Root cause investigation of deviations in protein chromatography based on mechanistic models and artificial neural networks

Regression Metamodel-Based Digital Twin for an Industrial Dynamic Crossflow Filtration Process

Dynamic crossflow filtration (DCF) is the state-of-the-art technology for solid-liquid separation from viscous and sensitive feed streams in the food and biopharma industry. Up to now, the potential of industrial processes is often not fully exploited, because fixed recipes are usually applied to run the processes. In order to take the varying properties of biological feed materials into account, we aim to develop a digital twin of an industrial brownfield DCF plant, allowing to optimize setpoint decisions in almost real time. The core of the digital twin is a mechanistic-empirical process model combining fundamental filtration laws with process expert knowledge. The effect of variation in the selected process and model parameters on plant productivity has been assessed using a model-based design-of-experiments approach, and a regression metamodel has been trained with the data. A cyclic program that bidirectionally communicates with the DCF asset serves as frame of the digital twin. It monitors the process dynamics membrane torque and transmembrane pressure and feeds back the optimum permeate flow rate setpoint to the physical asset in almost real-time during process runs. We considered a total of 24 industrial production batches from the filtration of grape juice from the years 2022 and 2023 in the study. After implementation of the digital twin on site, the campaign mean productivity increased by 15% over the course of the year 2023. The presented digital twin framework is a simple example how an industrial established process can be controlled by a hybrid model-based algorithm. With a digital process dynamics model at hand, the presented metamodel optimization approach can be easily transferred to other (bio)chemical processes.

Parameter-by-parameter estimation method for adsorption isotherm in hydrophobic interaction chromatography

Hydrophobic interaction chromatography (HIC) is used as a critical polishing step in the downstream processing of biopharmaceuticals. Normally the process development of HIC is a cumbersome and time-consuming task, and the mechanical models can provide a powerful tool to characterize the process, assist process design and accelerate process development. However, the current estimation of model parameters relies on the inverse method, which lacks an efficient and logical parameter estimation strategy. In this study, a parameter-by-parameter (PbP) method based on the theoretical derivation and simplifying assumptions was proposed to estimate the Mollerup isotherm parameters for HIC. The method involves three key steps: (1) linear regression (LR) to estimate the salt-protein interaction parameter and the equilibrium constant; (2) linear approximation (LA) to estimate the stoichiometric parameter and the maximum binding capacity; and (3) inverse method to estimate the protein-protein interaction parameter and the kinetic coefficient. The results indicated that the LR step should be used for dilution condition (loading factor below 5%), while the LA step should be conducted when the isotherm is in the transition or nonlinear regions. Six numerical experiments were conducted to implement the PbP method. The results demonstrated that the PbP method developed allows for the systematic estimation of HIC parameters one-by-one, effectively reducing the number of parameters required for inverse method estimation from six to two. This helps prevent non-identifiability of structural parameters. The feasibility of the PbP-HIC method was further validated by real-world experiments. Moreover, the PbP method enhances the mechanistic understanding of adsorption behavior of HIC and shows a promising application to other stoichiometric displacement model-derived isotherms.

Machine learning in bioprocess development: from promise to practice

Fostered by novel analytical techniques, digitalization, and automation, modern bioprocess development provides large amounts of heterogeneous experimental data, containing valuable process information. In this context, data-driven methods like machine learning (ML) approaches have great potential to rationally explore large design spaces while exploiting experimental facilities most efficiently. Herein we demonstrate how ML methods have been applied so far in bioprocess development, especially in strain engineering and selection, bioprocess optimization, scale-up, monitoring, and control of bioprocesses. For each topic, we will highlight successful application cases, current challenges, and point out domains that can potentially benefit from technology transfer and further progress in the field of ML.

