Cell tracking is a key computational task in live-cell microscopy, but fully automated analysis of high-throughput imaging requires reliable and, thus, uncertainty-aware data analysis tools, as the amount of data recorded within a single experiment exceeds what humans are able to overlook. We here propose and benchmark various methods to reason about and quantify uncertainty in linear assignment-based cell tracking algorithms. Our methods take inspiration from statistics and machine learning, leveraging two perspectives on the cell tracking problem explored throughout this work: Considering it as a Bayesian inference problem and as a classification problem. Our methods admit a framework-like character in that they equip any frame-to-frame tracking method with uncertainty quantification. We demonstrate this by applying it to various existing tracking algorithms including the recently presented Transformer-based trackers. We demonstrate empirically that our methods yield useful and well-calibrated tracking uncertainties.

SUMMARY: 13C-based metabolic flux analysis is a cornerstone of quantitative systems biology, yet its increasing data complexity and methodological diversity place high demands on simulation software. We introduce 13CFLUX(v3), a third-generation simulation platform that combines a high-performance C++ engine with a convenient Python interface. The software delivers substantial performance gains across isotopically stationary and nonstationary analysis workflows, while remaining flexible to accommodate diverse labeling strategies and analytical platforms. Its open-source availability facilitates seamless integration into computational ecosystems and community-driven extension. By supporting multi-experiment integration, multi-tracer studies, and advanced statistical inference such as Bayesian analysis, 13CFLUX provides a robust and extensible framework for modern fluxomics research. AVAILABILITY AND IMPLEMENTATION: Sources and containers are provided at https://jugit.fz-juelich.de/IBG-1/ModSim/Fluxomics/13CFLUX, and scripts to replicate results in the supplementary data at https://github.com/JuBiotech/Supplement-to-Stratmann-et-al.-Bioinformatics-2025.

SUMMARY: Effective collaboration between developers of Bayesian inference methods and users is key to advance our quantitative understanding of biosystems. We here present hopsy, a versatile open-source platform designed to provide convenient access to powerful Markov chain Monte Carlo sampling algorithms tailored to models defined on convex polytopes (CP). Based on the high-performance C++ sampling library HOPS, hopsy inherits its strengths and extends its functionalities with the accessibility of the Python programming language. A versatile plugin-mechanism enables seamless integration with domain-specific models, providing method developers with a framework for testing, benchmarking, and distributing CP samplers to approach real-world inference tasks. We showcase hopsy by solving common and newly composed domain-specific sampling problems, highlighting important design choices. By likening hopsy to a marketplace, we emphasize its role in bringing together users and developers, where users get access to state-of-the-art methods, and developers contribute their own innovative solutions for challenging domain-specific inference problems. AVAILABILITY AND IMPLEMENTATION: Sources, documentation and a continuously updated list of sampling algorithms are available at https://jugit.fz-juelich.de/IBG-1/ModSim/hopsy, with Linux, Windows and MacOS binaries at https://pypi.org/project/hopsy/.

Thinning is a sub-sampling technique to reduce the memory footprint of Markov chain Monte Carlo. Despite being commonly used, thinning is rarely considered efficient. For sampling constraint-based models, a highly relevant use-case in systems biology, we here demonstrate that thinning boosts computational and, thereby, sampling efficiencies of the widely used Coordinate Hit-and-Run with Rounding (CHRR) algorithm. By benchmarking CHRR with thinning with simplices and genome-scale metabolic networks of up to thousands of dimensions, we find a substantial increase in computational efficiency compared to unthinned CHRR, in our examples by orders of magnitude, as measured by the effective sample size per time (ESS/t), with performance gains growing with polytope (effective network) dimension. Using a set of benchmark models we derive a ready-to-apply guideline for tuning thinning to efficient and effective use of compute resources without requiring additional coding effort. Our guideline is validated using three (out-of-sample) large-scale networks and we show that it allows sampling convex polytopes uniformly to convergence in a fraction of time, thereby unlocking the rigorous investigation of hitherto intractable models. The derivation of our guideline is explained in detail, allowing future researchers to update it as needed as new model classes and more training data becomes available. CHRR with deliberate utilization of thinning thereby paves the way to keep pace with progressing model sizes derived with the constraint-based reconstruction and analysis (COBRA) tool set. Sampling and evaluation pipelines are available at https://jugit.fz-juelich.de/IBG-1/ModSim/fluxomics/chrrt.

