Mathematical models play a central role in biological systems engineering, supporting the generation of new testable hypotheses and novel ways of intervention, as well as providing mechanistic explanations of experimental results. Kinetic (i.e. dynamic) models are particularly important since they can explain and predict the functional behavior that emerges from the time-varying concentrations in cellular components. However, there are currently many pitfalls and challenges for kinetic model building:
algorithms for systematic model building techniques are still largely missing or not suitable for the noisy and partially observed nature of data in biology. Current model-building practices mostly ignore crucial steps such as identifiability analysis, optimal experimental design (OED) and uncertainty quantification, which are fundamental for a proper modeling cycle;
the existing algorithms and related software are fragmented: different communities have developed tools for different aspects of the modeling cycle, and thus no coherent and integrated framework exists that a broad community of biologists can apply in a straight forward manner;
application of existing algorithms requires expert modeling knowledge, making them extremely difficult to apply by biologists;
the existing models are unfit for engineering purposes: they can justify function but cannot automate the development of new systems (automation); they can uncover network structure from large data sets but cannot be utilized/employed for optimization and control strategies.
In an interdisciplinary joint effort of modelers, systems and software engineers, as well as experimentalists, the SyMBioSys consortium intends to address these challenges by providing an innovative research and training programme for young researchers.
The main objective of SyMBioSys is to provide a new generation of innovative and entrepreneurial early-stage researchers (ESRs) that will develop cutting-edge kinetic models for biological processes via systems engineering research and will exploit these for designing novel biotechnological applications. To achieve this, the ESRs projects will be integrated in such a way that all will collaboratively contribute to the building and usage of proper kinetic models of complex biological systems.
In this context, the SyMBioSys consortium is supporting the establishment of a leading EU research community at the interface of engineering, biological and computational sciences. The network brings together leading academic institutions (five universities and two research centres) with strong record in biological process systems engineering research; four companies, which are key players in development of simulation software for process systems engineering, metabolic engineering and industrial biotechnology; and two partner organizations with expertise in bioinformatics and synthetic biology, and management, innovation and entrepreneurship. SyMBioSys provides a training network for 15 ESRs spread over five research work packages (WPs 1-5), complemented by an intensive programme of taught courses and networking events. Each ESR will be involved in an Individual Research Project comprising partners from both Academia and Industry.The combination of industrial and academic partners will deliver scientific courses in Experimentation (Hypotheses Generation, Dynamic Measurements…), Engineering (Model-based Design, Optimization and Control…) and Product Development (Software Design) as well as transferable skills courses (entrepreneurship, IP management…).
There are still significant challenges and crossing barriers to overcome in biological systems engineering.
In order to overcome these barriers, the SyMBioSys consortium will focus its efforts on:
- developing proper kinetic models for complex biological systems, since the existing kinetic models in biology cannot yet deal with the true complexity of biological systems and have thus very limited predictive power;
- developing new algorithms for analysis and refinement of the kinetic models developed, including model inference and calibration, identifiability analysis, optimal experimental design and model reduction, and discrimination;
- developing new algorithms for exploitation of the new kinetic models, including dynamic optimization (to predict the dynamic behavior in biological networks) and mixed-integer optimal control (to identify design principles or optimal ways of intervention);
- developing high-level and user-friendly software tools, implementing the new algorithms above, to be used by a large community of biologists/experimentalists. Existing software developed by several of the industrial partners will be extended with the new methods to be developed by the academic partners; active user comments will feed back into the model and algorithm development to verify performance. The ultimate goal is to create computational workflows by extending and integrating existing state-of-the-art simulation and optimization software;
- real-life application of the kinetic models, methods and software developed to provide experimentally-verified solutions to real-life biotechnological and biomedical problems.
Funded by the Horizon 2020 Framework Programme of the European Union