BioDynaMo: a modular platform for high-performance agent-based simulation

Agent-based modeling in cancer biomedicine: applications and tools for calibration and validation

Computational models are not just appealing because they can simulate and predict the development of biological phenomena across multiple spatial and temporal scales, but also because they can integrate information from well-established in vitro and in vivo models and test new hypotheses in cancer biomedicine. Agent-based models and simulations are especially interesting candidates among computational modeling procedures in cancer research due to the capability to, for instance, recapitulate the dynamics of neoplasia and tumor - host interactions. Yet, the absence of methods to validate the consistency of the results across scales can hinder adoption by turning fine-tuned models into black boxes. This review compiles relevant literature that explores strategies to leverage high-fidelity simulations of multi-scale, or multi-level, cancer models with a focus on verification approached as simulation calibration. We consolidate our review with an outline of modern approaches for agent-based models' validation and provide an ambitious outlook toward rigorous and reliable calibration.

Machine Learning for Predicting Biologic Agent Efficacy in Ulcerative Colitis: An Analysis for Generalizability and Combination with Computational Models

Machine learning (ML) has been applied to predict the efficacy of biologic agents in ulcerative colitis (UC). ML can offer precision, personalization, efficiency, and automation. Moreover, it can improve decision support in predicting clinical outcomes. However, it faces challenges related to data quality and quantity, overfitting, generalization, and interpretability. This paper comments on two recent ML models that predict the efficacy of vedolizumab and ustekinumab in UC. Models that consider multiple pathways, multiple ethnicities, and combinations of real-world and clinical trial data are required for optimal shared decision-making and precision medicine. This paper also highlights the potential of combining ML with computational models to enhance clinical outcomes and personalized healthcare. Key Insights: (1) ML offers precision, personalization, efficiency, and decision support for predicting the efficacy of biologic agents in UC. (2) Challenging aspects in ML prediction include data quality, overfitting, and interpretability. (3) Multiple pathways, multiple ethnicities, and combinations of real-world and clinical trial data should be considered in predictive models for optimal decision-making. (4) Combining ML with computational models may improve clinical outcomes and personalized healthcare.

More Is Faster: Why Population Size Matters in Biological Search

Many biological scenarios have multiple cooperating searchers, and the timing of the initial first contact between any one of those searchers and its target is critically important. However, we are unaware of biological models that predict how long it takes for the first of many searchers to discover a target. We present a novel mathematical model that predicts initial first contact times between searchers and targets distributed at random in a volume. We compare this model with the extreme first passage time approach in physics that assumes an infinite number of searchers all initially positioned at the same location. We explore how the number of searchers, the distribution of searchers and targets, and the initial distances between searchers and targets affect initial first contact times. Given a constant density of uniformly distributed searchers and targets, the initial first contact time decreases linearly with both search volume and the number of searchers. However, given only a single target and searchers placed at the same starting location, the relationship between the initial first contact time and the number of searchers shifts from a linear decrease to a logarithmic decrease as the number of searchers grows very large. More generally, we show that initial first contact times can be dramatically faster than the average first contact times and that the initial first contact times decrease with the number of searchers, while the average search times are independent of the number of searchers. We suggest that this is an underappreciated phenomenon in biology and other collective search problems.

Mechanistic model of radiotherapy-induced lung fibrosis using coupled 3D agent-based and Monte Carlo simulations

BACKGROUND

Mechanistic modelling of normal tissue toxicities is unfolding as an alternative to the phenomenological normal tissue complication probability models. The latter, currently used in the clinics, rely exclusively on limited patient data and neglect spatial dose distribution information. Among the various approaches, agent-based models are appealing as they provide the means to include patient-specific parameters and simulate long-term effects in complex systems. However, Monte Carlo tools remain the state-of-the-art for modelling radiation transport and provide measurements of the delivered dose with unmatched precision.

METHODS

In this work, we develop and characterize a coupled 3D agent-based - Monte Carlo model that mechanistically simulates the onset of the radiation-induced lung fibrosis in an alveolar segment. To the best of our knowledge, this is the first such model.

