Speakers
Tim Rudge
School of Computing, Newcastle University
An experimental and computational synthetic biologist with an interest in engineering gene regulatory networks for controlling the behaviour of bacterial biofilms.
AI-assisted Deep Learning for designing gene regulatory networks
Gene regulatory networks (GRNs), and specifically transcription networks have been engineered with a variety of behaviours that can be used for purposes including programmed therapeutic delivery, control of bioprocesses, and optimising metabolic pathways. I will present our work on design of data infrastructures for engineering GRNs, and show how we are using AI to make these infrastructures accessible to a broad range of users (both human and machine). I will demonstrate how our approach allows the use of natural language to train Deep Learning models that can help to predict GRN behaviour.
Thomas E. Gorochowski
School of Biological Sciences, University of Bristol, UK
Thomas Gorochowski is a Professor of Biological Engineering at the University of Bristol, holds a Royal Society University Research Fellowship, Turing Fellowship, is Co-Director of the Bristol BioDesign Institute and Deputy Director of the BBSRC/EPSRC Engineering Biology Centre of Doctoral Training. Coming from a background in computer science, he has since transitioned into the area of bioengineering, working across industry as a Marie Curie Fellow at DSM in the Netherlands and academia as part of the Synthetic Biology Centre at the Massachusetts Institute of Technology in the USA. In 2016, he founded the Biocompute Lab at the University of Bristol, which aims to better understand the computational architecture of biological systems to enable the more rationally reprogramming of biology. He currently leads the national UKRI CYBER Engineering Biology Mission Award to help de-risk environmental applications of synthetic biology and is the Bristol lead for the EPSRC EEBio Programme Grant that aims to improve the robustness of synthetic biology through the use of control engineering approaches.
Abstract
High-throughput experiments are revolutionising our understanding of biological complexity and offer a rich foundation on which to establish data-centric and mechanistic models of living cells. In this talk, I will present some of the sequencing methodologies my group has been developing to aid in the reprogramming of cells, providing broad and detailed information about cell-wide transcriptional and translational processes and some of the insights this type of data has gleaned. I will also, discuss some of the challenges associated when working with heterogeneous data for biological design and efforts we have contributed to (e.g., Synthetic Biology Open Language) to improve data interoperability across the field of engineering biology.
Carmen Mei-Ling Frischknecht-Gruber
Zurich University of Applied Sciences – School of Engineering
Carmen Frischknecht-Gruber is a research associate in the Safety Critical Systems Research Lab, which is part of the Institute of Applied Mathematics and Physics at the Zurich University of Applied Sciences (ZHAW). She holds an MSc in Artificial Intelligence from the University of Bath. Her work focuses on safety and functional safety in AI and machine learning, with a particular interest in using artificial life (ALife) methods to enhance the robustness and reliability of AI systems. To ensure transparency, she also applies explainable AI (XAI) techniques as part of her broader research efforts. In addition to her research, Frischknecht-Gruber actively works on the certification of AI systems, providing guidance on how to meet current standards and prepare for upcoming regulations such as the EU AI Act. She has worked on projects involving anomaly detection and predictive maintenance to improve overall system safety and reliability, contributing to the development of AI technologies that are both innovative and safe.
Dr. Mathias Weyland
Zurich University of Applied Sciences – School of Engineering
Matt Weyland is a Senior Lecturer at the Zurich University of Applied Sciences, specializing in interdisciplinary research at the intersection of Computer Science, Mathematics, Physics, Medicine, and Biology. He holds a Master's degree in Bioinformatics and completed his studies at the AILab of the University of Zurich, where he honed his expertise in robotics, computer vision, computational modeling and data analysis. Following his academic training, Matt gained valuable industry experience by developing lab equipment and robotics for the biomechanical field, further deepening his understanding of practical applications in science and technology. Afterward, Matt returned to academia and pursued a PhD in Chemistry, where he developed models and simulations to predict how cancer cells react to various therapies and how data acquired in the lab can be used to calibrate these models.
