Thomas Gaudelet

About me

Hello and welcome!

I am currently a Ph.D Candidate at University College of London supervised by Professor Nataša Pržulj.

My research interests lie within the fields of mathematics and computing. I focus on developing models for mining, analysing and integrating large scale data using graph theory and machine learning approaches. I use multi-scale molecular and clinical data to uncover new biological knowledge with an eye towards improving patient diagnosis, prognosis, and treatment.

I am a graduate from the University of Oxford, MSc. in Applied Mathematics, and from Ecole Centrale de Lyon, BEng/MEng in general engineering.



T. Gaudelet, N. Malod-Dognin, J. Sànchez-Valle, V. Pancaldi, A. Valencia, and N. Pržulj, Unveiling new disease, pathway, and gene associations via multi-scale neural networks, arXiv preprint arXiv:1901.10005


Diseases involve complex processes and modifications to the cellular machinery. The gene expression profile of the affected cells contains characteristic patterns linked to a disease. Hence, biological knowledge pertaining to a disease can be derived from a patient cell's profile, improving our diagnosis ability, as well as our grasp of disease risks. This knowledge can be used for drug re-purposing, or by physicians to evaluate a patient's condition and co-morbidity risk.
Here, we look at differential gene expression obtained from microarray technology for patients diagnosed with various diseases. Based on this data and cellular multi-scale organization, we aim to uncover disease--disease links, as well as disease--gene and disease--pathways associations. We propose neural networks with structures inspired by the multi-scale organization of a cell. We show that these models are able to correctly predict the diagnosis for the majority of the patients. Through the analysis of the trained models, we predict and validate disease--disease, disease--pathway, and disease--gene associations with comparisons to known interactions and literature search, proposing putative explanations for the novel predictions that come from our study.


T. Gaudelet, N. Malod-Dognin, and N. Pržulj, Higher-order molecular organization as a source of biological function, Bioinformatics, Volume 34, Issue 17, 1 September 2018, Pages i944–i953.


Molecular interactions have widely been modelled as networks. The local wiring patterns around molecules in molecular networks are linked with their biological functions. However, networks model only pairwise interactions between molecules and cannot explicitly and directly capture the higher-order molecular organization, such as protein complexes and pathways. Hence, we ask if hypergraphs (hypernetworks), that directly capture entire complexes and pathways along with protein–protein interactions (PPIs), carry additional functional information beyond what can be uncovered from networks of pairwise molecular interactions. The mathematical formalism of a hypergraph has long been known, but not often used in studying molecular networks due to the lack of sophisticated algorithms for mining the underlying biological information hidden in the wiring patterns of molecular systems modelled as hypernetworks.
We propose a new, multi-scale, protein interaction hypernetwork model that utilizes hypergraphs to capture different scales of protein organization, including PPIs, protein complexes and pathways. In analogy to graphlets, we introduce hypergraphlets, small, connected, non-isomorphic, induced sub-hypergraphs of a hypergraph, to quantify the local wiring patterns of these multi-scale molecular hypergraphs and to mine them for new biological information. We apply them to model the multi-scale protein networks of bakers yeast and human and show that the higher-order molecular organization captured by these hypergraphs is strongly related to the underlying biology. Importantly, we demonstrate that our new models and data mining tools reveal different, but complementary biological information compared with classical PPI networks. We apply our hypergraphlets to successfully predict biological functions of uncharacterized proteins.


H. Tsuchiya, K. Manabe, T. Gaudelet, T. Moriya, K. Suwabe, M. Tenjimbayashi, K. Kyong, F. Gillot, and S. Shiratori. Improvement of heat transfer by promoting dropwise condensation using electrospun polytetrafluoroethylene thin films, New Journal of Chemistry 41, no. 3 (2017): 982–991.


