Search

Chemical, manufacturing, and control development timelines occupy a significant part of vaccine end-to-end development. In the on-going race for accelerating timelines, in silico process development constitutes a viable strategy that can be achieved through an artificial intelligence (AI)-driven or a mechanistically oriented approach. In this opinion, we focus on the mechanistic option and report on the modeling competencies required to achieve it. By inspecting the most frequent vaccine process units, we identify fluid mechanics, thermodynamics and transport phenomena, intracellular modeling, hybrid modeling and data science, and model-based design of experiments as the pillars for vaccine development. In addition, we craft a generic pathway for accommodating the modeling competencies into an in silico process development strategy.

To date, a large number of experiments are performed to develop a biochemical process. The generated data is used only once, to take decisions for development. Could we exploit data of already developed processes to make predictions for a novel process, we could significantly reduce the number of experiments needed. Processes for different products exhibit differences in behaviour, typically only a subset behave similar. Therefore, effective learning on multiple product spanning process data requires a sensible representation of the product identity. We propose to represent the product identity (a categorical feature) by embedding vectors that serve as input to a Gaussian Process regression model. We demonstrate how the embedding vectors can be learned from process data and show that they capture an interpretable notion of product similarity. The improvement in performance is compared to traditional one-hot encoding on a simulated cross product learning task. All in all, the proposed method could render possible significant reductions in wet-lab experiments., 14 pages, 5 figures

The acetone-butanol-ethanol (ABE) fermentation from corn stover was considered. We propose, study and optimise, via process simulation (using MATLAB® and Aspen Plus®) the use of fermentation tanks-in-series and in situ product recovery by vacuum evaporation. As the operating time of continuous fermentation processes is usually limited at less than 500 h, shutdown and start-up of the reactors are considered and optimised. A multi-objective optimisation methodology that considers economics as well as the energy requirements of the process was used. The optimal configuration was found to be five fermentation tanks-in-series where the first and the last reactors are operated at atmospheric pressure, and the intermediate vessels are operated under vacuum. The economic potential using this configuration was found to be 45% higher than that of vacuum fermenters operating in parallel (continuous operating mode). Also, the total fuel requirements for ABE recovery and purification system were as low as 7 MJ kg−1 butanol, a reduction of between 4 and 33% when compared to a parallel configuration of batch, fed-batch or continuous fermenters. The energy efficiency of this recovery and reaction system was as high as 74% when co-generation is considered.

Digital twins (DTs) represent a vividly emerging technology in the manufacturing industry strongly motivated by the goals of industry 4.0. It strives for smart factories with completely interconnected devices and data archives as well as their automated and adaptive operation based on autonomous decisions of smart algorithms. In this article, we want to provide our vision on the broad potential, critical components as well as current technological and organizational challenges of DTs in biopharma. We put a profound focus on the capability of hybrid modeling as a core technology to realize the capabilities of DTs in the biopharma environment, which is governed by considerable levels of heterogeneity, complexity and uncertainty.

In bioprocesses, specific process responses such as the biomass cannot typically be measured directly on‐line, since analytical sampling is associated with unavoidable time delays. Accessing those responses in real‐time is essential for Quality by Design and process analytical technology concepts. Soft sensors overcome these limitations by indirectly measuring the variables of interest using a previously derived model and actual process data in real time. In this study, a biomass soft sensor based on 2D‐fluorescence data and process data, was developed for a comprehensive study with a 20‐L experimental design, for Escherichia coli fed‐batch cultivations. A multivariate adaptive regression splines algorithm was applied to 2D‐fluorescence spectra and process data, to estimate the biomass concentration at any time during the process. Prediction errors of 4.9% (0.99 g/L) for validation and 3.8% (0.69 g/L) for new data (external validation), were obtained. Using principal component and parallel factor analyses on the 2D‐fluorescence data, two potential chemical compounds were identified and directly linked to cell metabolism. The same wavelength pairs were also important predictors for the regression‐model performance. Overall, the proposed soft sensor is a valuable tool for monitoring the process performance on‐line, enabling Quality by Design.