Hybrid model development for parameter estimation and process optimization of hydrophobic interaction chromatography

Hydrophobic Interaction Chromatography (HIC) is often employed as a polishing step to remove aggregates for the purification of therapeutic proteins in the biopharmaceutical industry. To accelerate the process development and save the costs of performing time- and resource-intensive experiments, advanced model-based process design and optimization are necessary. Due to the unclear adsorption mechanism of the salt-dependent interaction between the protein and resin, the development of an accurate mechanistic model to describe the complex HIC behavior is challenging. In this work, an isotherm derived from Wang et al. is modified by adding three extra parameters together with an equilibrium dispersive model to represent the HIC process. To reduce the development effort of isotherm equations and extract missing information from the available data, a hybrid model is constructed by combining a simple and well-known multi-component Langmuir isotherm (MCL) with a neural network (NN). It is observed that the structure of the hybrid model is of critical importance to the accuracy of the developed model. During parameter estimation, a regularization strategy is incorporated to prevent overfitting. Furthermore, the impact of NN structures and regularization rates are comprehensively investigated. One of the interesting findings was that a simple NN with one hidden layer with two nodes and sigmoid as the activation function, significantly outperforms the mechanistic model, with a 62% improvement in accuracy in calibration and 31.4% in validation. To ensure the generalizability of the developed hybrid model, an in-silico dataset is generated using the mechanistic model to test the extrapolation capability of the hybrid model. Process optimization is also carried out to find the optimal operating conditions under product quality constraints using the developed hybrid model.

Artificial intelligence and machine learning applications in biopharmaceutical manufacturing

Artificial intelligence and machine learning (AI-ML) offer vast potential in optimal design, monitoring, and control of biopharmaceutical manufacturing. The driving forces for adoption of AI-ML techniques include the growing global demand for biotherapeutics and the shift toward Industry 4.0, spurring the rise of integrated process platforms and continuous processes that require intelligent, automated supervision. This review summarizes AI-ML applications in biopharmaceutical manufacturing, with a focus on the most used AI-ML algorithms, including multivariate data analysis, artificial neural networks, and reinforcement learning. Perspectives on the future growth of AI-ML applications in the area and the challenges of implementing these techniques at manufacturing scale are also presented.

The use of predictive models to develop chromatography-based purification processes

Chromatography is the workhorse of biopharmaceutical downstream processing because it can selectively enrich a target product while removing impurities from complex feed streams. This is achieved by exploiting differences in molecular properties, such as size, charge and hydrophobicity (alone or in different combinations). Accordingly, many parameters must be tested during process development in order to maximize product purity and recovery, including resin and ligand types, conductivity, pH, gradient profiles, and the sequence of separation operations. The number of possible experimental conditions quickly becomes unmanageable. Although the range of suitable conditions can be narrowed based on experience, the time and cost of the work remain high even when using high-throughput laboratory automation. In contrast, chromatography modeling using inexpensive, parallelized computer hardware can provide expert knowledge, predicting conditions that achieve high purity and efficient recovery. The prediction of suitable conditions in silico reduces the number of empirical tests required and provides in-depth process understanding, which is recommended by regulatory authorities. In this article, we discuss the benefits and specific challenges of chromatography modeling. We describe the experimental characterization of chromatography devices and settings prior to modeling, such as the determination of column porosity. We also consider the challenges that must be overcome when models are set up and calibrated, including the cross-validation and verification of data-driven and hybrid (combined data-driven and mechanistic) models. This review will therefore support researchers intending to establish a chromatography modeling workflow in their laboratory.

