Fraunhofer Institute for Industrial Mathematics ITWMWe support startups, small and medium-sized enterprises (SMEs), and large corporations through our research as a development partner and provider of methodological expertise. Our goal is to accelerate, secure, and virtualize development and manufacturing processes in biotechnology, the pharmaceutical industry, and the life sciences. Our focus is on mathematically grounded models, explainable Artificial Intelligence (AI), and decision-relevant software prototypes – beyond black-box approaches.
Keyvisual Anwendungsfeld Biotechnologie, Pharma und Life Sciences
Acceleration of Biotech Processes Driven by Quality by Design
How can model-based approaches be used to accelerate and validate development and manufacturing processes in biotechnology, pharmaceuticals, and the life sciences? Our overview page highlights approaches, methods, and real-world application examples, as well as how we support companies.
© iStockphoto
Applied Mathematics, Models and Simulation for Biotechnology, Pharmaceuticals and Life Sciences
Model-Driven Acceleration of Processes in the Life Sciences
Using model-based methods, we help companies better understand complex biotechnological processes, shorten development times, and validate decisions at an early stage. In doing so, we combine mathematical modelling, data-driven analyses, hybrid models that integrate physics and Machine Learning, and simulation-based process development into practical solutions for research, development, and production.
Our Core Methodological Portfolio
Model-Based Process Development
We develop mechanistic and hybrid models for processes in biotechnology, pharmaceuticals, and medical technology. We combine a physical understanding of these processes with data-driven methods to provide robust and interpretable models for development, scaling, and production.
Our areas of focus include:
- Mixing and formulation processes
- Bioreactors and fermentation processes
- Chromatography, for example for binding, elution, and breakthrough curves
- Filtration and membrane-based separation
In addition, we provide support for parameter identification – even with scarce, incomplete, or noisy data. Our models are designed for traceability and regulatory compliance, creating a transparent basis for decision-making in research and production.
Model-Based Design of Experiments (Quality by Design)
Using model-based experimental design and Quality by Design methods, we help make development processes more efficient and robust. Through mathematical models and simulation-based analyses, we reduce the number of experiments required and accelerate development cycles.
In doing so, we provide support in areas such as:
- Design of Experiments (DoE) under real-world constraints
- Reducing the number of experiments and development costs
- Establishing design spaces
- Robustness and sensitivity analyses
This creates a solid basis for decision-making in process development, scaling, and regulatory documentation.
Optimization and Decision Support
We develop methods for the multi-criteria optimization of complex processes and help companies make informed decisions based on transparent models.
This includes, among other things:
- Multi-criteria optimization
- »What-if« and scenario analyses
- Transparent decision-making frameworks instead of pure automation
Our goal is to make complex relationships understandable and to support decision-making processes with data and models – without removing people from the process.
Hybrid AI and Agent-Based Development Cycles
We combine mathematical modeling with artificial intelligence methods to develop hybrid approaches for the next generation of data-driven development processes. In doing so, we specifically embed AI into physical and mechanistic models to create explainable, controllable, and robust systems.
Our development cycles follow an iterative approach:
Model → Experiment → Update → Decision
The focus is on transparent results and scientifically sound decision support – not black-box systems that cannot be interpreted.
It is not just about mathematics or artificial intelligence – but about the intelligent combination of both approaches.
What Sets Us Apart
| Many Providers | Fraunhofer ITWM |
| Black-Box-AI | Explainable Hybrid Models |
| Off-the-shelf Software | Customized Methods |
| Consulting without Action | Development Partner with Implementation |
| Full Automation | Dicision Support |
Rethinking Development Processes Together
We make processes easier to understand, decisions more robust, and development faster – based on mathematics, data, and domain expertise. We combine mathematical modeling, simulation-based methods, explainable AI, and a deep understanding of processes to create practical solutions for biotechnology, pharmaceuticals, and the life sciences. Our goal is to make complex relationships transparent and to help companies make well-informed and transparent decisions.
We see ourselves as a long-term development partner – not merely a software or consulting provider. Together with our partners, we develop new methods, test innovative approaches, and translate scientific insights into concrete applications for research, development, and production.
Let’s work together to accelerate development processes, reduce risks, and unlock new data-driven decision-making pathways.
Our Targer Audiences
Start-ups
Start-ups in Biotechnology and Life Sciences
Start-ups in the biotechnology, pharmaceutical, and life sciences sectors often face the challenge of having to make sound decisions with limited data, scarce resources, and significant time and financial pressures. At the same time, they need well-founded information for investors, partners, and regulatory authorities to validate development processes and enable growth.
