We come to an age where new technologies and innovative thinking are profoundly changing biopharmaceutical drug production industry holistically, giving rise to “the factory of the future.” What will the factory of the future with respect to its structure, technologies, processes, and drug product delivery to the patient be? What enablers will manufacturers need in terms of strategy, leadership, employee skills, IT infrastructure, regulatory oversight, and suppliers to make this a reality? The factory of the future is an innovation vision on how manufacturers should enhance production by transforming manufacturing in three dimensions: plant structure, plant digitization, and plant processes.

Currently, “Industry 4.0” and “Digital Twins” are inflationary used words in the bioprocess industry, but not a single fully automated bioprocess for the production of biopharmaceuticals was filed so far. The talk will highlight the current limitations of digital twins but also demonstrate with several up- and downstream showcases the potential of bioprocess digital twins to accelerate process development and fully automated manufacturing.

Transgene optimization often takes a “one size fits all” approach that uses the same optimization process irrespective of a transgene’s desired expression profile. Rather, factors unique to each vector’s use case should be considered during optimization. We have developed the Expression Cassette Optimization (ECO) platform as an AI-driven optimization tool that considers the gene transfer vector, target tissue, and real-world manufacturing constraints to output optimized sequences tailored to each transgene’s desired expression profile. The ECO pipeline interfaces with our discovery-to-GMP pipeline to enable in-house production of high-titer vector by leveraging fixed bed bioreactor technology from the Corning Ascent platform.

Bioprocesses are highly complex, level of automation is moderate, and there is constant pressure on improving efficiency and costs. On top, climate change and shortage of resources dictate reduction of the environmental footprint of bioprocesses and manufacturing plants. Two extreme production scenarios are applied in biopharmaceutical industry: a fully-disposable factory with the characteristics of full flexibility and speed or a large-scale fixed plant with high capacity. Time ahead solutions and ideas will be presented how new processes and environmental friendliness can be married for the benefit of patient, supply reliability, and economics.

Live virus vaccine production often requires the use of adherent cell lines, which necessitates considerable hands-on manipulation and operator staffing. Quantifying key process attributes in real-time without sampling presents opportunities for process automation and the enrichment of process understanding. By conjunction of multiple capacitance frequencies and process outcomes, as well as application of multivariate correlative statistics, we intend to show the wealth of detail that can be extracted from a single bioreactor batch and how to apply that detail to process decision-making.

Subcutaneous injection of monoclonal antibodies (mAbs) and the use of devices are becoming more common options for delivering therapeutics. In order to reduce the volume being administered, desired antibody (mAb) concentrations can range from 100 mg/mL to over 300 mg/mL. High mAb concentrations, in traditional TFF processing, are problematic due to challenges that include: inaccurate mAb concentration measurements due to high viscosity retentate solutions, unrepresentative samples, limited dynamic range of common UV instruments and potentially variable excipient or buffer concentrations arising from asymmetric distribution across the TFF membrane. These challenges can further complicate downstream formulation, the control strategy of the step and exacerbate mAb stability issues at high concentrations. 

This presentation will highlight the use of integrated, real-time PAT techniques designed to address these new processing risks. We will provide examples demonstrating the use of real-time infrared and Raman spectroscopy to make accurate protein concentration measurements and monitor buffers and excipients change regardless of viscosity and protein concentration. In addition, we will also illustrate the use of in situ particle measurements to detect the onset of micron-sized particles during up-concentration.

Ensuring consistent high yields and product quality are key challenges in biomanufacturing. Even minor deviations in critical process parameters (CPPs) such as media and feed compositions can significantly affect product critical quality attributes (CQAs). In this regard, machine learning (ML) approaches offer immense potential in handling non-linear datasets and thus are able to identify new CPPs that could effectively predict the CQAs. ML techniques can also be synergized with mechanistic models as a ‘hybrid ML’ or ‘white box ML’ to identify how CPPs affect the product yield and quality mechanistically, thus enabling rational design and control of the bioprocess.

In this talk, we will highlight certain applications of machine learning for the holistic modeling of bioprocesses. As a new approach, we will shed light on the role of recurrent neural networks (RNNs) in predicting upstream cultivation processes. We will discuss the basics as well as the scientific implications for interpreting the results.

