BIOECODBTL has as its General Objective the development of a modeling and optimization framework for the Design-Build-Test-Learn cycle of the industrial biomanufacturing pipeline, increasing in that way the efficiency, flexibility, and agility of the pipeline in the context of the bioeconomy, allowing a more sustainable use of biodiversity resources.

BIOECODBTL will achieve its overall goal by implementing the following Scientific and Technical Specific Objectives:

1. Determine the circular bioeconomy metabolic design space.

   We want to address the questions of what compounds can be produced in a circular bioeconomy fashion through biobased biomanufacturing? What are the expected titers? To that end, we will perform a modular expansion of the metabolic space of the microbial host through a reaction rules-based algorithm to determine the full set of external production circuits (the plant) that can be plugged into the desired host and molecules that can be reached through production. Moreover, we will implement an algorithm for the enumeration of the production pathways in the resulting graph. We will also compute the modular expansion of the host metabolic space in order to determine the molecules that can be detected through biosensor circuits. In that way, we will provide the enumeration of the biosensor pathways in the resulting graph. Similarly, we will perform the enumeration of the potential regulated pathways in a host through the combination of a production and a biosensor circuit.

2. Build machine-learning approaches for predicting bio-based targets.

   Based on deep-learning approaches and molecular fingerprints, we will develop predictive frameworks for bio-based chemicals, including antioxidant and nutritional activity, antibacterial, therapeutic, as well as in developing novel materials with enhanced properties. The approach will consist of computing a molecular similarity metric between the molecules and their embedding in a space of desired dimensionality, streamlining the process for deep-learning approaches.

3. Digital twin of the circular biomanufacturing pipeline for digitalization of the process.

   As shown preliminarily in our studies, we will explore the biodesign space of genetic parts that can be implemented in bioproduction processes, including biosensors and dynamic regulation through optimal design of experiments, in order to predict and identify the best targets to be implemented in a bioproduction pathway. To that end, we will implement a Digital Twin of the full Design-Build-Test-Learn approach based on the modelling of the tasks and interactions between the steps in the process. Such an approach will involve both dynamic modeling of the biological and fermentation processes through differential equations, as well as discrete evento simulation of the steps of the biomanufacturing chain. In that way, our tool will assist biomanufacturing facilities to optimize their industrial processes.