Hybrid semi-parametric modeling, combining mechanistic and machine-learning methods, has proven to be a powerful method for process development. This paper proposes bootstrap aggregation to increase the predictive power of hybrid semi-parametric models when the process data are obtained by statistical design of experiments. A fed-batch Escherichia coli optimization problem is addressed, in which three factors (biomass growth setpoint, temperature, and biomass concentration at induction) were designed statistically to identify optimal cell growth and recombinant protein expression conditions. Synthetic data sets were generated applying three distinct design methods, namely, Box-Behnken, central composite, and Doehlert design. Bootstrap-aggregated hybrid models were developed for the three designs and compared against the respective non-aggregated versions. It is shown that bootstrap aggregation significantly decreases the prediction mean squared error of new batch experiments for all three designs. The number of (best) models to aggregate is a key calibration parameter that needs to be fine-tuned in each problem. The Doehlert design was slightly better than the other designs in the identification of the process optimum. Finally, the availability of several predictions allowed computing error bounds for the different parts of the model, which provides an additional insight into the variation of predictions within the model components.

This work aims to model, simulate and provide insights into the dynamics and control of COVID-19 infection rates. Using an established epidemiological model augmented with a time-varying disease transmission rate allows daily model calibration using COVID-19 case data from countries around the world. This hybrid model provides predictive forecasts of the cumulative number of infected cases. It also reveals the dynamics associated with disease suppression, demonstrating the time to reduce the effective, time-dependent, reproduction number. Model simulations provide insights into the outcomes of disease suppression measures and the predicted duration of the pandemic. Visualisation of reported data provides up-to-date condition monitoring, while daily model calibration allows for a continued and updated forecast of the current state of the pandemic.

Digital twins (DTs) are expected to render process development and life-cycle management much more cost-effective and time-efficient. A DT definition, a brief retrospect on their history and expectations for their deployment in today's business environment, and a detailed financial assessment of their attractive economic benefits are provided in this chapter. The argument that restrictive guidelines set forth by regulatory agencies would hinder the adoption of DTs in the (bio)pharmaceutical industry is revisited, concluding that those companies who collaborate with the agencies to further their technical capabilities will gain significant competitive advantage. The analyzed process development examples show high methodological readiness levels but low systematic adoption of technology. Given the technical feasibilities, financial opportunities, and regulatory encouragement, concerns regarding intellectual property and data sharing, though required to be taken into account, will at best delay an industry-wide adoption of DTs. In conclusion, it is expected that a strategic investment in DTs now will gain an advantage over competition that will be difficult to overcome by late adopters.

We propose a mathematical model for investment evaluation of process modeling on both monetary and diffusion-of-innovation fronts. Our methodology builds upon Discounted Cash Flow analysis and Bayesian inference and utilizes the Rogers diffusion of innovation paradigm for computing lower expected returns.

In silico process development constitutes a viable option for accelerating CMC development timelines and can be achieved through either a hybrid-modeling-driven or an Artificial Intelligence (AI)-driven avenue. Each pathway has its own pros and cons but the biggest difference is that the former can be developed in-house whereas the latter requires inter-corporation collaboration. Motivated by the precedence of inter-corporation data ecosystems targeting drug discovery, in this paper we bring forward the case of the AI-driven approach to in silico process development in terms of both technical feasibility and scientific soundness. Our analysis asserts that methodologically, AI is now mature to be employed for the task at hand, provided that the overall exercise is driven by multi-disciplinary experts. Further, in silico process development should be understood as a collection of models, with each process unit being endowed with its own model.

In this work, sparse regression using a penalized least absolute deviations objective function is considered. Regression model sparsity is promoted using a L 0 - pseudo norm penalty (the cardinality of the model parameter vector). Implemented using mixed integer linear programming (MILP) it is demonstrated that the use of the L 0 - norm (without approximation) enables efficient and accurate solutions to sparse regression problems of practical size. For model development with a large number of potential model parameters (or features) methods to relax the MILP are also developed; using nonlinear function approximations to the L 0 - norm, penalty terms are linearized and solved using sequential linear programming. Experimental results (using both simulated and real data) demonstrate that these algorithms are also computationally efficient producing accurate and parsimonious model structures. Applications considered are the development of a calibration model for prediction with Near Infrared (NIR) data and the development of a model for the prediction of chemical toxicity - a quantitative structure activity relationship (QSAR).

