The field of metabolic modelling at the genome scale continues to expand, with increasing numbers of models being reconstructed, shared and curated. This expansion creates a need to adopt common principles regarding transparency and versioning, in addition to standardisation of file formats, annotations, and model testing. Here, we present a repository standard and template for Git-based genome-scale metabolic models (GEMs) supporting both newly reconstructed and actively curated models, following FAIR principles (findability, accessibility, interoperability, and reusability), and incorporating research software best practices. Standard-GEM exposes GEMs across web services and platforms in the metabolic modelling field and enables automatic validation of GEMs. The use of this template for new models and its adoption reduces the effort required to share, validate and reuse GEMs.

Schlüsselwörter: genome-scale metabolic model, open source, versioning, Git

As biomedical knowledge keeps growing, resources storing available information multiply and grow in size and complexity. Such resources can be in the format of molecular interaction maps, which represent cellular and molecular processes under normal or pathological conditions. However, these maps can be complex and hard to navigate, especially to novice users. Large Language Models (LLMs), particularly in the form of agentic frameworks, have emerged as a promising technology to support this exploration. In this article, we describe a user-driven process of prototyping, development, and user testing of Llemy, an LLM-based system for exploring these molecular interaction maps. By involving domain experts from the very first prototyping in the form of a hackathon and collecting both fine-grained and general feedback on more refined versions, we were able to evaluate the perceived utility and quality of the developed system, in particular for summarising maps and pathways, as well as prioritise the development of future features. We recommend continued user-driven development and benchmarking to keep the community engaged. This will also facilitate the transition towards open-weight LLMs to support the needs of the open research environment in an ever-changing technology landscape.

Schlüsselwörter: SBGN, LLM

Computational models are essential for studying complex systems which, particularly in clinical settings, need to be quality-approved and transparent. To enhance the communication of a model’s features and capabilities, we propose an adaptation of the Findability, Accessibility, Interoperability and Reusability (FAIR) indicators published by the Research Data Alliance to assess models encoded in domain-specific standards, such as those established by COMBINE. The assessments guide FAIRification and add value to models.

Schlüsselwörter: FAIR, computational models

Constraint-based modeling is a well-established modeling methodology used to study biological networks at both the medium-scale and genome-scale. Due to their large size and complexity, such steady-state flux models are typically ana- lyzed using constraint-based optimization techniques, such as Flux Balance Anal- ysis (FBA). The Flux Balance Constraints (FBC) Package extends SBML Level 3 to provide a standardized format for encoding, exchanging, and annotating constraint-based models. It includes support for modeling concepts such as objective functions, flux bounds, and annotation of model components that facilitate reaction balanc- ing. Version two extended the original release by adding support for encoding gene-protein associations. Version three builds upon and maintains backwards compatibility with Version two by introducing new elements and attributes that include: user-defined constraints, a new quadratic variable type and key-value pair annotation that can be used to store additional information relevant to constraint-based modeling. The only changes to existing attributes are that a Species charge is no longer restricted to an integer value and chemicalFormula can now include generic components. In addition to providing the elements necessary to uniquely encode constraint- based models, the FBC package provides an open platform that facilitates the continued, cross-community development of an interoperable, constraint-based model encoding format.

Schlüsselwörter: SBML, FBC

Parameter estimation is a central task in data-driven computational modeling, and a clear, unambiguous, and commonly understood specification of parameter estimation problems is essential for reproducibility and interoperability. PEtab provides a standardized framework for defining such problems in systems biology and systems medicine, particularly for ordinary differential equation models. This document describes the second version of the PEtab format, PEtab 2.0. PEtab 2.0 simplifies and clarifies how estimation problems are defined, while also extending the standard to support more complex and realistic experimental setups.

