Lack of reproducibility in science is a relevant problem (https://doi.org/10.1038/533452a). Methods for FAIR experiments and models can help solve it. FAIR in this context means findable, accessible, interoperable and reusable.

To increase the speed during the development of bioprocesses, e.g., the production of drugs in microbes, we apply mathematical models describing the microbial growth. Those allow us to extract more information from our experimental data, and to optimize the production of the desired active ingredients.

Models get better the more data we use to train them. However, reuse of data is limited due to the aforementioned difficulties in reproducing experiments as well as transferring knowledge between different process steps or laboratories.

We identified reasons on four levels in our automated bioprocesses:

1. Laboratory devices: 
   • incompatible interfaces
   • software cannot communicate
2. Scientists in the lab: 
   • reasoning and assumptions are not documented
3. Manually and automatically generated data:
   • missing (meta)data
   • different formats
   • not open access
   • not digital
4. Mechanistic and data-driven models:
   • unclear semantics
   • not all parameters and parameter estimation procedure recorded

The goal of the project FIERCS is the FAIR digitalisation in biotechnology. For this purpose, methods will developed that enable the acquisition of all data relevant to understand and replicate an experiment - from the planning, to the implementation by humans and machines, to the analysis of the data by software.

The challenges here are:

• Heterogenous data
• Must be transdisciplinary understandable (Scientists are from diverse fields: microbiology, chemical engineering, mathematics, computer science, ... )
• People, device and diverse algorithms as well as models work together

An ontology will developed to a clear and complete representation of all relationships in the research example. The expected results are:

• Improvement in the collaboration between humans and algorithms
• Traceable and unambiguous origin and progression of decisions
• More efficient data use
• Decrease in resource use
• Acceleration of process development

This project is funded as part of the DiGiTal graduate program, which brings together 13 scientists and artists from Berlin to advance digitalization and support women in their academic careers (https://www.tu.berlin/en/gleichstellung/digital).