Our Contribution: Reliable Decision-Making Foundations and Reduced Development Risks
We help young companies build reliable knowledge faster and reduce development risks early on. To do this, we develop robust minimal and effect models that deliver valuable insights even with limited data. Using model-based experimental design, we help reduce costly and time-consuming experiments in a targeted manner and make development processes more efficient.
In addition, we conduct feasibility, scaling, and risk analyses to provide a sound basis for decision-making regarding technology development, process transfer, and investment decisions. In this way, we lay the groundwork for faster yet reliable product and process development.
We help you build reliable knowledge faster – for confident decisions, more efficient processes, and shorter development times.
SMEs and CDMOs
SMEs and CDMOs Facing Increasing Demands
Small and medium-sized enterprises (SMEs) and Contract Development and Manufacturing Organizations (CDMOs) often possess extensive tacit process knowledge, yet they simultaneously face high testing costs and increasing regulatory requirements. Issues such as Quality by Design (QbD), traceability, and robust process control are becoming increasingly important.
Our Contribution: Model-Based Process Optimization
We provide support through model-based process development and optimization for upstream and downstream processes. Using mathematical models, Design of Experiments (DoE) approaches, and simulation-based design spaces, we create robust Quality by Design workflows that make development processes more efficient and regulatory-compliant.
In addition, we develop hybrid models that combine physical relationships with Machine Learning methods. This enables robust and traceable decision-making even in complex process environments. Our approach integrates seamlessly into existing workflows – without fundamentally disrupting operations.
Let’s work together to make your processes more efficient, robust, and scalable – from development through production!
Corporations
Corporations with Complex Processes and New Methods
Large corporations in the pharmaceutical, biotechnology, and life sciences industries operate with highly complex processes, vast amounts of data, and established organizational structures. At the same time, there is a growing need for new methods that go beyond established standard software and enable innovative development and decision-making processes.
Our Contribution: Developing and Testing New Methods for Complex Systems
We work with you to develop and test new methods for process development, process analysis, and decision support. Our focus is on scientifically sound approaches that can be flexibly adapted to existing structures and requirements.
Instead of off-the-shelf standard products, we develop prototypical implementations that allow new procedures, modeling approaches, and data-driven methods to be quickly evaluated and tested. At the same time, we support knowledge transfer and the integration of new methods into existing development and production environments.
Together with you, we develop the next generation of development and decision-making processes – transparent, traceable, and scientifically sound.
Application Examples from Biotechnology, Pharmaceuticals, and the Life Sciences
Real-World Use Cases
AAV-Manufacturing
AAV-Manufacturing Processes (Gene Therapy)
Background: Complex AAV Manufacturing and High Regulatory Uncertainty in Gene Therapy
The manufacturing of adeno-associated viruses (AAV) for gene therapies involves highly complex upstream, downstream, and cryopreservation processes. At the same time, only a limited number of costly experiments are available, while regulatory requirements and uncertainties are particularly high in early development phases.
Our Contribution: Model-Based Process Development and Experimental Design for Aav Production Processes
We develop mechanistic and hybrid process models for the various steps of AAV manufacturing – from upstream and downstream processes to cryopreservation. Using model-based experimental design in the spirit of Quality by Design (QbD), we help to specifically reduce the number of necessary experiments and efficiently allocate development resources.
In addition, we conduct robustness and sensitivity analyses to identify stable process windows and better understand critical process parameters.
Benefits: Faster Development and Validated Scale-up Decisions in Gene Therapy
Through our model-based approach, we accelerate process development and foster a deeper understanding of complex interrelationships within manufacturing processes. At the same time, we establish a robust basis for decision-making for GMP-compliant development stages as well as for further scaling and process validation.
Cryopreservation
Cryopreservation of Biological Products
Background: Uncertain Stability of Biological Products During Cryopreservation and Storage
In the cryopreservation of biological products – that is, the controlled storage of biological products at extremely low temperatures – there are often complex and difficult-to-trace interactions between process parameters, cryogenic media, and product stability. Traditional trial-and-error approaches in this context are time-consuming, costly, and often yield insights with limited generalizability.
Our Contribution: Model-Based Analysis and DoE for Cryopreservation Processes and Media Optimization
We analyze cryopreservation processes using model-based methods and combine mathematical models with Design of Experiments (DoE) approaches. Using DoE under real-world constraints, we systematically investigate the influence of media and process parameters on stability and product quality.