RESULTS

Our model replicates extracellular matrix patterns, radiation-induced lung fibrosis severity indexes and functional subunits survivals that show qualitative agreement with experimental studies and are consistent with our past results. Moreover, in accordance with experimental results, higher functional subunits survival and lower radiation-induced lung fibrosis severity indexes are achieved when a 5-fractions treatment is simulated. Finally, the model shows increased sensitivity to more uniform protons dose distributions with respect to more heterogeneous ones from photon irradiation.

CONCLUSIONS

This study lays thus the groundwork for further investigating the effects of different radiotherapeutic treatments on the onset of radiation-induced lung fibrosis via mechanistic modelling.

Is it selfish to be filamentous in biofilms? Individual-based modeling links microbial growth strategies with morphology using the new and modular iDynoMiCS 2.0

Microbial communities are found in all habitable environments and often occur in assemblages with self-organized spatial structures developing over time. This complexity can only be understood, predicted, and managed by combining experiments with mathematical modeling. Individual-based models are particularly suited if individual heterogeneity, local interactions, and adaptive behavior are of interest. Here we present the completely overhauled software platform, the individual-based Dynamics of Microbial Communities Simulator, iDynoMiCS 2.0, which enables researchers to specify a range of different models without having to program. Key new features and improvements are: (1) Substantially enhanced ease of use (graphical user interface, editor for model specification, unit conversions, data analysis and visualization and more). (2) Increased performance and scalability enabling simulations of up to 10 million agents in 3D biofilms. (3) Kinetics can be specified with any arithmetic function. (4) Agent properties can be assembled from orthogonal modules for pick and mix flexibility. (5) Force-based mechanical interaction framework enabling attractive forces and non-spherical agent morphologies as an alternative to the shoving algorithm. The new iDynoMiCS 2.0 has undergone intensive testing, from unit tests to a suite of increasingly complex numerical tests and the standard Benchmark 3 based on nitrifying biofilms. A second test case was based on the "biofilms promote altruism" study previously implemented in BacSim because competition outcomes are highly sensitive to the developing spatial structures due to positive feedback between cooperative individuals. We extended this case study by adding morphology to find that (i) filamentous bacteria outcompete spherical bacteria regardless of growth strategy and (ii) non-cooperating filaments outcompete cooperating filaments because filaments can escape the stronger competition between themselves. In conclusion, the new substantially improved iDynoMiCS 2.0 joins a growing number of platforms for individual-based modeling of microbial communities with specific advantages and disadvantages that we discuss, giving users a wider choice.

Digitize your Biology! Modeling multicellular systems through interpretable cell behavior

Cells are fundamental units of life, constantly interacting and evolving as dynamical systems. While recent spatial multi-omics can quantitate individual cells' characteristics and regulatory programs, forecasting their evolution ultimately requires mathematical modeling. We develop a conceptual framework-a cell behavior hypothesis grammar-that uses natural language statements (cell rules) to create mathematical models. This allows us to systematically integrate biological knowledge and multi-omics data to make them computable. We can then perform virtual "thought experiments" that challenge and extend our understanding of multicellular systems, and ultimately generate new testable hypotheses. In this paper, we motivate and describe the grammar, provide a reference implementation, and demonstrate its potential through a series of examples in tumor biology and immunotherapy. Altogether, this approach provides a bridge between biological, clinical, and systems biology researchers for mathematical modeling of biological systems at scale, allowing the community to extrapolate from single-cell characterization to emergent multicellular behavior.

BMX: Biological modelling and interface exchange

High performance computing has a great potential to provide a range of significant benefits for investigating biological systems. These systems often present large modelling problems with many coupled subsystems, such as when studying colonies of bacteria cells. The aim to understand cell colonies has generated substantial interest as they can have strong economic and societal impacts through their roles in in industrial bioreactors and complex community structures, called biofilms, found in clinical settings. Investigating these communities through realistic models can rapidly exceed the capabilities of current serial software. Here, we introduce BMX, a software system developed for the high performance modelling of large cell communities by utilising GPU acceleration. BMX builds upon the AMRex adaptive mesh refinement package to efficiently model cell colony formation under realistic laboratory conditions. Using simple test scenarios with varying nutrient availability, we show that BMX is capable of correctly reproducing observed behavior of bacterial colonies on realistic time scales demonstrating a potential application of high performance computing to colony modelling. The open source software is available from the zenodo repository https://doi.org/10.5281/zenodo.8084270 under the BSD-2-Clause licence.

