On fait la science avec des faits, comme on fait une maison avec des pierres :
mais une accumulation de faits n'est pas plus une science qu'un tas de pierres n'est une maison.

Henri Poincaré, La science et l'hypothèse.

Presentation

I currently work as Research Engineer in the EPI Lifeware team at Inria Saclay. I mainly contribute to the maintenance and development of Biocham (Gitlab), a software that revolve around the famous Chemical Reaction Networks (CRN) of systems biology, promoting them as a Programing Language and allowing high level manipulation (parameter search, graph comparison, function compilation, etc.). As a result of this development, I also do some theoretical reasoning around these functionalities: complexity proof, expressivity, etc.

From a research point of view, my main interest is about from the constraints imposed on the biological systems by the physical or even the mathematical laws. This topic covers for example the study of evolutionary spandrels like the antagonism discovered and proved in the perfect adaptation of the immune system or more recently the comparison of CRN produced for the same task (an oscillating clock like CRN), by artificial evolution, direct compilation and also hopefully darwinian evolution.

On top of my engineering and research job, I am invested in scientific mediation both through in-depth work (like translation of Wikipedia articles) and particular events (Fête de la Science, programme Chiche, semaine des mathématiques, etc.).

Research interests

Biochemical computing
What does computation mean in the context of Biochemical Reactions Networks?
What are the possible models for a biochemical machine and what is the expressivity of these models?
Mathematical and physical constraints of the living systems
Given a particular task (synchronization, discrimination, etc.), how may a living system perform it ?
If there is several possibilities, which ones will be selected by natural evolution, why?
Natural and artificial evolution
What are the typical cases where an evolutionary algorithm would be relevant?
Is it possible to distinguish the products of an evolutionary method from ones produce through other means (algorithmic compilation, human invention, etc.)?
If it is the case, how can we characterize evolutionary products from a statistical point of view?

Software development

  • Biocham (dépot git, documentation) -- A platform for the modelling analyzing of biochemical reactions networks (CRN).
  • Nested evolutionary algorithm -- Proof of concept of a two layers evolutionary algorithm that use CMA-ES to determine the parameters used to evaluate the fitness of an evolving population of PODE.
  • Phi-evo (Dépot git, Documentation) -- An evolutionary algorithm for bio-molecular interactions selected toward a high-level physically defined functionality (oscillation, mutual-information, etc.).
  • Evolutionary algorithm of spin-like protein model -- An evolutionary algorithm imposing a variable fitness function upon a rough model of protein to see how the statistics of the environment is imprinted into the architecture of the protein by the evolutionary process.

Education and previous positions

  • 2018-… : Post Doctoral Fellowship and engineer position at Inria Saclay
  • 2015-17 : Post Doctoral Fellowship at Mcgill University (Montreal QC) on the artificial evolution of biological reactions network.
  • 2012-15 : PhD Thesis of Statistical Physics of Darwinian evolution (Modèles d’évolution de protéines en environnement variable) under the supervision of Florent Krzakala and Olivier Rivoire.
  • 2011-12 : Master in Theoretical and Statistical Physics at the ENS-Ulm.
  • X08 : Ancien élève de l'École polytechnique.

Teaching

  • Constraints & Relational Programing in the École polytechnique, tutorial session for the bachelor and engineer school (3 years, 2019,2021,2022 : 50h).
  • Optic and Wave Propagation several lectures, tutorial and practical works given in the prépa intégéré de Grenoble INP as part of my PhD thesis instructorship (2 years: 2014, 2015, 128h).
  • Physics tutorial in the École polytechnique for quantum and statistical physics during my master (1 year: 2012, 64h).

Publications

Hemery, Mathieu, Fages, François. On Estimating Derivatives of Input Signals in Biochemistry. In CMSB 2023 - 21st International Conference on Computational Methods in Systems Biology, volume 14137 of LNCS. Springer-Verlag, 2023. [ preprint ]

Mathieu Hemery, François Fages. Algebraic Biochemistry: a Framework for Analog Online Computation in Cells. In CMSB'22: Proceedings of the twentieth international conference on Computational Methods in Systems Biology, volume 13447 of LNCS. Springer-Verlag, 2022. [ preprint ] [ slides ] [ video ]

Mathieu Hemery, François Fages, Sylvain Soliman. A Polynomialization Algorithm for Elementary Functions and ODEs, and their Compilation into Chemical Reaction Networks. In CASC'21: Computer Algebra in Scientific Computing, proceedings of short papers, 2021. [ preprint ]

Mathieu Hemery, François Fages, Sylvain Soliman. Compiling Elementary Mathematical Functions into Finite Chemical Reaction Networks via a Polynomialization Algorithm for ODEs. In CMSB'21: Proceedings of the nineteenth international conference on Computational Methods in Systems Biology, volume 12881 of LNCS. Springer-Verlag, 2021. [ preprint ]

Mathieu Hemery, François Fages, Sylvain Soliman. On the Complexity of Quadratization for Polynomial Differential Equations. In CMSB'20: Proceedings of the eighteenth international conference on Computational Methods in Systems Biology, LNCS. Springer-Verlag, 2020. [ preprint ]

Elisabeth Degrand, Mathieu Hemery, François Fages. On Chemical Reaction Network Design by a Nested Evolution Algorithm. In CMSB'19: Proceedings of the seventeenth international conference on Computational Methods in Systems Biology, volume 11773 of LNCS. Springer-Verlag, 2019. [ preprint ]

Hemery, Mathieu, François, Paul. In Silico Evolution of Biochemical Log-Response. The Journal of Physical Chemistry B, 2019. [ preprint ]

Henry, Adrien, Hemery, Mathieu, François, Paul. varphi-evo: A program to evolve phenotypic models of biological networks. PLoS computational biology, 14(6):e1006244, 2018.

François, Paul, Hemery, Mathieu, Johnson, Kyle A, Saunders, Laura N. Phenotypic spandrel: absolute discrimination and ligand antagonism. Physical Biology, 13(6):066011, 2016.

Hemery, Mathieu. Modeles d'evolution de proteeines en environnement variable.. Thèse de doctorat, Universite Pierre et Marie Curie, 2015. [ preprint ]

Hemery, Mathieu, Rivoire, Olivier. Evolution of sparsity and modularity in a model of protein allostery. Physical review E, 91(4):042704, 2015.

Candiature CRCN/ISFP 2023

  • Phenotypic spandrel: absolute discrimination and ligand antagonism : pdf_article
  • Algebraic Biochemistry: A Framework for Analog Online Computation in Cells : pdf_article

HAL_check_list?