Schlüsselwörter: Parameter estimation, PETab

Stochastic simulations underpin computational research in fields from systems biology and epidemiology to finance and physics-informed machine learning, yet their reproducibility is hard to quantify because each run yields different outcomes. Existing practices to improve computational reproducibility focus on sharing code, simulation seeds, data, and software environments but do not ensure independent reproduction of results, limiting scientific rigor. Here we introduce the Empirical Characteristic Function Equality Convergence Test (EFECT), a universal computational framework for quantifying the statistical reproducibility of stochastic simulations based on empirical characteristic functions. EFECT defines a normalized EFECT error that measures distributional differences between two sets of simulation outputs over a model- and scale-independent range, and an EFECT convergence point that specifies the minimum sample size required to achieve a desired reproducibility threshold at a chosen significance level. EFECT applies to any simulation whose results can be represented as bounded, real-valued data, regardless of modelling formalism or source of stochasticity. To facilitate widespread adoption and data exchange, we implemented EFECT in an open-source software library and a compact, machine-readable standardized report. We rigorously evaluated the framework across more than 40 test cases spanning stochastic differential equations, agent- based models, Boolean networks, and partial differential equations. Applying EFECT to published models in pandemic epidemiology, financial stochastic volatility, and physics-informed neural networks reveals that commonly used sample sizes are often underpowered: substantial parameter differences can go undetected unless thousands of replicates are simulated. By providing a domain-agnostic statistical reproducibility metric, reusable software libraries, and a standardized report format, EFECT offers a practical framework for embedding quantitative reproducibility and replicability checks into stochastic simulation workflows across data-intensive disciplines. Thus, EFECT delivers the capability to make scientific inferences and policy decisions on the basis of reliable, reproducible stochastic simulations.

Schlüsselwörter: COMBINE, reproducibility

Background: Hydrochlorothiazide is a widely prescribed thiazide diuretic used for the treatment of hypertension. Cardiac, hepatic, and renal impairments are common comorbidities in hypertensive patients and may contribute to substantial interindividual variability in pharmacokinetic (PK) and pharmacodynamic (PD) responses. This variability complicates dose optimization, particularly in clinically heterogeneous patient populations. Methods: We developed a whole-body physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) model of hydrochlorothiazide integrating oral absorption, systemic distribution, renal elimination, renal sodium, chloride, and fluid handling, and blood-pressure regulation. The model was calibrated and evaluated using systematically curated PK and PD data from 25 clinical studies in healthy individuals and patients with hypertension, cardiac impairment, hepatic impairment, or renal impairment. Results: Model predictions were consistent with observed dose-proportional PK and PD and captured variability in hydrochlorothiazide exposure and response across simulated comorbid conditions. Renal impairment was identified as the main driver of altered hydrochlorothiazide exposure and reduced urinary excretion. By contrast, hepatic and cardiac impairment affected PK primarily through secondary effects on renal function, rather than through direct effects on drug metabolism or absorption. Conclusions: The developed PBPK/PD model provides a mechanistic framework for understanding hydrochlorothiazide pharmacology across heterogeneous clinical conditions. It supports model-informed assessment of variability in exposure and response and may contribute to individualized antihypertensive therapy in patients with comorbidities.

Schlüsselwörter: digital twin, hydrochlorothiazide, hypertension, physiologically based pharmacokinetic/pharmacodynamic model (PBPK/PD), pharmacokinetics, pharmacodynamics, personalized medicine

Logical models are widely used to study regulatory and signaling systems, yet their reuse, annotation, and exchange across tools remain challenging. Although SBML Level 3 Qualitative Models (SBML-qual) provides a standard representation, its XML-based syntax is difficult to inspect and edit directly. Here we introduce TabularQual, a spreadsheet-based, community-driven representation for Boolean and multi-valued logical models, together with a bidirectional converter between spreadsheets and SBML-qual. The converter is accessible programmatically and via a web interface to support diverse user workflows. We further describe two integration strategies that enable existing modeling tools to operate with TabularQual, either through SBML-qual exchange or via direct support. Case studies using the Stress Knowledge Map and CaSQ demonstrate how this integration supports model construction, curation, and reuse. TabularQual provides a practical bridge between human-readable model representations and standardized executable formats, supporting reproducibility, interoperability, and community-driven model development.