Efficient simulation of neural development using shared memory parallelization

The Neural Development Simulator, NeuroDevSim, is a Python module that simulates the most important aspects of brain development: morphological growth, migration, and pruning. It uses an agent-based modeling approach inherited from the NeuroMaC software. Each cycle has agents called fronts execute model-specific code. In the case of a growing dendritic or axonal front, this will be a choice between extension, branching, or growth termination. Somatic fronts can migrate to new positions and any front can be retracted to prune parts of neurons. Collision detection prevents new or migrating fronts from overlapping with existing ones. NeuroDevSim is a multi-core program that uses an innovative shared memory approach to achieve parallel processing without messaging. We demonstrate linear strong parallel scaling up to 96 cores for large models and have run these successfully on 128 cores. Most of the shared memory parallelism is achieved without memory locking. Instead, cores have only write privileges to private sections of arrays, while being able to read the entire shared array. Memory conflicts are avoided by a coding rule that allows only active fronts to use methods that need writing access. The exception is collision detection, which is needed to avoid the growth of physically overlapping structures. For collision detection, a memory-locking mechanism was necessary to control access to grid points that register the location of nearby fronts. A custom approach using a serialized lock broker was able to manage both read and write locking. NeuroDevSim allows easy modeling of most aspects of neural development for models simulating a few complex or thousands of simple neurons or a mixture of both.

Code available at

https://github.com/CNS-OIST/NeuroDevSim.

Gell: A GPU-powered 3D hybrid simulator for large-scale multicellular system

As a powerful but computationally intensive method, hybrid computational models study the dynamics of multicellular systems by evolving discrete cells in reacting and diffusing extracellular microenvironments. As the scale and complexity of studied biological systems continuously increase, the exploding computational cost starts to limit large-scale cell-based simulations. To facilitate the large-scale hybrid computational simulation and make it feasible on easily accessible computational devices, we develop Gell (GPU Cell), a fast and memory-efficient open-source GPU-based hybrid computational modeling platform for large-scale system modeling. We fully parallelize the simulations on GPU for high computational efficiency and propose a novel voxel sorting method to further accelerate the modeling of massive cell-cell mechanical interaction with negligible additional memory footprint. As a result, Gell efficiently handles simulations involving tens of millions of cells on a personal computer. We compare the performance of Gell with a state-of-the-art paralleled CPU-based simulator on a hanging droplet spheroid growth task and further demonstrate Gell with a ductal carcinoma in situ (DCIS) simulation. Gell affords ~150X acceleration over the paralleled CPU method with one-tenth of the memory requirement.

Retinal self-organization: a model of retinal ganglion cells and starburst amacrine cells mosaic formation

Individual retinal cell types exhibit semi-regular spatial patterns called retinal mosaics. Retinal ganglion cells (RGCs) and starburst amacrine cells (SACs) are known to exhibit such layouts. Mechanisms responsible for the formation of mosaics are not well understood but follow three main principles: (i) homotypic cells prevent nearby cells from adopting the same type, (ii) cell tangential migration and (iii) cell death. Alongside experiments in mouse, we use BioDynaMo, an agent-based simulation framework, to build a detailed and mechanistic model of mosaic formation. We investigate the implications of the three theories for RGC's mosaic formation. We report that the cell migration mechanism yields the most regular mosaics. In addition, we propose that low-density RGC type mosaics exhibit on average low regularities, and thus we question the relevance of regular spacing as a criterion for a group of RGCs to form a RGC type. We investigate SAC mosaics formation and interactions between the ganglion cell layer (GCL) and inner nuclear layer (INL) populations. We propose that homotypic interactions between the GCL and INL populations during mosaics creation are required to reproduce the observed SAC mosaics' characteristics. This suggests that the GCL and INL populations of SACs might not be independent during retinal development.