Enterobacter A47 is a bacterium that produces high amounts of a fucose-rich exopolysaccharide (EPS) from glycerol residue of the biodiesel industry. The fed-batch process is characterized by complex non-linear dynamics with highly viscous pseudo-plastic rheology due to the accumulation of EPS in the culture medium. In this paper, we study hybrid modeling as a methodology to increase the predictive power of models for EPS production optimization. We compare six hybrid structures that explore different levels of knowledge-based and machine-learning model components. Knowledge-based components consist of macroscopic material balances, Monod type kinetics, cardinal temperature and pH (CTP) dependency and power-law viscosity models. Unknown dependencies are set to be identified by a feedforward artificial neural network (ANN). A semiparametric identification schema is applied resorting to a data set of 13 independent fed-batch experiments. A parsimonious hybrid model was identified that describes the dynamics of the 13 experiments with the same parameterization. The final model is specific to Enterobacter A47 but can be easily extended to other microbial EPS processes.

Background: Flux analyses, such as Metabolic Flux Analysis (MFA), Flux Balance Analysis (FBA), Flux Variability Analysis (FVA) or similar methods, can provide insights into the cellular metabolism, especially in combination with experimental data. The most common integration of extracellular concentration data requires the estimation of the specific fluxes (/rates) from the measured concentrations. This is a time-consuming, mathematically ill-conditioned inverse problem, raising high requirements for the quality and quantity of data. Method: In this contribution, a time integrated flux analysis approach is proposed which avoids the error-prone estimation of specific flux values. The approach is adopted for a Metabolic time integrated Flux Analysis and (sparse) time integrated Flux Balance/Variability Analysis. The proposed approach is applied to three case studies: (1) a simulated bioprocess case studying the impact of the number of samples (experimental points) and measurements&rsquo, noise on the performance, (2) a simulation case to understand the impact of network redundancies and reaction irreversibility, and (3) an experimental bioprocess case study, showing its relevance for practical applications. Results: It is observed that this method can successfully estimate the time integrated flux values, even with relatively low numbers of samples and significant noise levels. In addition, the method allows the integration of additional constraints (e.g., bounds on the estimated concentrations) and since it eliminates the need for estimating fluxes from measured concentrations, it significantly reduces the workload while providing about the same level of insight into the metabolism as classic flux analysis methods.

In this study, step variations in temperature, pH, and carbon substrate feeding rate were performed within five high cell density Escherichia coli fermentations to assess whether intraexperiment step changes, can principally be used to exploit the process operation space in a design of experiment manner. A dynamic process modeling approach was adopted to determine parameter interactions. A bioreactor model was integrated with an artificial neural network that describes biomass and product formation rates as function of varied fed-batch fermentation conditions for heterologous protein production. A model reliability measure was introduced to assess in which process region the model can be expected to predict process states accurately. It was found that the model could accurately predict process states of multiple fermentations performed at fixed conditions within the determined validity domain. The results suggest that intraexperimental variations of process conditions could be used to reduce the number of experiments by a factor, which in limit would be equivalent to the number of intraexperimental variations per experiment. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:1343-1352, 2016.

In recent years, multivariate data analysis (MVDA) and modeling approaches have found increasing applications for upstream bioprocess studies (e.g., monitoring, development, optimization, scale-up, etc.). Many of these studies look at variations in the concentrations of metabolites and cell-based measurements. However, these measures are subject to system inherent variations (e.g., changes in metabolic activity) but also intentional operational changes. It is proposed to perform MVDA and modeling on data representative of the underlying biological system operation, that is, the specific rates, which are per se independent of the scale, operational strategy (e.g., batch, fed-batch), and biomass content. Two industrial case studies are highlighted to showcase the approach: one is HEK medium performance comparison study and the other is CHO scale-up/-down study. It is shown that analyzing processes in this way reveals insights into behavior of the underlying biological system, which cannot to the same degree be deducted from the analysis of concentrations.

