Workshop Metabolism and mathematical models: Two for a tango – 4th Edition

General presentation

Title: Workshop Metabolism and mathematical models: Two for a tango – 4th Edition

Dates: November 19-20, 2024

Location: This workshop will be held in a virtual way.

The topic of this workshop is metabolism in general, with a special focus, although not exclusive, on parasitology. Besides an exploration of the biological, biochemical and biomedical aspects, the workshop will also aim at presenting some of the mathematical modelling, algorithmic theory and software development that have become crucial to explore such aspects.

This workshop is being organised in the context of two projects, both with the Inria European Team Erable. One of the projects involves a partnership with the University of São Paulo (USP), in São Paulo, Brazil, more specifically the Institute of Mathematics and Statistics (IME) and the Institute of Biomedical Sciences – Inria Associated Team Capoeira – and the other involves the Inesc-ID/IST in Portugal, ETH in Zürich and EMBL in Heidelberg – H2020 Twinning Project Olissipo.

The workshop is open to all members of these two projects but also, importantly, to the community in general.

Keynote lecturers

Ross Carlson, Montana State University, USA

Title talk: Cell geometry and membrane protein crowding are major constraints on bacterial phenotype including growth rate, respiration efficiency, and maintenance energy generation.

Short abstract: Bacterial cell geometry is highly regulated and constrains the surface area available for acquiring nutrients as well as the volume available for synthesizing proteins. The surface area to volume (SA:V) ratio of all documented bacteria decreases with growth rate due to increases in cell size. However, the membrane protein content of Escherichia coli increases with growth rate creating a positive correlation between membrane enzyme capacity and growth rate. Despite its central role in cell biology, the intersection of membrane protein capacity, cell geometry, and central metabolism has not been defined with a predictive and quantitative theory. Here, we present a biophysical basis for maximum growth rate, overflow metabolism, electron transport chain efficiency, and maintenance energy flux. The theory successfully predicts the phenotypes of two genetically similar E. coli K-12 strains, MG1655 and NCM3722, that have SA:V ratios that vary up to 30%, maximum growth rates on salts medium that differ by 40%, and growth rates at which overflow phenotypes occur differing by 80%. These analyses do not consider the cytosolic proteome, demonstrating the predictive power of the surface area and membrane protein crowding alone. Cell geometry and membrane protein crowding are significant biophysical constraints on phenotype and provide a theoretical framework for improved understanding and control of cell biology.

Jurgen Haanstra, VU Amsterdam, The Netherlands

Title talk: Simulating (currently) impossible experiments – a tale of two parasites.

Short abstract: Energy metabolism is pivotal to all life. A highly conserved pathway in metabolism is glycolysis: the breakdown of glucose to generate ATP and precursors for other metabolic pathways. While the pathway is very similar in many organisms, there are sometimes intriguing differences in the exact set-up of the pathway in different organisms. How are those differences important for the functioning of the pathway?
In my talk I will focus on glycolysis in Trypanosoma brucei and in Schistosoma mansoni, two different parasites. T. brucei has compartmentalised parts of glycolysis inside an organelle and S. mansoni has unusual regulation of lactate dehydrogenase, the enzyme that converts the pyruvate that comes out of glycolysis to lactate. The role of each of these phenomena is challenging to probe experimentally in the wetlab and we have successfully employed computational modelling to predict their roles in glucose metabolism..

John Kececioglu, The University of Arizona, USA

Title talk: Inferring pathways in metabolic networks via optimal factories and hyperpaths.

Short abstract: Fundamental to systems and synthetic biology is the task of inferring pathways of reactions in metabolic networks. Two key models for pathway inference are factories and hyperpaths: systems of reactions that ultimately produce target substances, starting from the available source compounds. Factories take into account the stoichiometries of reactions to produce the targets while conserving or not depleting intermediate metabolites; hyperpaths ignore stoichiometry, and yield an ordered cascade of reactions in which the input reactants to each reaction are produced as output products of prior reactions; finding optimal factories and hyperpaths are both NP-complete. We present recent results on the first practical exact algorithms for optimal factories and hyperpaths, which are fast in practice (as demonstrated through comprehensive experiments over the standard pathway databases in the literature), and freely available in the tools Freeia and Mmunin.
This is joint work with Spencer Krieger, published in RECOMB 2023, RECOMB 2024, and their special issues in the Journal of Computational Biology.

Diego Oyarzún, University of Edinburgh, Scotland

Title talk: Integration of machine learning and genome-scale metabolic models for improved phenotypic prediction.