Cell tracking is an essential tool in live-cell imaging to determine single-cell features, such as division patterns or elongation rates. Unlike in common multiple object tracking, in microbial live-cell experiments cells are growing, moving, and dividing over time, to form cell colonies that are densely packed in mono-layer structures. With increasing cell numbers, following the precise cell-cell associations correctly over many generations becomes more and more challenging, due to the massively increasing number of possible associations. To tackle this challenge, we propose a fast parameter-free cell tracking approach, which consists of activity-prioritized nearest neighbor assignment of growing (expanding) cells and a combinatorial solver that assigns splitting mother cells to their daughters. As input for the tracking, Omnipose is utilized for instance segmentation. Unlike conventional nearest-neighbor-based tracking approaches, the assignment steps of our proposed method are based on a Gaussian activity-based metric, predicting the cell-specific migration probability, thereby limiting the number of erroneous assignments. In addition to being a building block for cell tracking, the proposed activity map is a standalone tracking-free metric for indicating cell activity. Finally, we perform a quantitative analysis of the tracking accuracy for different frame rates, to inform life scientists about a suitable (in terms of tracking performance) choice of the frame rate for their cultivation experiments, when cell tracks are the desired key outcome.

SUMMARY: Random flux sampling is a powerful tool for the constraint-based analysis of metabolic networks. The most efficient sampling method relies on a rounding transform of the constraint polytope, but no available rounding implementation can round all relevant models. By removing redundant polytope constraints on the go, PolyRound simplifies the numerical problem and rounds all the 108 models in the BiGG database without parameter tuning, compared to ∼50% for the state-of-the-art implementation. AVAILABILITY AND IMPLEMENTATION: The implementation is available on gitlab: https://gitlab.com/csb.ethz/PolyRound. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

SUMMARY: The C++ library Highly Optimized Polytope Sampling (HOPS) provides implementations of efficient and scalable algorithms for sampling convex-constrained models that are equipped with arbitrary target functions. For uniform sampling, substantial performance gains were achieved compared to the state-of-the-art. The ease of integration and utility of non-uniform sampling is showcased in a Bayesian inference setting, demonstrating how HOPS interoperates with third-party software. AVAILABILITY AND IMPLEMENTATION: Source code is available at https://github.com/modsim/hops/, tested on Linux and MS Windows, includes unit tests, detailed documentation, example applications and a Dockerfile. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: The validity of model based inference, as used in systems biology, depends on the underlying model formulation. Often, a vast number of competing models is available, that are built on different assumptions, all consistent with the existing knowledge about the studied biological phenomenon. As a remedy for this, Bayesian Model Averaging (BMA) facilitates parameter and structural inferences based on multiple models simultaneously. However, in fields where a vast number of alternative, high-dimensional and non-linear models are involved, the BMA-based inference task is computationally very challenging. RESULTS: Here we use BMA in the complex setting of Metabolic Flux Analysis (MFA) to infer whether potentially reversible reactions proceed uni- or bidirectionally, using 13C labeling data and metabolic networks. BMA is applied on a large set of candidate models with differing directionality settings, using a tailored multi-model Markov Chain Monte Carlo (MCMC) approach. The applicability of our algorithm is shown by inferring the in vivo probability of reaction bidirectionalities in a realistic network setup, thereby extending the scope of 13C MFA from parameter to structural inference. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Microfluidic platforms for live-cell analysis are in dire need of automated image analysis pipelines. In this context, producing reliable tracks of single cells in colonies has proven to be notoriously difficult without manual assistance, especially when image sequences experience low frame rates. RESULTS: With Uncertainty-Aware Tracking (UAT), we propose a novel probabilistic tracking paradigm for simultaneous tracking and estimation of tracking-induced errors in biological quantities derived from live-cell experiments. To boost tracking accuracy, UAT relies on a Bayesian approach which exploits temporal information on growth patterns to guide the formation of lineage hypotheses. A biological study is presented, in which UAT demonstrates its ability to track cells, with comparable to better accuracy than state-of-the-art trackers, while simultaneously estimating tracking-induced errors. AVAILABILITY AND IMPLEMENTATION: Image sequences and Java executables for reproducing the results are available at https://doi.org/10.5281/zenodo.1299526. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Streptomycetes are filamentous microorganisms of high biotechnological relevance, especially for the production of antibiotics. In submerged cultures, the productivity of these microorganisms is closely linked to their growth morphology. Microfluidic lab-on-a-chip cultivation systems, coupled with automated time-lapse imaging, generate spatio-temporal insights into the mycelium development of streptomycetes, therewith extending the biotechnological toolset by spatio-temporal screening under well-controlled and reproducible conditions. However, the analysis of the complex mycelial structure formation is limited by the extent of manual interventions required during processing of the acquired high-volume image data. These interventions typically lead to high evaluation times and, therewith, limit the analytic throughput and exploitation of microfluidic-based screenings. RESULTS: We present the tool mycelyso (MYCElium anaLYsis SOftware), an image analysis system tailored to fully automated hyphae-level processing of image stacks generated by time-lapse microscopy. With mycelyso, the developing hyphal streptomycete network is automatically segmented and tracked over the cultivation period. Versatile key growth parameters such as mycelium network structure, its development over time, and tip growth rates are extracted. Results are presented in the web-based exploration tool mycelyso Inspector, allowing for user friendly quality control and downstream evaluation of the extracted information. In addition, 2D and 3D visualizations show temporal tracking for detailed inspection of morphological growth behaviors. For ease of getting started with mycelyso, bundled Windows packages as well as Docker images along with tutorial videos are available. CONCLUSION: mycelyso is a well-documented, platform-independent open source toolkit for the automated end-to-end analysis of Streptomyces image stacks. The batch-analysis mode facilitates the rapid and reproducible processing of large microfluidic screenings, and easy extraction of morphological parameters. The objective evaluation of image stacks is possible by reproducible evaluation workflows, useful to unravel correlations between morphological, molecular and process parameters at the hyphae- and mycelium-levels with statistical power.