Exploring the potential of machine learning for more efficient development and production of biopharmaceuticals

Principles of Industry 4.0 direct us to predict how pharmaceutical operations and regulations may exist with automation, digitization, artificial intelligence (AI), and real time data acquisition. Machine learning (ML), a sub-discipline of AI, involves the use of statistical tools to extract the desired information either through understanding the underlying patterns in the information or by development of mathematical relationships among the critical process parameters (CPPs) and critical quality attributes (CQAs) of biopharmaceuticals. ML is still in its infancy for directly supporting the quality-by-design based development and manufacturing of biopharmaceuticals. However, adoption of ML-based models in place of conventional multi-variate-data-analysis (MVDA) is increasing with the accumulation of large-scale data. This has been majorly contributed by the real-time monitoring of process variables and quality attributes of products through the implementation of process analytical technology in biopharmaceutical manufacturing. All aspects of healthcare, from drug design to product distribution, are complex and multidimensional. Thus, ML-based approaches are being applied to achieve sophistication, accuracy, flexibility and agility in all these areas. This review discusses the potential of ML for addressing the complex issues in diverse areas of biopharmaceutical development, such as biopharmaceuticals design and assessment of early stage development, upstream and downstream process development, analysis, characterization and prediction of post translational modifications (PTMs), formulation and stability studies. Moreover, the challenges in acquisition, cleaning and structuring the bioprocess data, which is one of the major hurdles in implementation of ML in biopharma industry, have also been discussed. Regulatory perspectives on implementation of AI/ML in the biopharma sector have also been briefly discussed. This article is a bird's eye view on the recent developments and applications of ML in overcoming the challenges for adopting "Industry - 4.0" in the biopharma industry.

Smart Process Development: Application of machine-learning and integrated process modeling for inclusion body purification processes

The development of a biopharmaceutical production process usually occurs sequentially, and tedious optimization of each individual unit operation is very time consuming. Here, the conditions established as optimal for one step serve as input for the following step. Yet, this strategy does not consider potential interactions between a priori distant process steps and therefore cannot guarantee for optimal overall process performance. To overcome these limitations, we established a SMART approach to develop and utilize integrated process models using machine learning techniques and genetic algorithms. We evaluated the application of the data-driven models to explore potential efficiency increases and compared them to a conventional development approach for one of our development products. First, we developed a data-driven integrated process model using Gradient Boosting Machines and Gaussian Processes as machine learning techniques and a genetic algorithm as recommendation engine for two downstream unit operations, namely solubilization and refolding. Through projection of the results into our large-scale facility, we predicted a two-fold increase in productivity. Second, we extended the model to a three-step model by including the capture chromatography. Here, depending on the selected baseline-process chosen for comparison, we obtained between 50 and 100% increase in productivity. These data show the successful application of machine learning techniques and optimization algorithms for downstream process development. Finally, our results highlight the importance of considering integrated process models for the whole process chain, including all unit operations.

Modeling the impact of amino acid substitution in a monoclonal antibody on cation exchange chromatography

A vital part of biopharmaceutical research is decision making around which lead candidate should be progressed in early-phase development. When multiple antibody candidates show similar biological activity, developability aspects are taken into account to ease the challenges of manufacturing the potential drug candidate. While current strategies for developability assessment mainly focus on drug product stability, only limited information is available on how antibody candidates with minimal differences in their primary structure behave during downstream processing. With increasing time-to-market pressure and an abundance of monoclonal antibodies (mAbs) in development pipelines, developability assessments should also consider the ability of mAbs to integrate into the downstream platform. This study investigates the influence of amino acid substitutions in the complementarity-determining region (CDR) of a full-length IgG1 mAb on the elution behavior in preparative cation exchange chromatography. Single amino acid substitutions within the investigated mAb resulted in an additional positive charge in the light chain (L) and heavy chain (H) CDR, respectively. The mAb variants showed an increased retention volume in linear gradient elution compared with the wild-type antibody. Furthermore, the substitution of tryptophan with lysine in the H-CDR3 increased charge heterogeneity of the product. A multiscale in silico analysis, consisting of homology modeling, protein surface analysis, and mechanistic chromatography modeling increased understanding of the adsorption mechanism. The results reveal the potential effects of lead optimization during antibody drug discovery on downstream processing.