Schlüsselwörter: Logical models, qual, SBML, standardization, COMBINE

A computational model in the form of a whole-body physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) model of dapagliflozin was developed to systematically evaluate the influence of patient-specific factors on drug disposition. Based on curated data from 28 clinical studies, the model simulates the absorption, distribution, metabolism and excretion (ADME) of the drug as well as its pharmacodynamics. The model accounts for variability in renal and hepatic function, and effects of food intake. The model is implemented in the Systems Biology Markup Language (SBML) standard. Analysis were performed utilizing the libroadrunner simulation library. Here, we demonstrate the computational reproducibility of the key findings from the primary publication, thereby verifying the consistency of the model implementation with the published results.

Schlüsselwörter: Dapagliflozin, PBPK/PD, SBML, Pharmacokinetics, Pharmacodynamics, Computational Model

Sirolimus is a potent mTOR inhibitor used primarily to prevent acute rejection in transplant recipients. Its clinical management is challenging because of its narrow therapeutic index, low and highly variable oral bioavailability, and pronounced inter-individual variability driven by CYP3A4/5-mediated metabolism and P-glycoprotein efflux. Extensive partitioning into erythrocytes further complicates its disposition and necessitates therapeutic drug monitoring. Here, we developed a mechanistic whole-body physiologically based pharmacokinetic (PBPK) digital twin of rapamycin that integrates complex absorption kinetics, nonlinear distribution, and first-pass metabolism. The SBML-encoded model was calibrated and evaluated against a comprehensive library of curated clinical pharmacokinetic data, comprising studies primarily in healthy volunteers and stable renal transplant recipients. The dataset covers diverse ethnic populations, cohorts with varying degrees of renal and hepatic impairment, and individuals with relevant genetic polymorphisms. The digital twin captured overall trends in rapamycin blood concentrations across a wide range of doses and dosing regimens. Simulations showed good agreement with observed data under hepatic and renal impairment, as well as under fasted and fed conditions. Furthermore, the model reproduced the magnitude of drug--drug interactions involving potent CYP3A4 inhibitors, CYP3A4 inducers, and concomitant immunosuppressive agents. This SBML-based digital twin provides a quantitative framework for characterizing sirolimus dose dependency and the multifactorial effects of intrinsic and extrinsic factors on systemic exposure. By providing the model in a standards-based, executable format together with simulation scripts and curated pharmacokinetic datasets, this work supports independent reproduction, transparent model evaluation, and systematic reuse in accordance with FAIR principles.

Schlüsselwörter: rapamycin, sirolimus, PBPK modeling, digital twin, mTOR inhibitor, drug–drug interactions, organ impairment, personalized medicine

Background: Thrombotic events increase with age, necessitating anticoagulants with reliable pharmacokinetic (PK) and pharmacodynamic (PD) profiles. Apixaban has important therapeutic advantages, but individualised dosing remains challenging because exposure and response are influenced by renal and hepatic function, food intake, body weight, and other patient-specific factors. Existing physiologically based pharmacokinetics/pharmacodynamics (PBPK/PD) models are limited by data sources, transparency, and incomplete representation of metabolites and pharmacodynamics. Methods: A systematic literature review identified 35 apixaban PK/PD clinical studies, which were curated and used for model development, parameter optimisation, and evaluation. We developed an expanded whole-body PBPK/PD model of apixaban with explicit metabolite representation and enhanced pharmacodynamic components. The model follows a modular structure and is encoded in SBML to support interoperability and reproducibility. Results: The model reproduced observed clinical PK/PD data across all 35 studies, covering diverse doses, regimens, and populations. Simulations captured apixaban PK and PD under normal conditions and clinically relevant scenarios, including renal and hepatic impairment, fasted and fed states, and obesity. Conclusions: This open PBPK/PD digital twin provides quantitative insight into determinants of apixaban exposure and response. All model files, documentation, simulation scripts, and curated datasets are openly available under MIT and CC-BY licenses following FAIR principles.