Short abstract: Genome-scale metabolic models are powerful tools for understanding cellular physiology. Flux balance analysis (FBA), in particular, is an optimization-based approach widely employed for predicting metabolic phenotypes. In this talk I will discuss our recent work on machine learning in tandem with FBA for improved predictive power. I will discuss the use of surrogate machine learning models to integrate kinetic pathway models with the whole metabolism of a microbial host; this approach allows to predict local pathway dynamics in response to perturbations at the genome-scale, such as gene deletions or drug treatments. I will then discuss a number of recent machine learning algorithms for predicting essentiality of metabolic genes that do not assume fitness optimality of deletion strains, as it is commonly required by current techniques. These results highlight the many opportunities offered by the integration of data-driven approaches with well established biological modelling strategies.
References
[1] Merzbacher, Mac Aodha, and Oyarzún, ‘Modelling dynamic host-pathway interactions at the genome scale’. bioRxiv, 2024. doi: 10.1101/2024.04.09.588720.
[2] Freischem, Barahona, and Oyarzún, ‘Prediction of Gene Essentiality Using Machine Learning and Genome-Scale Metabolic Models’. Foundations of Systems Biology and Engineering, 2022. doi: 10.1016/j.ifacol.2023.01.006
[3] Hasibi, Michoel, and Oyarzún, ‘Integration of Graph Neural Networks and Genome-Scale Metabolic Models for Predicting Gene Essentiality’. npj Systems Biology & Applications, 2024, 10.1038/s41540-024-00348-2

Samraat Pawar, Imperial College London, UK

Title talk: Thermal-metabolic constraints on complex ecosystem dynamics (AKA how can we *really* engineer microbiomes?).

Short abstract: Metabolically-driven interactions between biological species in any local ecosystem typically form large, complex networks that generate interesting and often unpredictable system dynamics. A new Mount Everest for theoretical biology, with a wide range of applications, is the development of methods to control and engineer these complex, nonlinear systems. To this end, we need to understand a fundamental problem that has occupied Biologists, and in particular, Ecologists, for almost two centuries: how do these seemingly improbable systems assemble and persist in (an increasingly) fluctuating and uncertain real world? In this talk, I will outline the challenge, and present recent progress towards the goal of engineering microbiomes under a ubiquitous source of environmental fluctuations: temperature.

Jürgen Zanghellini, University of Vienna, Austria

Title talk: Boosting biotechnological production process with constraint-based process modeling.

Short abstract: Biotechnological processes for producing high-value compounds are often inefficient and costly. Addressing these challenges requires innovative solutions, and the reconstruction, analysis, and optimization of digital twins has emerged as a powerful paradigm. These models provide deeper insights into intracellular metabolism and fermentation processes, enabling significant improvements in efficiency.
In this talk, we will introduce how advanced computational tools such as dynamic control flux balance analysis (dcFBA) and OptFed are driving notable enhancements in production titer, productivity, and yield. By optimizing bioprocesses, these tools make production more economically viable across various applications, often outperforming traditional methods.
For example, we applied dcFBA to optimize 2,3-butanediol production, developing a two-stage fed-batch process that improved productivity by 104%. Simulations suggested that a continuous reactor setup could further triple productivity. In another study, we optimized a fed-batch process for plasmid DNA production, using a three-stage process with sulfate starvation, which increased the specific supercoiled plasmid DNA yield by 33%. OptFed, when applied to recombinant protein production, led to a 19% enhancement in specific yield.
These examples demonstrate the transformative potential of computational modeling in biotechnology. By harnessing such tools, industries can achieve significant cost savings, enhance sustainability, and unlock new possibilities for large-scale production of biopharmaceuticals and other high-value compounds.

Discussions on open questions

For this fourth edition of the workshop, besides the keynote talks, there will be also two slots, one per day, for a discussion on two specific open questions:

• Discussion 1: Which are the most relevant limitations of metabolic computational models and which are the bottlenecks to improve their development?
• Discussion 2: How robust is the current predictive power of computational tools? Will the field mature enough to make most of wet science obsolete?

Program

In order to cover a larger audience, the workshop will take place in the afternoon, CET time (French time), on both days.

Registration

Registration is free but mandatory. To register, go to this link.

Organisation

Sabine Peres, University of Lyon 1, CNRS, and Inria
Marie-France Sagot, Inria, CNRS and University of Lyon 1 UMR 5558, France
Ariel M. Silber, Department of Parasitology, Institute of Biomedical Sciences, University of São Paulo (USP), São Paulo, Brazil
Susana Vinga, Instituto Superior Técnico and INESC-ID, Lisbon, Portugal, and coordinator of the H2020 Twinning project Olissipo

Participants

Information on the participants will be made available later.

Funding

The Inria Associated Team Capoeira has received funding from Fapesp in Brazil and from Inria in France.

The H2020 Twinning Project Olissipo has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 951970.

One of the organisers of the workshop, Ariel M. Silber, is also supported by the project “A Global Network fo Neglected Tropical Diseases” of the GCRF-UKRI research programme.