Hybrid Models for the simulation and prediction of chromatographic processes for protein capture

The biopharmaceutical industries are continuously faced with the pressure to reduce the development costs and accelerate development time scales. The traditional approach of heuristic-based or platform process-based optimization is soon getting obsolete, and more generalized tools for process development and optimization are required to keep pace with the emerging trends. Thus, advanced model-based methods that can reduce the can ensure accelerated development of robust processes with minimal experiments are necessary. Though mechanistic models for chromatography are quite popular, their success is limited by the need to have accurate knowledge of adsorption isotherms and mass transfer kinetics. As an alternative, in this work, a hybrid modeling approach is proposed. Thereby, the chromatographic unit behavior is learned by a combination of neural network and mechanistic model while fitting suitable experimental breakthrough curves. Since this approach does not require identifying suitable mechanistic assumptions for all the phenomena, it can be developed with lower effort. Thus, allowing the scientists to concentrate their focus on process development. The performance of the hybrid model is compared with the mechanistic Lumped kinetic Model for in-silico data and experiments conducted on a system of industrial relevance. The flexibility of the hybrid modeling approach results in about three times higher accuracies compared to Lumped Kinetic Model. This is validated for five different isotherm models used to simulate data, with the hybrid model showing about two to three times lower prediction errors in all the cases. Not only in prediction, but we could also show that the hybrid model is more robust in extrapolating across process conditions with about three times lower error than the LKM. Additionally, it could be demonstrated that an appropriately tailored formulation of the hybrid model can be used to generate representations for the underlying principles such as adsorption equilibria and mass transfer kinetics.

Developments and opportunities in continuous biopharmaceutical manufacturing

Today’s biologics manufacturing practices incur high costs to the drug makers, which can contribute to high prices for patients. Timely investment in the development and implementation of continuous biomanufacturing can increase the production of consistent-quality drugs at a lower cost and a faster pace, to meet growing demand. Efficient use of equipment, manufacturing footprint, and labor also offer the potential to improve drug accessibility. Although technological efforts enabling continuous biomanufacturing have commenced, challenges remain in the integration, monitoring, and control of traditionally segmented unit operations. Here, we discuss recent developments supporting the implementation of continuous biomanufacturing, along with their benefits.

Mechanistic Modeling of Preparative Column Chromatography for Biotherapeutics

Chromatography has long been, and remains, the workhorse of downstream processing in the production of biopharmaceuticals. As bioprocessing has matured, there has been a growing trend toward seeking a detailed fundamental understanding of the relevant unit operations, which for some operations include the use of mechanistic modeling in a way similar to its use in the conventional chemical process industries. Mechanistic models of chromatography have been developed for almost a century, but although the essential features are generally understood, the specialization of such models to biopharmaceutical processing includes several areas that require further elucidation. This review outlines the overall approaches used in such modeling and emphasizes current needs, specifically in the context of typical uses of such models; these include selection and improvement of isotherm models and methods to estimate isotherm and transport parameters independently. Further insights are likely to be aided by molecular-level modeling, as well as by the copious amounts of empirical data available for existing processes.

Distinct and Quantitative Validation Method for Predictive Process Modelling in Preparative Chromatography of Synthetic and Bio-Based Feed Mixtures Following a Quality-by-Design (QbD) Approach

Process development, especially in regulated industries, where quality-by-design approaches have become a prerequisite, is cost intensive and time consuming. A main factor is the large number of experiments needed. Process modelling can reduce this number significantly by replacing experiments with simulations. However, this requires a validated model. In this paper, a process and model development workflow is presented, which focuses on implementing, parameterizing, and validating the model in four steps. The presented methods are laid out to gain, create, or generate the maximum information and process knowledge needed for successful process development. This includes design of experiments and statistical evaluations showing process robustness, sensitivity of target values to process parameters, and correlations between process and target values. Two case studies are presented. An ion exchange capture step for monoclonal antibodies focusing on high accuracy and low feed consumption; and one case study for small molecules focusing on rapid process development, emphasizing speed of parameter determination.