Science revolves around the best way of conducting an experiment to obtain insightful results. Experiments with maximal information content can be found by computational experimental design (ED) strategies that identify optimal conditions under which to perform the experiment. Several criteria have been proposed to measure the information content, each emphasizing different aspects of the design goal, i.e., reduction of uncertainty. Where experiments are complex or expensive, second sight is at the budget governing the achievable amount of information. In this context, the design objectives cost and information gain are often incommensurable, though dependent. By casting the ED task into a multiple-criteria optimization problem, a set of trade-off designs is derived that approximates the Pareto-frontier which is instrumental for exploring preferable designs. In this work, we present a computational methodology for multiple-criteria ED of information-rich experiments that accounts for virtually any set of design criteria. The methodology is implemented for the case of 13C metabolic flux analysis (MFA), which is arguably the most expensive type among the 'omics' technologies, featuring dozens of design parameters (tracer composition, analytical platform, measurement selection etc.). Supported by an innovative visualization scheme, we demonstrate with two realistic showcases that the use of multiple criteria reveals deep insights into the conflicting interplay between information carriers and cost factors that are not amendable to single-objective ED. For instance, tandem mass spectrometry turns out as best-in-class with respect to information gain, while it delivers this information quality cheaper than the other, routinely applied analytical technologies. Therewith, our Pareto approach to ED offers the investigator great flexibilities in the conception phase of a study to balance costs and benefits.

MOTIVATION: Single cell time-lapse microscopy is a powerful method for investigating heterogeneous cell behavior. Advances in microfluidic lab-on-a-chip technologies and live-cell imaging render the parallel observation of the development of individual cells in hundreds of populations possible. While image analysis tools are available for cell detection and tracking, biologists are still confronted with the challenge of exploring and evaluating this data. RESULTS: We present the software tool Vizardous that assists scientists with explorative analysis and interpretation tasks of single cell data in an interactive, configurable and visual way. With Vizardous, lineage tree drawings can be augmented with various, time-resolved cellular characteristics. Associated statistical moments bridge the gap between single cell and the population-average level. AVAILABILITY AND IMPLEMENTATION: The software, including documentation and examples, is available as executable Java archive as well as in source form at https://github.com/modsim/vizardous. CONTACT: [email protected]. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: The precise quantification of intracellular metabolic flow rates is of fundamental importance in bio(techno)logy and medical research. The gold standard in the field is metabolic flux analysis (MFA) with 13C-labeling experiments. 13C-MFA workflows orchestrate several, mainly human-in-the-loop, software applications, integrating them with plenty of heterogeneous information. In practice, this had posed a major practical barrier for evaluating, interpreting and understanding isotopic data from carbon labeling experiments. RESULTS: Graphical modeling, interactive model exploration and visual data analysis are the key to overcome this limitation. We have developed a first-of-its-kind graphical tool suite providing scientists with an integrated software framework for all aspects of 13C-MFA. Almost 30 modules (plug-ins) have been implemented for the Omix visualization software. Several advanced graphical workflows and ergonomic user interfaces support major domain-specific modeling and proofreading tasks. With that, the graphical suite is a productivity enhancing tool and an original educational training instrument supporting the adoption of 13C-MFA applications in all life science fields. AVAILABILITY: The Omix Light Edition is freely available at http://www.omix-visualization.com CONTACT: [email protected], [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