Furthermore, we quantify robustness and uncertainties to ensure that development decisions are transparent and data-driven.
Benefits: Reduced Experimentation and Transferable Insights for Biopharmaceutical Products
Our approaches help significantly reduce the number of experiments required and shorten development times. At the same time, we create transparent decision-making frameworks and generate transferable insights that can also be applied to other products and processes.
Synthesis Processes
Pharmaceutical Synthesis Processes
Background: Complex Chemical Syntheses with Scaling and Sustainability Requirements
Pharmaceutical synthesis processes are often characterized by a wide variety of variants, complex scaling issues, and increasing demands for efficiency, resource conservation, and sustainability. At the same time, process design and production conditions must be rigorously evaluated at an early stage.
Our Contribution: Multi-Criteria Optimization of Pharmaceutical Synthesis Processes
We combine mechanistic models with data-driven approaches to better understand and specifically optimize chemical synthesis processes. Using multi-criteria optimization, we simultaneously consider factors such as yield, product quality, costs, and energy consumption.
In addition, we conduct scenario and scaling analyses to transparently evaluate different process variants and identify risks associated with scaling up to larger production scales at an early stage.
Benefits: More Efficient Process Design and Reliable Scale-up Decisions
Our methods provide a sound basis for decision-making in process design and help shorten development times. At the same time, they increase the reliability and traceability of scale-up decisions and support more efficient and sustainable manufacturing processes.
Decision Support
Decision Support in Medical Technology and Clinical Settings
Background: Data-Driven Decision-Making Processes in Medical Technology and Clinical Settings Characterized by High Uncertainty
Decision-making processes in medical technology and clinical applications are often based on heterogeneous, sometimes incomplete data, while simultaneously being subject to significant professional and regulatory accountability. Transparency and traceability play a central role in this context.
Our Contribution: Explainable AI and Mathematical Models for Clinical Decision Support
We develop mathematical models and explainable AI methods to support decision-making in complex clinical and medical technology applications. By simulating different alternatives and scenarios, we create transparent decision-making frameworks for planning, evaluation, and risk assessment.
The focus is on comprehensible models and explainable results – not black-box approaches that cannot be interpreted.
Benefits: Transparent and Robust Decision-Making Processes in Medical Technology and Clinical Settings
Our approaches improve the quality of planning and decision-making and support robust, data-driven decisions while maintaining or increasing safety. At the same time, they promote transparency and trust in complex decision-making processes.
Chromatography
Improvement of Chromatography and Purification Processes
Background: High Experimental Effort in Chromatography and Protein Purification
The design and scaling of chromatography and purification processes often involve a significant amount of experimental work. In particular, evaluating loading capacity, breakthrough, elution, and process stability requires numerous time- and cost-intensive experimental runs.
Our Contribution: Model-Based Simulation and DoE for Chromatography Processes
We develop model-based simulations for chromatography and purification processes and systematically analyze loading, breakthrough, and elution behavior. In addition, we use Design of Experiments (DoE) methods to investigate material, media, and process parameters in a structured and efficient manner.
Furthermore, we quantify uncertainties and robustness to identify stable and traceable process windows.
Benefits: More Efficient Process Development and Robust Design Spaces for Biopharmaceutical Production
Our approaches reduce the number of necessary experiments and create reproducible design spaces for development and scaling. This increases confidence in scale-up decisions and makes process development more efficient and robust.
Process Optimization
Process Optimization for Bioreactors with Agitators
Background: Scaling Issues and Inhomogeneities in Agitated Bioreactors
In agitated bioreactors, issues such as inconsistent homogeneity, scaling effects, and quality risks in agitation, mixing, and aeration processes can affect process stability. Particularly during scale-up, uncertainties often arise regarding oxygen transfer, mass distribution, and reproducible process conditions.
Our Contribution: Flow and Mixing Models for Bioreactors and Oxygen Transfer
We combine flow, mixing, and parameter studies – for example, on oxygen transfer and homogeneity in bioreactors – with mechanistic models and real-world process data. This allows for the transparent analysis and evaluation of complex interactions within the processes.
In addition, we conduct scenario and sensitivity analyses to identify critical influencing factors early on and derive robust operating windows.
Benefits: Robust Bioreactor Scale-up Processes and Reduced Production Risks
Our model-based approaches support more robust process design and help reduce risks during scale-up. At the same time, they improve process understanding and provide a solid basis for decision-making in development and production.