Schlüsselwörter: apixaban; anticoagulation, pharmacokinetics/pharmacodynamics, PBPK/PD modelling, digital twin, personalised medicine, special populations, FAIR data, reproducible modelling

Automated analysis of hepatic vascular structures and lobules within whole-slide histological images is critical for ensuring accurate and timely morphometric evaluations and facilitating advancements in computational liver histology. Nonetheless, the intricate morphology of the tissue, variability in staining techniques, and the requirements for high-resolution imaging present substantial challenges to the precision of segmentation processes. We present a robust deep-learning pipeline using adaptive patch extraction and specialized U-Net architectures for segmenting vessels, bile ducts, and lobules in Glutamine Synthetase and Picro-Sirius-Red stained porcine liver sections. Our architecture incorporates a weight-boosted nnU-Net framework to effectively manage class imbalances and improve the detection of smaller vascular structures. Geometric data transformations enhanced the robustness and generalizability of the segmentation models. Evaluations conducted through five-fold cross-validation, as well as assessments utilizing independent test datasets, resulted in Dice similarity scores: 0.960 for lobules, 0.801 for central veins, 0.909 for hepatic arteries, 0.609 for portal veins, and 0.710 for bile ducts. The developed segmentation pipeline additionally supports comprehensive morphometric analyses of structural parameters, including number and size (diameter, area) of vascular structures, bile ducts, and lobules. For e.g., the diameter of hepatic arteries ranges between 30-90 µm. These findings underscore the practical relevance of adaptable segmentation frameworks in advancing computational histological analysis of liver tissue.

Schlüsselwörter: Hepatic Vascular Segmentation, Lobule Detection, Histological Image, Analysis, Deep Learning, Lobule Segmentation

Bayesian methods offer a principled framework for parameter estimation and uncertainty quantification in physiologically-based pharmacokinetic (PBPK) modeling, yet the sensitivity of these approaches to key workflow hyperparameters remains insufficiently characterized. This thesis presents a reproducible methodology for evaluating Bayesian Markov Chain Monte Carlo (MCMC) parameter estimation across three pharmacokinetic models of increasing complexity: a simple chain model, a basic pharmacokinetic model, and a clinically validated PBPK model of indocyanine green (ICG). Synthetic pharmacokinetic datasets were generated from known parameter distributions and used to fit Bayesian MCMC samplers under systematically varied hyperparameters — number of timepoints, number of samples, coefficient of variation, and prior type. Posterior accuracy, point bias, computational cost, and sampling efficiency were evaluated across all conditions. Among the four hyperparameters investigated, prior specification emerged as the dominant determinant of posterior accuracy: well-specified priors yielded parameter estimates close to the true values across all models, while biased and incorrect priors introduced substantial and progressive deviations. The remaining hyperparameters had limited influence on estimation accuracy but the number of timepoints and samples directly increased computational cost. The entire workflow — built on SBML, libRoadRunner, PEtab, and pyPESTO — is fully open-source and FAIR-compliant, providing a transparent and reusable framework for validating Bayesian approaches in pharmacokinetic modeling.

Schlüsselwörter: Baysian, parameter optimization, SBML, PEtab

Guidelines for managing scientific data have been established under the FAIR principles requiring that data be Findable, Accessible, Interoperable, and Reusable. In many scientific disciplines, especially computational biology, both data and models are key to progress. For this reason, and recognizing that such models are a very special type of 'data', we argue that computational models, especially mechanistic models prevalent in medicine, physiology and systems biology, deserve a complementary set of guidelines. We propose the CURE principles, emphasizing that models should be Credible, Understandable, Reproducible, and Extensible. We delve into each principle, discussing verification, validation, and uncertainty quantification for model credibility; the clarity of model descriptions and annotations for understandability; adherence to standards and open science practices for reproducibility; and the use of open standards and modular code for extensibility and reuse. We outline recommended and baseline requirements for each aspect of CURE, aiming to enhance the impact and trustworthiness of computational models, particularly in biomedical applications where credibility is paramount. Our perspective underscores the need for a more disciplined approach to modeling, aligning with emerging trends such as Digital Twins and emphasizing the importance of data and modeling standards for interoperability and reuse. Finally, we emphasize that given the non-trivial effort required to implement the guidelines, the community moves to automate as many of the guidelines as possible.