SUMMARY: (13)C-based metabolic flux analysis ((13)C-MFA) is the state-of-the-art method to quantitatively determine in vivo metabolic reaction rates in microorganisms. 13CFLUX2 contains all tools for composing flexible computational (13)C-MFA workflows to design and evaluate carbon labeling experiments. A specially developed XML language, FluxML, highly efficient data structures and simulation algorithms achieve a maximum of performance and effectiveness. Support of multicore CPUs, as well as compute clusters, enables scalable investigations. 13CFLUX2 outperforms existing tools in terms of universality, flexibility and built-in features. Therewith, 13CFLUX2 paves the way for next-generation high-resolution (13)C-MFA applications on the large scale. AVAILABILITY AND IMPLEMENTATION: 13CFLUX2 is implemented in C++ (ISO/IEC 14882 standard) with Java and Python add-ons to run under Linux/Unix. A demo version and binaries are available at www.13cflux.net.

The MapReduce pattern popularized by Google has successfully been utilized in several scientific applications. In this paper, it is investigated whether a MapReduce approach utilizing on-demand resources from a Cloud is beneficial to perform simulation tasks in the area of Systems Biology and whether it can be seamlessly integrated into a service-oriented scientific workflow framework. In particular, an Amazon Elastic Map Reduce Cloud implementation of the 13C-MFA (Metabolix Flux Analysis) Monte Carlo bootstrap approach aimed at the integration into an existing BPEL-based scientific workflow system is presented. A comparison of a 64 node MapReduce cluster with a single node computation approach reveals a total performance gain up to a factor of 14, with a total cost for on-demand resources of $11. The most critical factor in terms of performance is I/O, i.e. our application suffers from the fact that I/O operations on many small files are expensive using Amazon S3 and the Hadoop DFS.

BACKGROUND: Stimulus Response Experiments to unravel the regulatory properties of metabolic networks are becoming more and more popular. However, their ability to determine enzyme kinetic parameters has proven to be limited with the presently available data. In metabolic flux analysis, the use of 13C labeled substrates together with isotopomer modeling solved the problem of underdetermined networks and increased the accuracy of flux estimations significantly. RESULTS: In this contribution, the idea of increasing the information content of the dynamic experiment by adding 13C labeling is analyzed. For this purpose a small example network is studied by simulation and statistical methods. Different scenarios regarding available measurements are analyzed and compared to a non-labeled reference experiment. Sensitivity analysis revealed a specific influence of the kinetic parameters on the labeling measurements. Statistical methods based on parameter sensitivities and different measurement models are applied to assess the information gain of the labeled stimulus response experiment. CONCLUSION: It was found that the use of a (specifically) labeled substrate will significantly increase the parameter estimation accuracy. An overall information gain of about a factor of six is observed for the example network. The information gain is achieved from the specific influence of the kinetic parameters towards the labeling measurements. This also leads to a significant decrease in correlation of the kinetic parameters compared to an experiment without 13C-labeled substrate.

BACKGROUND: Metabolic Flux Analysis (MFA) based on isotope labeling experiments (ILEs) is a widely established tool for determining fluxes in metabolic pathways. Isotope labeling networks (ILNs) contain all essential information required to describe the flow of labeled material in an ILE. Whereas recent experimental progress paves the way for high-throughput MFA, large network investigations and exact statistical methods, these developments are still limited by the poor performance of computational routines used for the evaluation and design of ILEs. In this context, the global analysis of ILN topology turns out to be a clue for realizing large speedup factors in all required computational procedures. RESULTS: With a strong focus on the speedup of algorithms the topology of ILNs is investigated using graph theoretic concepts and algorithms. A rigorous determination of all cyclic and isomorphic subnetworks, accompanied by the global analysis of ILN connectivity is performed. Particularly, it is proven that ILNs always brake up into a large number of small strongly connected components (SCCs) and, moreover, there are natural isomorphisms between many of these SCCs. All presented techniques are universal, i.e. they do not require special assumptions on the network structure, bidirectionality of fluxes, measurement configuration, or label input. The general results are exemplified with a practically relevant metabolic network which describes the central metabolism of E. coli comprising 10390 isotopomer pools. CONCLUSION: Exploiting the topological features of ILNs leads to a significant speedup of all universal algorithms for ILE evaluation. It is proven in theory and exemplified with the E. coli example that a speedup factor of about 1000 compared to standard algorithms is achieved. This widely opens the door for new high performance algorithms suitable for high throughput applications and large ILNs. Moreover, for the first time the global topological analysis of ILNs allows to comprehensively describe and understand the general patterns of label flow in complex networks. This is an invaluable tool for the structural design of new experiments and the interpretation of measured data.