Fast and Versatile Chromatography Process Design and Operation Optimization with the Aid of Artificial Intelligence

AI in analytical chemistry: Advancements, challenges, and future directions

This article explores the influence and applications of Artificial Intelligence (AI) in analytical chemistry, highlighting its potential to revolutionize the analysis of complex data sets and the development of innovative analytical methods. Additionally, it discusses the role of AI in interpreting large-scale data and optimizing experimental processes. AI has been fundamental in managing heterogeneous data and in advanced analysis of complex spectra in areas such as spectroscopy and chromatography. The article also examines the historical development of AI in chemistry, its current challenges, including the interpretation of AI models and the integration of large volumes of data. Finally, it forecasts future trends and the potential impact of AI on analytical chemistry, emphasizing the need for ethical and secure approaches in the use of AI.

A parameter estimation method for chromatographic separation process based on physics-informed neural network

Chromatographic separation processes are most often modeled in the form of partial differential equations (PDEs) to describe the complex adsorption equilibria and kinetics. However, identifying parameters in such a model requires substantial computational effort. In this work, a novel parameter estimation approach using a Physics-informed Neural Network (PINN) model is developed and tested for a binary component system. Numerical accuracy of our PINN model is confirmed by validating its simulations against those of the finite element method (FEM). Furthermore, model parameters in the kinetic model are estimated by the PINN model with sufficient accuracy from the observed data at the column outlet, where parameter fitting error can be reduced by up to 35.0 % from the conventional method. In a comparison with the conventional numerical method, our approach can reduce the computational time by up to 95 %. The robustness of the PINN model has also been demonstrated by estimating model parameters from noisy artificial experimental data.

Current state of implementation of in silico tools in the biopharmaceutical industry-Proceedings of the 5th modeling workshop

The fifth modeling workshop (5MW) was held in June 2023 at Favrholm, Denmark and sponsored by Recovery of Biological Products Conference Series. The goal of the workshop was to assemble modeling practitioners to review and discuss the current state, progress since the last fourth mini modeling workshop (4MMW), gaps and opportunities for development, deployment and maintenance of models in bioprocess applications. Areas of focus were four categories: biophysics and molecular modeling, mechanistic modeling, computational fluid dynamics (CFD) and plant modeling. Highlights of the workshop included significant advancements in biophysical/molecular modeling to novel protein constructs, mechanistic models for filtration and initial forays into modeling of multiphase systems using CFD for a bioreactor and mapped strategically to cell line selection/facility fit. A significant impediment to more fully quantitative and calibrated models for biophysics is the lack of large, anonymized datasets. A potential solution would be the use of specific descriptors in a database that would allow for detailed analyzes without sharing proprietary information. Another gap identified was the lack of a consistent framework for use of models that are included or support a regulatory filing beyond the high-level guidance in ICH Q8-Q11. One perspective is that modeling can be viewed as a component or precursor of machine learning (ML) and artificial intelligence (AI). Another outcome was alignment on a key definition for "mechanistic modeling." Feedback from participants was that there was progression in all of the fields of modeling within scope of the conference. Some areas (e.g., biophysics and molecular modeling) have opportunities for significant research investment to realize full impact. However, the need for ongoing research and development for all model types does not preclude the application to support process development, manufacturing and use in regulatory filings. Analogous to ML and AI, given the current state of the four modeling types, a prospective investment in educating inter-disciplinary subject matter experts (e.g., data science, chromatography) is essential to advancing the modeling community.

Scalable mRNA Machine for Regulatory Approval of Variable Scale between 1000 Clinical Doses to 10 Million Manufacturing Scale Doses

The production of messenger ribonucleic acid (mRNA) and other biologics is performed primarily in batch mode. This results in larger equipment, cleaning/sterilization volumes, and dead times compared to any continuous approach. Consequently, production throughput is lower and capital costs are relatively high. Switching to continuous production thus reduces the production footprint and also lowers the cost of goods (COG). During process development, from the provision of clinical trial samples to the production plant, different plant sizes are usually required, operating at different operating parameters. To speed up this step, it would be optimal if only one plant with the same equipment and piping could be used for all sizes. In this study, an efficient solution to this old challenge in biologics manufacturing is demonstrated, namely the qualification and validation of a plant setup for clinical trial doses of about 1000 doses and a production scale-up of about 10 million doses. Using the current example of the Comirnaty BNT162b2 mRNA vaccine, the cost-intensive in vitro transcription was first optimized in batch so that a yield of 12 g/L mRNA was achieved, and then successfully transferred to continuous production in the segmented plug flow reactor with subsequent purification using ultra- and diafiltration, which enables the recycling of costly reactants. To realize automated process control as well as real-time product release, the use of appropriate process analytical technology is essential. This will also be used to efficiently capture the product slug so that no product loss occurs and contamination from the fill-up phase is <1%. Further work will focus on real-time release testing during a continuous operating campaign under autonomous operational control. Such efforts will enable direct industrialization in collaboration with appropriate industry partners, their regulatory affairs, and quality assurance. A production scale-operation could be directly supported and managed by data-driven decisions.

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.