Schlüsselwörter: FAIR, CURE, reproducibility

Background/Objectives: Empagliflozin is an SGLT2 inhibitor prescribed for the management of type 2 diabetes mellitus, lowering blood glucose by increasing urinary glucose excretion (UGE) through inhibition of renal glucose reabsorption. PK/PD responses vary substantially across patient populations, complicating dose selection under altered organ function. Here, we developed a whole-body PBPK/PD digital twin integrating absorption, distribution, metabolism, and excretion with explicit modeling of renal glucose handling via the renal threshold for glucose. Methods: The model represents empagliflozin and its glucuronide metabolite, is implemented in SBML, and was calibrated and evaluated against curated PK/PD data from 27 clinical studies spanning healthy individuals, patients with type 2 diabetes, and cohorts with renal or hepatic impairment. Results: The model accurately captured observed clinical PK/PD data across all 27 studies, spanning a wide range of doses, dosing regimens, and patient populations. Good agreement between simulations and observations was obtained under normal and impaired renal and hepatic function, as well as under fasted and fed conditions, demonstrating the model's ability to reproduce empagliflozin disposition and pharmacodynamic response across clinically relevant scenarios. Conclusions: This SBML-based PBPK/PD digital twin provides quantitative insight into empagliflozin dose dependency and the impact of renal impairment, hepatic impairment, and food intake on PK/PD across clinically relevant populations. All model files, simulation scripts, and curated datasets are openly available in accordance with FAIR principles.

Schlüsselwörter: digital twin, empagliflozin, Type 2 diabetes mellitus, physiologically based pharmacokinetic/pharmacodynamic model (PBPK/PD), pharmacokinetics, pharmacodynamics, personalized medicine

Background/Objectives: Dapagliflozin is an SGLT2 inhibitor prescribed for the management of type 2 diabetes mellitus. The drug lowers blood glucose levels by increasing urinary glucose excretion (UGE). Despite established efficacy, dapagliflozin demonstrates significant inter-individual variability in pharmacokinetics (PK) and pharmacodynamics (PD), with potential impact on treatment outcomes. Methods: To evaluate the sources of variability and to support patient stratification and model-informed individualized therapy, we developed a physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) model of dapagliflozin using curated data from 28 clinical studies. This framework integrates absorption, distribution, metabolism, excretion, and pharmacodynamics, and accounts for key determinants of variability including renal and hepatic function, and food effects. Results: The simulations reproduced dose-dependent pharmacokinetics with predicted Cmax and AUC values typically within 10–15% of observed data. Renal impairment reduced UGE by 40–60% despite modest changes in plasma exposure, while hepatic impairment produced only small shifts in PK and PD. The model also reproduced the fed-state reduction of peak concentrations, consistent with the 30–50% decrease reported clinically. Conclusions: All model files, code, and curated datasets are openly available in line with FAIR standards and Open Science practices, enabling transparent and reproducible analyses and providing a mechanistic basis for individualized therapy in type 2 diabetes.

Schlüsselwörter: dapagliflozin, type 2 diabetes mellitus, physiologically based pharmacokinetic/pharmacodynamic model (PBPK/PD), pharmacokinetics, pharmacodynamics, personalized medicine

