Seminar by Dr Filippo Menolascina, Reader and Director of Industry Engagement, University of Edinburgh

Abstract

In this talk, I will present a computational and experimental framework to automate modelling in synthetic biology. Leveraging principles from System Identification, Real-Time Control, Optimal Experimental Design and in vivo experiments in microfluidics, I will show how we achieved automatic model selection and model calibration in both a frequentist and Bayesian framework in E. coli and S. cerevisiae. Given a set experimental budget (e.g., experiment hours), our results show that online Optimal Experimental Design can reduce parameter estimation errors by 80% compared to traditional, intuition-driven inputs. I will briefly discuss how our activities in the space of lab automation led to spinning out OGI Bio, an Edinburgh-based start-up focusing on smart bioreactors.

I will bring together online identification and control to introduce the concept of Cybergenetic Twin: a hybrid biological and digital twin of bioprocesses that overcomes the historical limitations of Model Predictive Control in biochemical engineering by explicitly using real-time information on intracellular dynamics. Using data from a small-scale Taxol (Paclitaxel®) precursor bioproduction trial, I will show how data from our Cybergenetic Twin can be used to train a Neural Optimal Controller that outperforms the state of the art in maximising yield, a Critical Quality Attribute. I will conclude by reflecting on how Cybergenetic Twins can help Contract Development and Manufacturing Organisations establish a Quality by Design framework that enables Real Time Release Testing of bioproducts which, at once, strengthens the economics of biomanufacturing for plant operators and allows medicines to become immediately available to our ultimate stakeholder: society.

 
Biography

Filippo Menolascina is a Reader at the University of Edinburgh, the Director of Industry Engagement of the School of Engineering and an Investment Executive in the University in house Venture Capital fund, Old College Capital. He is also co-founder of OGI Bio, a University of Edinburgh spin-out developing modular smart bioreactors that recently closed its seed round at a £4M valuation. An Electrical Engineer and Computer Scientist by training (BSc ’06, MSc ’08), Dr Menolascina obtained his PhD in 2011 by defending a thesis that provided the first demonstration of in vivo real-time control of a complex synthetic gene network. His doctoral work pioneered the field now known as cybergenetics. As a postdoc at the Massachusetts Institute of Technology Dr Menolascina extended these results to the control of complex traits emerging from biomolecular networks, providing the first demonstration of real-time control of aerotaxis in B. subtilis.

Dr Menolascina, a former EPSRC Innovation Fellow, leads the University of Edinburgh’s cSynBioSys group which combines in silico methods and in vivo experiments to pursue two complementary goals: elucidating the design principles of living systems - i.e., understanding the instruction-set of life, and using these primitives to re-program cells. The overarching goal of the cSynBioSys group is to develop a Model-Based Biosystem Engineering framework for Engineering Biology, i.e., a model-based approach to automating the design of synthetic circuits that aims to make engineered cells as easy to build and program as computers are today.