Upstream bioprocess characterization and optimization are time and resource-intensive tasks. Regularly in the biopharmaceutical industry, statistical design of experiments (DoE) in combination with response surface models (RSMs) are used, neglecting the process trajectories and dynamics. Generating process understanding with time-resolved, dynamic process models allows to understand the impact of temporal deviations, production dynamics, and provides a better understanding of the process variations that stem from the biological subsystem. The authors propose to use DoE studies in combination with hybrid modeling for process characterization. This approach is showcased on Escherichia coli fed-batch cultivations at the 20L scale, evaluating the impact of three critical process parameters. The performance of a hybrid model is compared to a pure data-driven model and the widely adopted RSM of the process endpoints. Further, the performance of the time-resolved models to simultaneously predict biomass and titer is evaluated. The superior behavior of the hybrid model compared to the pure black-box approaches for process characterization is presented. The evaluation considers important criteria, such as the prediction accuracy of the biomass and titer endpoints as well as the time-resolved trajectories. This showcases the high potential of hybrid models for soft-sensing and model predictive control.

The term hybrid modeling is often found to be ambiguous and misleading, as it could refer to hybrid discrete-continuous systems, gray-box modeling, semiparametric modeling, multi-scale modeling, block models, and other methodologies. Different additions to the term have been proposed, such as gray-box, neural (network), or semiparametric, to distinguish this approach clearly from other hybrids, yet these additions have raised discussions and concerns themselves. In essence, the type of a modeling that is referred to in this book as hybrid modeling combines model structures that have been distilled from a priori process knowledge with those whose structure is determined from data. This chapter provides an intuitive introduction to process modeling via fundamental, data-driven, and hybrid modeling methods to set the stage for the following chapters. A brief introduction is given into the idea and history of hybrid modeling. The key benefits of hybrid modeling are highlighted and illustrated using a simple example. Scenarios are outlined under which hybrid modeling can add value to process operation and design in relation to other modeling approaches, but cases in which other approaches should be preferred are also addressed. The challenges of developing hybrid models and potential problems with their application for operation are described in this chapter, and the stage is set for the following book chapters.

Predicting the activity of modified biological parts is difficult due to the typically large size of nucleotide sequences, resulting in combinatorial designs that suffer from the “curse of dimensionality” problem. Mechanistic design methods are often limited by knowledge availability. Empirical methods typically require large data sets, which are difficult and/or costly to obtain. In this study, we explore for the first time the combination of both approaches within a formal hybrid semiparametric framework in an attempt to overcome the limitations of the current approaches. Protein translation as a function of the 5’ untranslated region sequence in Escherichia coli is taken as case study. Thermodynamic modeling, partial least squares (PLS) and hybrid parallel combinations thereof are compared for different data sets and data partitioning scenarios. The results suggest a significant and systematic reduction of both calibration and prediction errors by the hybrid approach in comparison to standalone thermodynamic or PLS modeling. Although with different magnitudes, improvements are observed irrespective of sample size and partitioning method. All in all the results suggest an increase of predictive power by the hybrid method potentially leading to a more efficient design of biological parts.

This multiauthor review article aims to bring readers up to date with some of the current trends in the field of process analytical technology (PAT) by summarizing each aspect of the subject (sensor development, PAT based process monitoring and control methods) and presenting applications both in industrial laboratories and in manufacture e.g. at GSK, AstraZeneca and Roche. Furthermore, the paper discusses the PAT paradigm from the regulatory science perspective. Given the multidisciplinary nature of PAT, such an endeavour would be almost impossible for a single author, so the concept of a multiauthor review was born. Each section of the multiauthor review has been written by a single expert or group of experts with the aim to report on its own research results. This paper also serves as a comprehensive source of information on PAT topics for the novice reader.

