow cytometry reveals microbial dynamics influenced by concurrent natural and operational events in groundwater used for drinking water treatment

Detailed measurements of physical, chemical and biological dynamics in groundwater are key to understanding the important processes in place and their influence on water quality – particularly when used for drinking water. Measuring temporal bacterial dynamics at high frequency is challenging due to the limitations in automation of sampling and detection of the conventional, cultivation-based microbial methods. In this study, fully automated online flow cytometry was applied in a groundwater system for the first time in order to monitor microbial dynamics in a groundwater extraction well. Measurements of bacterial concentrations every 15 minutes during 14 days revealed both aperiodic and periodic dynamics that could not be detected previously, resulting in total cell concentration (TCC) fluctuations between 120 and 280 cells μL⁻¹. The aperiodic dynamic was linked to river water contamination following precipitation events, while the (diurnal) periodic dynamic was attributed to changes in hydrological conditions as a consequence of intermittent groundwater extraction. Based on the high number of measurements, the two patterns could be disentangled and quantified separately. This study i) increases the understanding of system performance, ii) helps to optimize monitoring strategies, and iii) opens the possibility for more sophisticated (quantitative) microbial risk assessment of drinking water treatment systems.

The feasibility of automated online flow cytometry for in-situ monitoring of microbial dynamics in aquatic ecosystems

Fluorescent staining coupled with flow cytometry (FCM) is often used for the monitoring, quantification and characterization of bacteria in engineered and environmental aquatic ecosystems including seawater, freshwater, drinking water, wastewater, and industrial bioreactors. However, infrequent grab sampling hampers accurate characterization and subsequent understanding of microbial dynamics in all of these ecosystems. A logic technological progression is high throughput and full automation of the sampling, staining, measurement, and data analysis steps. Here we assess the feasibility and applicability of automated FCM by means of actual data sets produced with prototype instrumentation. As proof-of-concept we demonstrate examples of microbial dynamics in (i) flowing tap water from a municipal drinking water supply network and (ii) river water from a small creek subject to two rainfall events. In both cases, automated measurements were done at 15-min intervals during 12-14 consecutive days, yielding more than 1000 individual data points for each ecosystem. The extensive data sets derived from the automated measurements allowed for the establishment of baseline data for each ecosystem, as well as for the recognition of daily variations and specific events that would most likely be missed (or miss-characterized) by infrequent sampling. In addition, the online FCM data from the river water was combined and correlated with online measurements of abiotic parameters, showing considerable potential for a better understanding of cause-and-effect relationships in aquatic ecosystems. Although several challenges remain, the successful operation of an automated online FCM system and the basic interpretation of the resulting data sets represent a breakthrough toward the eventual establishment of fully automated online microbiological monitoring technologies.