Toward Autonomous Production of mRNA-Therapeutics in the Light of Advanced Process Control and Traditional Control Strategies for Chromatography

mRNA-based therapeutics are predicted to have a bright future. Recently, a B2C study was published highlighting the critical bottlenecks of mRNA manufacturing. The study focused on supply bottlenecks of various chemicals as well as shortages of skilled personnel. The assessment of existing messenger ribonucleic acid (mRNA) vaccine processing shows the need for continuous manufacturing processes that are capable of about 80% chemical reduction and more than 70% personnel at factor five more efficient equipment utilization. The key technology to solve these problems is both a higher degree of automation and the maximization of process throughput. In this paper, the application of a quality-by-design process development approach is demonstrated, using process models as digital twins. Their systematic application leads to both robust optimized process parameters, with an increase in productivity of up to 108%, and sophisticated control concepts, preventing batch failures and minimizing the operating workload in terms of personnel and chemicals’ consumption. The approach thereby provides a data-driven decision basis for the industrialization of such processes, which fulfills the regulatory requirements of the approval authorities and paves the way for PAT integration. In the process investigated, it was shown that conventional PID-based controls can regulate fluctuations in the input streams sufficiently well. Model-based control based on digital twins may have potential above all in a further increase in productivity, but is not mandatory to implement for the industrialization of continuous mRNA manufacturing.

Process Automation and Control Strategy by Quality-by-Design in Total Continuous mRNA Manufacturing Platforms

Vaccine supply has a bottleneck in manufacturing capacity due to operation personnel and chemicals needed. Assessment of existing mRNA (messenger ribonucleic acid) vaccine processing show needs for continuous manufacturing processes. This is enabled by strict application of the regulatory demanded quality by design process based on digital twins, process analytical technology, and control automation strategies in order to improve process transfer for manufacturing capacity, reduction out-of-specification batch failures, qualified personnel training and number, optimal utilization of buffers and chemicals as well as speed-up of product release. In this work, process control concepts, which are necessary for achieving autonomous, continuous manufacturing, for mRNA manufacturing are explained and proven to be ready for industrialization. The application of the process control strategies developed in this work enable the previously pointed out benefits. By switching from batch-wise to continuous mRNA production as was shown in previous work, which was the base for this study, a potential cost reduction by a factor 5 (i.e., from EUR 0.380 per dose to EUR 0.085 per dose) is achievable. Mainly, based on reduction of personnel (factor 30) and consumable (factor 7.5) per campaign due to the significant share of raw materials in the manufacturing costs (74–97). Future research focus following this work may be on model-based predictive control to gain further optimization potential of potential batch failure and out of specification (OOS) number reduction.

Automation of Modeling and Calibration of Integrated Preparative Protein Chromatography Systems

With the increasing global demand for precise and efficient pharmaceuticals and the biopharma industry moving towards Industry 4.0, the need for advanced process integration, automation, and modeling has increased as well. In this work, a method for automatic modeling and calibration of an integrated preparative chromatographic system for pharmaceutical development and production is presented. Based on a user-defined system description, a system model was automatically generated and then calibrated using a sequence of experiments. The system description and model was implemented in the Python-based preparative chromatography control software Orbit.

Need for a Next Generation of Chromatography Models—Academic Demands for Thermodynamic Consistency and Industrial Requirements in Everyday Project Work

Process chromatography modelling for process development, design, and optimization as well as process control has been under development for decades. Still, the discussion of scientific potential and industrial applications needs is open to innovation. The discussion of next-generation modelling approaches starting from Langmuirian to steric mass action and multilayer or thermodynamic consistent real and ideal adsorption theory or colloidal particle adsorption approaches is continued.

Artificial Neural Network for Fast and Versatile Model Parameter Adjustment Utilizing PAT Signals of Chromatography Processes for Process Control under Production Conditions

Preparative chromatography is a well-established operation in chemical and biotechnology manufacturing. Chromatography achieves high separation performances, but often has to deal with the yield versus purity trade-off as the optimization criterium regarding through-put. The initial trade-off is often disturbed by the well-known phenomenon of chromatogram shifts over process lifetime, and has to be corrected by operators via adjustment of peak fraction cutting. Nevertheless, with regard to autonomous operation and batch to continuous processing modes, an advanced process control strategy is needed to identify and correct shifts from the optimal operation point automatically. Previous studies have already presented solutions for batch-to-batch variance and process control options with the aid of rigorous physico-chemical process modeling. These models can be implemented as distinct digital twins as well as statistical process operation data analyzers. In order to utilize such models for advanced process control (APC), the model parameters have to be updated with the aid of inline Process Analytical Technology (PAT) data to describe the actual operational status. This updating process also includes any operational change phenomena that occur, and its relation to their physico-chemical root cause. Typical phenomena are fluid dynamic changes due to packing breakage, channelling or compression as well as mass transfer and phase equilibrium-related separation performance decrease due to adsorbent aging or feed and buffer composition changes. In order to track these changes, an Artificial Neural Network (ANN) is trained in this work. The ANN training is in this first step, based on the simulation results of a distinct and previously experimentally validated process model. The model is implemented in the open source tool CasADi for Python. This allows the implementation of interfaces to process control systems, among others, with relatively low effort. Therefore, PAT signals can easily be incorporated for sufficient adjustment of the process model for appropriate process control. Further steps would be the implementation of optimization routines based on PAT and ANN predictions to derive optimal operation points with the model.

Machine Learning-Based Dynamic Modeling for Process Engineering Applications: A Guideline for Simulation and Prediction from Perceptron to Deep Learning

A misusage of machine learning (ML) strategies is usually observed in the process systems engineering literature. This issue is even more evident when dynamic identification is performed. The root of this problem is the gradient explode and vanishing issue related to the recurrent neural networks training. However, after the advent of deep learning, these issues were mitigated. Furthermore, the problem of data structuration is often overlooked during the machine learning model identification in this field. In this scenario, this work proposes a guideline for identifying ML models for the different applications in process systems engineering, which are usually for simulation or prediction purposes. While using the proposed guideline, the work also identifies a virtual analyzer for a pressure swing adsorption unit. In these types of adsorption separations, it is usual that the measurement of the main properties is not done online. Therefore, the virtual analyzer is another contribution of this manuscript. The overall results demonstrate that even though the test provides good performance during the ML model identification, its quality might degenerate over the application domain if the model application is overlooked.