Background/Objectives: Losartan, an angiotensin II receptor blocker (ARB) used to treat hypertension and heart failure, shows significant variability in pharmacokinetics (PK) and pharmacodynamics (PD) among individuals. Methods: In this study, we developed a physiologically based pharmacokinetic/ pharmacodynamic (PBPK/PD) model of losartan and its active metabolite, E3174, using curated data from 25 clinical trials. The model mechanistically describes the processes of absorption, hepatic metabolism, renal and fecal excretion, and pharmacodynamic blood pressure regulation. Simulation studies examined the effects of dose, hepatic and renal impairment, and genetic polymorphisms in cytochrome p450 2C9 (CYP2C9) and P-glycoprotein 1, also known as multidrug resistance protein 1 (MDR1) or ATP-binding cassette sub-family B member 1 (ABCB1), on the model. Results: The model successfully reproduced key PK/PD observations, including dose-dependent receptor blockade, attenuated responses with hepatic impairment, modest enhancement with renal impairment, and substantial variability in E3174 formation dependent on CYP2C9; the effects of ABCB1 were minimal. Specifically, dose dependency simulations confirmed the saturable nature of CYP2C9 metabolism, predicting a decreasing E3174-to-losartan ratio and a stronger, sustained suppression of blood pressure and aldosterone at higher doses. Hepatic impairment was predicted to lead to elevated losartan plasma concentrations (increased AUC) and attenuated metabolite formation, confirming the clinical need for dose reduction. Renal impairment simulations predicted stable losartan AUC but showed an overestimation of E3174 accumulation compared to observed data, where E3174 exposure remained stable. Genetic variability (CYP2C9) was the major determinant of response, with simulations confirming that reduced-function alleles lead to a 1.6- to 3-fold increase in losartan AUC and diminished blood pressure reduction. ABCB1 variability resulted in only minor modulation of systemic exposure and blood pressure effects. Conclusions: This mechanistic digital twin framework provides a quantitative basis for understanding variability in losartan therapy and supports its application in individualized dosing strategies.

Schlüsselwörter: losartan, angiotensin II receptor blocker, ARB, PBPK/PD modeling, pharmacokinetics, pharmacodynamics, SBML

A digital twin in the form of a whole-body physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) model of losartan was developed to systematically evaluate the influence of patient-specific factors on drug disposition and effects. Based on curated data from 25 clinical studies, the model simulates the absorption, distribution, metabolism and excretion (ADME) as well as pharmacological effects of the drug. The model accounts for variability caused by the differences in renal and hepatic function, and by genetic polymorphisms of CYP2C9 and ABCB1. The model is implemented in the Systems Biology Markup Language (SBML) standard. Simulations were performed utilising the libroadrunner library. Here, we demonstrate the computational reproducibility of the key findings from the primary publication, thereby verifying the consistency and reproducibility of the model implementation with the published results.

Schlüsselwörter: Losartan, PBPK/PD, SBML, Pharmacokinetics, Pharmacodynamics, Computational Model

Canagliflozin is an SGLT2 inhibitor prescribed for the management of type 2 diabetes mellitus, lowering blood glucose by increasing urinary glucose excretion (UGE). However, pharmacokinetic (PK) and pharmacodynamic (PD) responses vary across patient populations, complicating dose selection under altered organ function. Here, we developed a whole-body PBPK/PD digital twin of canagliflozin that integrates absorption, distribution, metabolism, and excretion, and explicitly models renal glucose handling. The model represents canagliflozin and its major metabolites (M5, M7, M9) and was calibrated and evaluated using curated PK/PD data from 22 clinical studies spanning healthy individuals, patients with type 2 diabetes, and cohorts with renal or hepatic impairment. Simulations reproduced dose-dependent exposure and UGE across single- and multiple-dose regimens. Renal impairment produced modest changes in parent drug exposure but markedly reduced UGE and increased metabolite exposure, consistent with reduced glucose filtration and impaired metabolite clearance. Under hepatic impairment, simulations predicted increased canagliflozin exposure with altered metabolite profiles, while PD effects were minimal, although evaluation was limited by sparse clinical PD endpoints. All model files, simulation scripts, and curated datasets are provided in open SBML workflows in accordance with FAIR principles, enabling reproducible simulations and reuse for model-informed analyses of canagliflozin PK/PD variability.