At the Zurich University of Applied Sciences, Matt continues to contribute to interdisciplinary research. He collaborates on projects requiring the development of models to predict cellular responses to cancer treatments, ensuring image quality in CT scan protocols, and exploring the interactions between artificial cells. Through these projects, Matt plays a crucial role in advancing scientific understanding across multiple domains.
Trustworthy Machine Learning in Safety-Critical Environments: Challenges and Strategies
As machine learning becomes increasingly integral to fields such as biological research and clinical applications, ensuring the safety and reliability of these systems is paramount. This presentation delves into the challenges of deploying machine learning in safety-critical environments, with a particular focus on bioengineering and medical applications. We explore strategies to enhance the trustworthiness of machine learning models, emphasizing robust development practices, risk assessment, and compliance with emerging regulatory frameworks, such as the EU AI Act. By drawing on collaborative initiatives with industry partners, we examine the roles, expectations, and challenges associated with the implementation of machine learning in a safe and responsible manner. In particular, we showcase how explainable AI (XAI) techniques can address key challenges and improve model reliability when applied to image quality assurance in medical computer tomography scans.
Our aim is to provide a comprehensive overview of the methodologies required to develop effective, interpretable machine learning models that meet the stringent demands of safety-critical applications while adhering to regulatory standards. In addition, we emphasize the importance of collaboration between research institutions, industry, and regulatory bodies to foster safe and innovative advancements in machine learning.
Dr Annalisa Occhipinti
Associate Professor, Teesside University
Dr Annalisa Occhipinti is an Associate Professor at Teesside University, where she works at the intersection between AI and Biology. Her current research topics involve the integration of metabolic modelling and multi-modal deep learning techniques for biomedical and bioprocessing applications.
Dr Occhipinti has recently received several awards for her outstanding contributions in the field of machine learning, big data, and cancer research, including an award from the Italian Embassy and the University of Cambridge. Dr Occhipinti is currently leading various research projects funded by BBSRC and EPSRC, and she extends her proficiency in AI for bioprocessing optimisation to collaborate with industrial partners and gain valuable data-driven insights.
Multimodal AI and metabolic modelling: a comprehensive framework for enhancing biological insights
Despite significant progress in the integration of AI and advanced medicine, the design and development of accurate and biologically interpretable models remain challenging and with a low probability of success in clinical practice. In this talk, Annalisa will introduce the latest AI and ML strategies that aim to improve cancer diagnosis and treatment by integrating data from multiple sources (e.g., transcriptomic, imaging, and clinical data) with metabolic modelling to generate accurate and biologically interpretable predictions. Using breast cancer as a paradigm, Annalisa will discuss the status of AI methodologies in cancer and synthetic biology, including the current challenges and opportunities for development.
Bo Wang, Ph.D.
National Cancer Institute, Laboratory of Pathology, US
Dr. Bo Wang currently holds appointment as Staff Scientist from National Cancer Institute (Bethesda, US) and an honorary title as Early Career Investigator from National Institute of Biomedical Imaging and Bioengineering (Bethesda, US). He also serves as Associate Editor for Computational and Structural Biotechnology Journal. As computational biophysicist, his research broadly engages molecular biology, system biology, genetics and genomics. He has demonstrated expertise in multiscale modeling with special interests in XAI and Digital Twins.