Hybrid computational models of multicellular tumour growth considering glucose metabolism

Cancer cells metabolize glucose through metabolic pathways that differ from those used by healthy and differentiated cells. In particular, tumours have been shown to consume more glucose than their healthy counterparts and to use anaerobic metabolic pathways, even under aerobic conditions. Nevertheless, scientists have still not been able to explain why cancer cells evolved to present an altered metabolism and what evolutionary advantage this might provide them. Experimental and computational models have been increasingly used in recent years to understand some of these biological questions. Multicellular tumour spheroids are effective experimental models as they replicate the initial stages of avascular solid tumour growth. Furthermore, these experiments generate data which can be used to calibrate and validate computational studies that aim to simulate tumour growth. Hybrid models are of particular relevance in this field of research because they model cells as individual agents while also incorporating continuum representations of the substances present in the surrounding microenvironment that may participate in intracellular metabolic networks as concentration or density distributions. Henceforth, in this review, we explore the potential of computational modelling to reveal the role of metabolic reprogramming in tumour growth.

An Agent-Based Model of Radiation-Induced Lung Fibrosis

Early- and late-phase radiation-induced lung injuries, namely pneumonitis and lung fibrosis (RILF), severely constrain the maximum dose and irradiated volume in thoracic radiotherapy. As the most radiosensitive targets, epithelial cells respond to radiation either by undergoing apoptosis or switching to a senescent phenotype that triggers the immune system and damages surrounding healthy cells. Unresolved inflammation stimulates mesenchymal cells' proliferation and extracellular matrix (ECM) secretion, which irreversibly stiffens the alveolar walls and leads to respiratory failure. Although a thorough understanding is lacking, RILF and idiopathic pulmonary fibrosis share multiple pathways and would mutually benefit from further insights into disease progression. Furthermore, current normal tissue complication probability (NTCP) models rely on clinical experience to set tolerance doses for organs at risk and leave aside mechanistic interpretations of the undergoing processes. To these aims, we implemented a 3D agent-based model (ABM) of an alveolar duct that simulates cell dynamics and substance diffusion following radiation injury. Emphasis was placed on cell repopulation, senescent clearance, and intra/inter-alveolar bystander senescence while tracking ECM deposition. Our ABM successfully replicates early and late fibrotic response patterns reported in the literature along with the ECM sigmoidal dose-response curve. Moreover, surrogate measures of RILF severity via a custom indicator show qualitative agreement with published fibrosis indices. Finally, our ABM provides a fully mechanistic alveolar survival curve highlighting the need to include bystander damage in lung NTCP models.

Thermodynamically-motivated chemo-mechanical models and multicellular simulation to provide new insight into active cell and tumour remodelling

Computational models can shape our understanding of cell and tissue remodelling, from cell spreading, to active force generation, adhesion, and growth. In this mini-review, we discuss recent progress in modelling of chemo-mechanical cell behaviour and the evolution of multicellular systems. In particular, we highlight recent advances in (i) free-energy based single cell models that can provide new fundamental insight into cell spreading, cancer cell invasion, stem cell differentiation, and remodelling in disease, and (ii) mechanical agent-based models to simulate large numbers of discrete interacting cells in proliferative tumours. We describe how new biological understanding has emerged from such theoretical models, and the trade-offs and constraints associated with current approaches. Ultimately, we aim to make a case for why theory should be integrated with an experimental workflow to optimise new in-vitro studies, to predict feedback between cells and their microenvironment, and to deepen understanding of active cell behaviour.

