Smart guide to synthetic biology

Researchers develop user-friendly software system to improve biological systems

Machine learning is transforming all areas of biological science and industry, but is typically reserved for only a few users in a limited number of scenarios. A research team at the Max Planck Institute for Terrestrial Microbiology led by Tobias Erb has developed METIS, a modular software system for optimizing biological systems. The research team demonstrates its user-friendliness and versatility using a large number of biological examples.

Design and optimization of biological systems are indispensable, especially in biotechnology and synthetic biology, but machine learning is now used in all areas of biology. However, it is obvious that the application and improvement of algorithms, ie calculation methods consisting of lists of statements, is not easily accessible. The reasons for this are not only a lack of programming knowledge, but also often insufficient experimentally determined data. At the interface between computational and experimental work, efficient approaches are therefore needed to close the gap between machine learning algorithms and their applications in biological systems.

Now a team at the Max Planck Institute for Terrestrial Microbiology led by Tobias Erb has succeeded in democratizing machine learning. In their current publication, the team presents its tool METIS together with cooperation partners at the INRAe Institute in Paris. Its application is so versatile and modular that it requires no computer knowledge and can be set up on different biological systems and with different laboratory devices. METIS is the abbreviation for Machine-learning Guided Experimental Trials for Improvement of Systems and named after the ancient goddess of wisdom and craft Μῆτις, lit. “wise advice”.

Less data required

Active learning, also known as optimal experimental design, uses machine learning algorithms to interactively suggest the next experimental design based on previous results – a valuable approach in the research lab, especially when a limited amount of experimentally labeled data is available. However, one of the biggest bottlenecks here is the amount of data, which is not always sufficient to train machine learning models. “While active learning already reduces the need for experimental data, we went a step further and examined different machine learning algorithms. Fortunately, we found a model that is even less dependent on data,” explains Amir Pandi, one of the lead authors the study.

To demonstrate the versatility of METIS, the team used it for a range of applications, including protein production optimization, genetic blueprints, combinatorial control of enzyme activity, and a complex CO2-Fixation metabolic cycle (CETCH cycle). For the CETCH cycle, they examined a combinatorial space of 1,025 conditions with only 1,000 experimental conditions and were able to find the most efficient CO2– Demonstrate the fixation cascade that has been described so far.

Optimization of biological systems

For application, the study offers novel tools to democratize and advance current developments in biotechnology, synthetic biology, genetic circuit design and metabolic engineering. “METIS enables researchers to optimize their already discovered or synthesized biological systems,” says Christoph Diehl, co-first author of the study. “But it’s also a combinatorial guide for exploring complex interactions and hypothesis-driven optimization. And what’s probably the biggest advantage: It can be a very helpful system for prototyping novel systems that don’t yet occur in nature.”

METIS is a modular tool that runs as a Google Colab Python notebook and can be accessed via a personal copy in a web browser, with no installation, registration or local processing required. The materials provided in the publication can help users to customize METIS for their applications.

#Smart #guide #synthetic #biology

Leave a Comment

Your email address will not be published.