A substantial drug release from poly(lactic-co-glycolic) acid (PLGA) micro- and nanoparticles can occur in the first hours of immersion, which is referred to as burst release. A strong burst release (when not intentional) is to be avoided as it decreases the efficacy of the treatment and could be dangerous to the host. In this work we analyze the total amount of drug released during burst and respective kinetics in relation to formulations characteristics, experimental conditions and drug molecular properties in 154 drug release experiments with 41 different drugs by partial least squares (PLS) and decision tree regression. The model created enables to quantify to which degree the physicochemical parameters control the burst release from PLGA particles. Our analysis shows that the amount of drug released during burst is mostly influenced by the formulation characteristics and the synthesis parameters, whereas the drug release kinetics is also influenced by the molecular properties of the drug. The variables that significantly influence the amount and kinetics of the burst release are discussed in detail and compared with findings from other researchers. The final regression models are shown to predict the release profile of a new drug, opening the possibility to be applied to systematically manipulate the burst release by means of designing an optimized drug delivery system.

This report highlights the drivers, challenges, and enablers of the hybrid modeling applications in biopharmaceutical industry. It is a summary of an expert panel discussion of European academics and industrialists with relevant scientific and engineering backgrounds. Hybrid modeling is viewed in its broader sense, namely as the integration of different knowledge sources in form of parametric and nonparametric models into a hybrid semi-parametric model, for instance the integration of fundamental and data-driven models. A brief description of the current state-of-the-art and industrial uptake of the methodology is provided. The report concludes with a number of recommendations to facilitate further developments and a wider industrial application of this modeling approach. These recommendations are limited to further exploiting the benefits of this methodology within process analytical technology (PAT) applications in biopharmaceutical industry.

In this age of technology, the vision of manufacturing industries built of smart factories is not a farfetched future. As a prerequisite for Industry 4.0, industrial sectors are moving towards digitalization and automation. Despite its tremendous growth reaching a sales value of worth $188 billion in 2017, the biopharmaceutical sector distinctly lags in this transition. Currently, the challenges are innovative market disruptions such as personalized medicine as well as increasing commercial pressure for faster and cheaper product manufacturing. Improvements in digitalization and data analytics have been identified as key strategic activities for the next years to face these challenges. Alongside, there is an emphasis by the regulatory authorities on the use of advanced technologies, proclaimed through initiatives such as Quality by Design (QbD) and Process Analytical Technology (PAT). In the manufacturing sector, the biopharmaceutical domain features some of the most complex and least understood processes. Thereby, process models that can transform process data into more valuable information, guide decision-making, and support the creation of digital and automated technologies are key enablers. This review summarizes the current state of model-based methods in different bioprocess related applications and presents the corresponding future vision for the biopharmaceutical industry to achieve the goals of Industry 4.0 while meeting the regulatory requirements.

Process modeling in pharma is gradually gaining momentum in process development but budget restrictions are growing. We first examine whether and how current practices rationalize within a decision process framework with a fictitious investor facing a decision problem subject to incomplete information. We then develop an algorithmic procedure for investment evaluation on both monetary and diffusion-of-innovation fronts. Our methodology builds upon discounted cash flow analysis and Bayesian inference and utilizes the Rogers diffusion of innovation paradigm for computing lower expected returns. We also introduce a set of intangible metrics for quantifying the level of diffusion of process modeling within an organization.

Statistical Design of Experiments (DoE) is a widely adopted methodology in upstream bioprocess development (and generally across industries) to obtain experimental data from which the impact of independent variables (factors) on the process response can be inferred. In this work, a method is proposed that reduces the total number of experiments suggested by a traditional DoE. The method allows the evaluation of several DoE combinations to be compressed into a reduced number of experiments, which is referred to as intensified Design of Experiments (iDoE). In this paper, the iDoE is used to develop a dynamic hybrid model (consisting of differential equations and a feedforward artificial neural network) for data generated from a simulated Escherichia coli fermentation. For the case study presented, the results suggest that the total number of experiments could be reduced by about 40% when compared to traditional DoE. An additional benefit is the simultaneous development of an appropriate dynamic model which can be used in both, process optimization and control studies.