Mathematical and computational models in spheroid-based biofabrication

Ubiquitous in embryonic development, tissue fusion is of interest to tissue engineers who use tissue spheroids or organoids as building blocks of three-dimensional (3D) multicellular constructs. This review presents mathematical models and computer simulations of the fusion of tissue spheroids. The motivation of this study stems from the need to predict the post-printing evolution of 3D bioprinted constructs. First, we provide a brief overview of differential adhesion, the main morphogenetic mechanism involved in post-printing structure formation. It will be shown that clusters of cohesive cells behave as an incompressible viscous fluid on the time scale of hours. The discussion turns then to mathematical models based on the continuum hydrodynamics of highly viscous liquids and on statistical mechanics. Next, we analyze the validity and practical use of computational models of multicellular self-assembly in live constructs created by tissue spheroid bioprinting. Finally, we discuss the perspectives of the field as machine learning starts to reshape experimental design, and modular robotic workstations tend to alleviate the burden of repetitive tasks in biofabrication. STATEMENT OF SIGNIFICANCE: Bioprinted constructs are living systems, which evolve via morphogenetic mechanisms known from developmental biology. This review presents mathematical and computational tools devised for modeling post-printing structure formation. They help achieving a desirable outcome without expensive optimization experiments. While previous reviews mainly focused on assumptions, technical details, strengths, and limitations of computational models of multicellular self-assembly, this article discusses their validity and practical use in biofabrication. It also presents an overview of mathematical models that proved to be useful in the evaluation of experimental data on tissue spheroid fusion, and in the calibration of computational models. Finally, the perspectives of the field are discussed in the advent of robotic biofabrication platforms and bioprinting process optimization by machine learning.

No model is perfect, but some are useful

Agent-based model should inform the action plan to curb algal blooms in Lake Erie.

Interrogating and Quantifying In Vitro Cancer Drug Pharmacodynamics via Agent-Based and Bayesian Monte Carlo Modelling

The effectiveness of chemotherapy in cancer cell regression is often limited by drug resistance, toxicity, and neoplasia heterogeneity. However, due to the significant complexities entailed by the many cancer growth processes, predicting the impact of interference and symmetry-breaking mechanisms is a difficult problem. To quantify and understand more about cancer drug pharmacodynamics, we combine in vitro with in silico cancer models. The anti-proliferative action of selected cytostatics is interrogated on human colorectal and breast adenocarcinoma cells, while an agent-based computational model is employed to reproduce experiments and shed light on the main therapeutic mechanisms of each chemotherapeutic agent. Multiple drug administration scenarios on each cancer cell line are simulated by varying the drug concentration, while a Bayesian-based method for model parameter optimisation is employed. Our proposed procedure of combining in vitro cancer drug screening with an in silico agent-based model successfully reproduces the impact of chemotherapeutic drugs in cancer growth behaviour, while the mechanisms of action of each drug are characterised through model-derived probabilities of cell apoptosis and division. We suggest that our approach could form the basis for the prospective generation of experimentally-derived and model-optimised pharmacological variables towards personalised cancer therapy.

A 3D Agent-Based Model of Lung Fibrosis

Understanding the pathophysiology of lung fibrosis is of paramount importance to elaborate targeted and effective therapies. As it onsets, the randomly accumulating extracellular matrix (ECM) breaks the symmetry of the branching lung structure. Interestingly, similar pathways have been reported for both idiopathic pulmonary fibrosis and radiation-induced lung fibrosis (RILF). Individuals suffering from the disease, the worldwide incidence of which is growing, have poor prognosis and a short mean survival time. In this context, mathematical and computational models have the potential to shed light on key underlying pathological mechanisms, shorten the time needed for clinical trials, parallelize hypotheses testing, and improve personalized drug development. Agent-based modeling (ABM) has proven to be a reliable and versatile simulation tool, whose features make it a good candidate for recapitulating emergent behaviors in heterogeneous systems, such as those found at multiple scales in the human body. In this paper, we detail the implementation of a 3D agent-based model of lung fibrosis using a novel simulation platform, namely, BioDynaMo, and prove that it can qualitatively and quantitatively reproduce published results. Furthermore, we provide additional insights on late-fibrosis patterns through ECM density distribution histograms. The model recapitulates key intercellular mechanisms, while cell numbers and types are embodied by alveolar segments that act as agents and are spatially arranged by a custom algorithm. Finally, our model may hold potential for future applications in the context of lung disorders, ranging from RILF (by implementing radiation-induced cell damage mechanisms) to COVID-19 and inflammatory diseases (such as asthma or chronic obstructive pulmonary disease).