Schlüsselwörter: Digital twin, Canagliflozin, Type 2 diabetes mellitus, Physiologically based pharmacokinetic/pharmacodynamic model (PBPK/PD), Pharmacokinetics, Pharmacodynamics, Personalized medicine

Introduction: Optimizing glimepiride therapy for type 2 diabetes (T2DM) is challenged by pronounced inter-individual variability in pharmacokinetics. Methods: We developed a whole-body physiologically based pharmacokinetic (PBPK) model as a digital twin of glimepiride, enabling systematic evaluation of how patient-specific factors influence drug disposition. Using curated data from 20 clinical studies, the digital twin mechanistically simulates glimepiride’s absorption, distribution, metabolism, and excretion (ADME). It accounts for key determinants of variability including renal and hepatic function, CYP2C9 genotype, and bodyweight. Results: The model accurately reproduced observed pharmacokinetics and quantified these factors’ impact on drug exposure. Increased glimepiride exposure was predicted in individuals with hepatic dysfunction or specific CYP2C9 variants, highlighting substantial genetic and physiological effects. Discussion: This digital twin provides mechanistic insights into pharmacokinetic variability and serves as an in silico platform for exploring individualized dosing and patient stratification strategies, laying the foundation for clinical decision support tools to improve T2DM management.

Schlüsselwörter: digital twin, diabetes, Glimepiride, PBPK, physiologically based pharmacokinetic model

A digital twin in the form of a whole-body physiologically based pharmacokinetic (PBPK) model of glimepiride was developed to systematically evaluate the influence of patient-specific factors on drug disposition. Based on curated data from 20 clinical studies, the model simulates the absorption, distribution, metabolism and excretion (ADME) of the drug while accounting for variability in renal and hepatic function, CYP2C9 genetic variants and bodyweight. The model is implemented in the Systems Biology Markup Language (SBML) standard and simulations are per- formed using scripts that utilise the libRoadRunner library to run simulations and generate results. Here, we demonstrate the computational reproducibility of the key findings from the primary publication, thereby verifying the consistency and reproducibility of the model implementation with the published results.

Schlüsselwörter: digital twin, diabetes, Glimepiride, PBPK, physiologically based pharmacokinetic model

Background: Artificial Intelligence (AI) and Digital Twins (DT) are transforming clinical decision-making, particularly in hepatology and hepatobiliary surgery. While these technologies promise more precise diagnostics and individualised treatment planning, their successful integration depends on clinician acceptance, trust, and regulatory clarity. Methods: We conducted an online survey among hepatologists, hepatobiliary surgeons, and related specialists to assess familiarity, perceived value, adoption of clinical decision support systems (CDSS), concerns regarding automation bias and liability, and attitudes toward patient-centred integration. Data from 18 respondents were analysed using descriptive statistics and visualised in Python. Results: Most clinicians rated AI (80%) and DT (79%) as valuable to very valuable. Adoption of CDSS was high, with 93% reporting current or intended use. However, significant concerns emerged: 71% feared automation bias, and 92% expressed uncertainty about legal protection. Additionally, 77% supported incorporating patient preferences into AI- and DT-assisted decisions. Conclusions: Clinicians demonstrate a strong interest in adopting AI and DT, but remain cautious due to concerns over trust, liability, and regulatory gaps. Clear guidelines, clinical validation, and patient-centred approaches are essential for safe and effective integration into hepatology practice.

Schlüsselwörter: Digital Twins, Artificial Intelligence, Liver, Hepatology, Hepatobiliary Surgery, Survey

Rivaroxaban, a direct oral anticoagulant (DOAC), is a widely used drug for the prevention and treatment of thromboembolic disorders. While its fixed dosing regimen and predictable anticoagulation profile improve clinical usability, significant inter-individual variability in pharmacokinetics (PK) and pharmacodynamics (PD) exists, particularly in patients with altered physiology or comorbidities. This study established an open database of rivaroxaban PK and PD data from 14 clinical studies. The data were used to develop and validate a physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) model of rivaroxaban that captures its absorption, distribution, metabolism, excretion, and coagulation effects. The model incorporates the effects of dose, food intake, hepatic and renal function, and allows simulation of activated Factor X (Factor Xa) inhibition, prothrombin time (PT), and activated partial thromboplastin time (aPTT). Simulations accurately reproduced observed plasma concentration-time and coagulation profiles across a range of doses, fed/fasting conditions, as well as with different degrees of cirrhosis and renal impairment. This mechanistic PBPK/PD model provides a framework for exploring variability in rivaroxaban response across diverse clinical scenarios and supports its use in personalized dosing strategies, clinical trial design, and