Abstract
A thorough understanding of the mechanistic insights of biological systems and key features of diseases are essential in biomedical research and clinical practice. Whereas various lab renovations and clinical breakthroughs have opened new opportunities for fulfill those purposes, the existing analysis strategies and means of visualization in reporting however are still far from meeting the real-world demands. On the other hand, interpreting lab results via conventional methods are typically guided by theoretical assumptions and model parameters. It therefore inadvertently presents knowledge gaps in a lot of use cases in connection with actual bench/clinical observations. In continuing efforts to address above challenges, here we present Genomics/Genetics - Molecular Biology – System Biology Suite (GMS_Suite) to better integrating state-of-art analysis modules with optimized visualizations on biomedical datasets. GMS_Suite provides end-to-end solutions on characterizations, ranging from -omics to imaging, suits the needs for rapid analysis and diagnosis. Notably, GMS_Suite leverages explainable AI modules in improving model predictions and interpretations thus enhancing experimental designs. GMS_Suite is also highly customizable to tailored problems - as demonstrated in the investigating of the amplification behavior of MYC onco-gene and differentiating features of lymphomas subtypes. We hope that GMS_suite might be valued as a handy toolkit in Basic and Clinical Research.
Prof. Rudolf M. Füchslin
Zurich University of Applied Sciences – School of Engineering
Ruedi Füchslin studied theoretical physics at ETH in Zürich. He got his PhD from the University of Zurich, where he wrote his thesis in the newly established group for computer-assisted physics. By chance, he had the opportunity to act as a consultant for the Institute for Forensic Medicine. This experience fostered his interest in biological and medical problems. After various postdoc positions on the interface between theoretical biology and engineering, he got a position as professor of applied complex systems science at the Zurich University of Applied Sciences. In addition, he is co-director of the European Centre for Living Technology in Venice, Italy and president of the Naturama Foundation. Besides research, he is interested in questions relating to the interplay between natural sciences and the humanities.
Application-oriented modelling in science, technology and decision support: Lessons from Knowledge Transfer across the boundaries of science
The interplay between fundamental and applied science, industrial R&D, and further parts of society requires carefully considering the role of the individual actors. This talk illustrates and discusses the tension between experiments and modelling with an emphasis first on embodied intelligence in chemical multi-compartment systems, which means compounds of different types of artificial and living cells and second on more lessons we learnt in the collaboration between academia, government, and industry.
Edda Klipp
Theoretical Biophysics, Humboldt-Universität zu Berlin, Berlin, Germany
Edda Klipp is full professor for Theoretical Biophysics at Humboldt-Universität zu Berlin. She received a doctoral degree in Theoretical Biophysics from Humboldt-Universität zu Berlin in 1994. She worked as junior group leader at the Max Planck Institute for Molecular Genetics in Berlin, where she focused on modeling cellular stress response and growth control. Klipp’s research interests lie in the area of systems biology, with a focus on mathematical modeling and simulation of complex biological systems. She has made significant contributions to the development of computational models of metabolism, signal transduction, and gene regulation, with a particular emphasis on the analysis of the dynamics and control of these systems. In 2009 she was awarded an honorary doctor of Göteborg University. 2015 she was awarded the Caroline-von-Humboldt professorship at Humboldt-Universität zu Berlin. In addition to her research activities, Klipp is also actively involved in the promotion of science education and outreach, and she has served as a mentor for many young scientists in the field of systems biology.
Understanding a life – integration of time-resolved population and single cell data for a yeast cell from birth to division
With the progress of genome-wide experimental approaches we witness the establishment of more and more libraries of genome-wide data for proteins or RNA or metabolites, especially for specific cell lines or for well-studied model organisms such as S. cerevisiae. However, the separated consideration of metabolic networks or gene regulation networks does not tell us how these networks are integrated to allow a cell to grow, divide and respond to changing environments.
We use the yeast Saccharomyces cerevisiae as the model organism for eukaryotic cells allowing to comprehensively analyzing regulatory networks and their integration with cellular physiology. We focus on processes during the lifetime of a single cell along one period of the cell division cycle and study the changes of metabolism, gene expression, or ion and nutrient transport during the growth of that cell.
We use a modular and iterative approach that allows for a systematic integration of cellular functions into a comprehensive model allowing to connect processes that are strongly interlinked in cellular life, but measured separately. The modular concept also permits to zoom in and out if different aspects of regulation or dynamics come into focus.