Background Non-negative linear combinations of elementary flux modes (EMs) describe all feasible reaction flux distributions for a given metabolic network under the quasi steady state assumption. However, only a small subset of EMs contribute to the physiological state of a given cell. Results In this paper, a method is proposed that identifies the subset of EMs that best explain the physiological state captured in reaction flux data, referred to as principal EMs (PEMs), given a pre-specified universe of EM candidates. The method avoids the evaluation of all possible combinations of EMs by using a branch and bound approach which is computationally very efficient. The performance of the method is assessed using simulated and experimental data of Pichia pastoris and experimental fluxome data of Saccharomyces cerevisiae. The proposed method is benchmarked against principal component analysis (PCA), commonly used to study the structure of metabolic flux data sets. Conclusions The overall results show that the proposed method is computationally very effective in identifying the subset of PEMs within a large set of EM candidates (cases with ~100 and ~1000 EMs were studied). In contrast to the principal components in PCA, the identified PEMs have a biological meaning enabling identification of the key active pathways in a cell as well as the conditions under which the pathways are activated. This method clearly outperforms PCA in the interpretability of flux data providing additional insights into the underlying regulatory mechanisms. Electronic supplementary material The online version of this article (doi:10.1186/s12859-016-1063-0) contains supplementary material, which is available to authorized users.

A general hybrid semi-parametric process control framework is proposed in this study. The motivation was the integration of different levels of knowledge systems into a general hybrid semi-parametric control structure, of which the general linear controller or the PID controller are, for instance, particular cases. Several hybrid semi-parametric control structure variants and tuning methods are benchmarked in relation to a simulated bioprocess control problem, namely closed-loop control of the biomass concentration through manipulation of the substrate feeding rate, coupled with the closed-loop control of the dissolved oxygen concentration through the stirring velocity. The results demonstrate that (i) due to the hybrid approach the control loop can be closed without any additional identification experiments; (ii) the incorporation of different types of knowledge can enhance the controller performance, when compared to structures without a priori knowledge; (iii) knowledge incorporation seems to facilitate the tuning of the controller; and (iv) the control action can be analyzed in relation to structural information incorporated into the hybrid controller.

Extraction of dilute metal ions from water was performed near room temperature with a variety of ionic liquids. Distribution coefficients are reported for fourteen metal ions extracted with ionic liquids containing cations 1-octyl-4-methylpyridinium[4MOPYR]+, 1-methyl-1-octylpyrrolidinium[MOPYRRO]+, or 1-methyl-1-octylpiperidinium[MOPIP]+ and anions tetrafluoroborate[BF4]+, trifluoromethyl sulfonate[TfO]+, or nonafluorobutyl sulfonate[NfO]+. Ionic liquids containing octylpyridinium cations are good for extracting mercury ions. However, other metal ions were not significantly extracted by any of these ionic liquids. Extractions were also performed with four new task-specific ionic liquids. When these liquids contain a disulfide functional group, they are efficient and selective for mercury and copper, whereas those containing a nitrile functional group are efficient and selective for silver and palladium.

The industrial production of complex biopharmaceuticals using recombinant mammalian cell lines is still mainly built on a quality by testing approach, which is represented by fixed process conditions and extensive testing of the end-product. In 2004 the FDA launched the process analytical technology initiative, aiming to guide the industry towards advanced process monitoring and better understanding of how critical process parameters affect the critical quality attributes. Implementation of process analytical technology into the bio-production process enables moving from the quality by testing to a more flexible quality by design approach. The application of advanced sensor systems in combination with mathematical modelling techniques offers enhanced process understanding, allows on-line prediction of critical quality attributes and subsequently real-time product quality control. In this review opportunities and unsolved issues on the road to a successful quality by design and dynamic control implementation are discussed. A major focus is directed on the preconditions for the application of model predictive control for mammalian cell culture bioprocesses. Design of experiments providing information about the process dynamics upon parameter change, dynamic process models, on-line process state predictions and powerful software environments seem to be a prerequisite for quality by control realization.