Whether it’s research for new drug discovery or ways of expediting time to market, many healthcare and life sciences companies are looking to cloud-based scientific computing to accelerate research and development. The applications are countless: from faster drug approvals to improved collaboration, from more rapid vaccine development to helping genomics companies meet strict compliance and security standards. The goal in every case: faster time to market, cost efficiency, performance enhancements, more automation, improved agility and collaboration, ongoing security and compliance, and most importantly the safety and efficacy of new treatments to improve patient outcomes.

Bionic Microcarriers-based Holistic Solution in Three-dimensional(3D) Cell Culture.

After discussing commonly applied scale-down technologies, the talk presents a newly developed single multi-compartment bioreactor (SCMB) that mimics large-scale mixing times and mechanical power input in a simple manner. A workflow is presented that enables the quantitative prediction of small-scale performance by using particularly-designed discs in a 3.5L stirred tank. The procedure could be successfully applied for aerated and non-aerated cultivation conditions of CHO cells.

Digital twins are perceived to be the core enabler for Pharma 4.0. So, if we have one, what is the benefit, where and how to deploy them? We show two applications: 1) The sound definition of the control strategy using integrated digital twins, which analyze the interactivities between unit operations. 2) The real-time deployment for direct optimization of specific productivity in the upstream process. Both increase productivity and smooth life-cycling.​

Detailed process-understanding is required to ensure high-quality products throughout the product development lifecycle. An innovative approach to achieving this balance is the application of digital twins in form of mechanistic modeling. This presentation shows the implementation of a digital twin concept in early-stage development including efficient workflows for fast model calibration and model application.

We present best practices for integrating lab instruments, apps, and data platforms, and building software to automate wet labs. This is the basis for Ganymede.bio, a modern cloud-based platform for wet lab integration and automation. We cover cases of common workflows such as flow cytometry, bioreactors, sequencing data, and LIMS/ELN tools, and share what approaches we've seen to how biotech companies build their data stack for bioprocessing, process development, and beyond.

Reproducing large molecules reliably at an industrial scale requires manufacturing capabilities with a high degree of sophistication as well as continued investments during commercialization. Understandably, these segments have been slow in adopting [innovative] technologies which are already well established in other areas & industries. As biopharma moves from niche to mainstream business, the industry is challenged to improve competitiveness and sustainability while ensuring high quality and delivery performance. We are pleased to invite you to our journey on harnessing the potential of data & digital to accelerate manufacturing efficiency. The talk will cover key topics related to data sourcing, data analysis, and deployment of visualization tools to make the output of advanced data & digital technologies accessible to relevant stakeholders within each department.

Several enabling technologies change the way we do pharmaceutical development: automation and digitalization, modeling and simulation. But especially the latter are not silver bullets, and it is important to have a closer look: why and when one might use them, which factors support a successful and sustainable implementation, and what it means for an organization to embrace these new ways of work.

Data-driven approaches offer a vast opportunity to leverage the large, rich datasets developed from bioprocessing systems. These approaches include statistical and artificial intelligence-based methods, but deep learning algorithms are perhaps best suited to the modeling, classification, prediction, and visualization of complex multivariate datasets. This presentation reviews the current opportunities, state-of-art, and barriers to the utilization of these approaches in bioprocess optimization. Specifically, techniques in dimensionality reduction, feature engineering, transfer learning, and various artificial neural networks will be presented. These techniques have the potential to reduce cost and increase scalability of development of new processes, materials, and product lines.

This session will focus on ways to accelerate analytical development to support faster development timelines. However, activities often performed during Phase 3, as assay validation, elucidation of structure and characterization of impurities, can sometimes generate surprises. This presentation is intended to provide recovery and bridging strategies when deficiencies are discovered late in the development lifecycle.

Monoclonal antibodies and novel bispecific DART molecules are being developed for a variety of indications including immuno-oncology. Sensitive, accurate, and high-throughput analytical techniques enable faster development timelines for therapeutic molecules. An overview of different approaches, methodologies, and supporting software for high-throughput process analytics and related challenges will be presented as case studies.

Clearance of residual host DNA is an important part of the biopharmaceutical process as host DNA can pose a potential risk to the patient. Using the KingFisher Presto integrated into a Hamilton liquid handling system, we have automated the entire residual DNA assay from sample preparation through qPCR plate preparation, significantly reducing FTE labor and allowing for a walk-up system for quicker turnaround and high-throughput for residual DNA results.