Vapor condensation is a crucial part of a broad range of industrial applications including power generation, water harvesting, and air conditioning. Hydrophobic and superhydrophobic surfaces promote dropwise condensation in vapor-filled environments and increase their heat transfer coefficients more than filmwise condensation on hydrophilic surfaces. Although dropwise condensation can lead to energy-efficient transfer, it is hard to achieve stable dropwise condensation in high-temperature environments. To decide the best conditions for achieving higher heat transfer is also difficult because the heat transfer coefficient is influenced by not only surface wettability but also surface structures of thin films and substrates. Herein, we fabricated thin films with different wettabilities and surface structures using polytetrafluoroethylene (PTFE) which show high heat resistance to determine the best conditions for heat transfer. Several different films were prepared by electrospinning a mixed solution of PTFE and polyvinyl alcohol on aluminum (Al) and copper (Cu) tubes. After annealing them, the PTFE thin films enhanced heat transfer performance and showed stable dropwise condensation in high-temperature environments. The films fabricated by electrospinning a solution containing 66 wt% PTFE displayed the highest heat transfer coefficients, with heat transfer coefficients 64% and 61% greater than those of uncoated Al and Cu tubes, respectively. That is because homogeneous superhydrophobic surfaces that showed the highest departure frequency of condensed water droplets were fabricated using 66 wt% PTFE. The results suggest that these electrospun PTFE thin films would demonstrate excellent potential for use on the surface of heat exchangers in various industries.



  • 2016-

    Ph.D Candidate
    University College of London

    Department of Computer Science.
    Supervised by Nataša Pržulj
    Title to be set.
  • 2014-

    Master of Science
    University of Oxford

    Mathematical Modelling and Scientific Computing
    Dissertation: investigated and implemented pattern recognition using multi-layer neural networks, highlighting their robustness.
  • 2011-

    MEng General Engineering
    Ecole Centrale Lyon

    Developed a software designed to read multimedia files during the first year project. During the second year, studied a solution to optimise the shape of vehicle components based on the level-set method.
  • 2008-

    Intensive Preparatory Courses
    Lycée Pierre de Fermat, Toulouse

    Grandes Ecoles entry exam preparatory courses: Mathematics, Physics, Computer Science.

Work experience

  • 2016

    R&D Intern
    Fuel3D Technologies

    Investigated and implemented algorithms, using Matlab and C++, to improve on the quality of the reconstruction of three dimensional surfaces using Vision and Photometric Stereo.
  • 2014

    Development Intern

    Developed the user interface of the software of passive magnetic ranging created by PathControl. It involved mainly Matlab and the associated GUI Toolbox.
  • 2013-

    Research Intern
    Keio University, Tokyo

    Investigated the condensation behaviour of water on PTFE-based coating. The objective was to enhance heat exchanges through the surface. We presented the results at the JSAP conference in Tokyo.
  • 2013


    Examined the interest of nanofilms/nanoparticles in energy generation and safety.

Past projects

Fuel3D Internship

Three-source Photometic Stereo.
(Report confidential)


We investigate different leads to improve photometric stereo for the reconstruction of three-dimensional shape based on a set of input pictures captured with Fuel3D Technologies Limited scanner SCANIFY®.

MSc Dissertation

Short-range Impact of Damage on Object Recognition in a trained Neuronal Network.


We investigate the impact of damage on a neuronal network trained to recognise objects. A neuronal network is formed by a set of nodes that represent neurons, and edges that represent the connections between the neurons. To set the stage, we review some existing models that describe the dynamics of individual neurons. We then focus on Integrate-and-Fire (IF) models that we use for this project. We then discuss the numerical method that we use for our model of interacting IF neurons. We then construct our network using a multilayer-network formalism. We consider a simple multilayer network architecture that represents the interactions between the retina and the primal visual cortex (V1) in the brain. We train the system to recognise objects and differentiate between them using the "continuous transformation learning rule", which is based on relative spike times of pre-synaptic and post-synaptic neurons. To highlight the robustness and stability of the trained neuronal network, we simulate damage affecting the connections in the network and measure the performance of the deteriorated systems. A network has good performances if the neurons in the V1 representation differentiate successfully between different objects. In this situation, a subset of the neurons responds strongly to the stimuli corresponding to one object and weakly to the other. Our study provides preliminary insights on the impact of damage on connections between two neuronal subsystems.

Pathcontrol Internship

GUI Development for Passive Magnetic Ranging.
(Report confidential)


The Passive Magnetic Ranging, or PMR, is a detection method based on the magnetic field produced by an object one tries to locate. The idea is to drive a magnetic object into the ground and to monitor its magnetic field. Then using the magnetic interferences produced by the target one can calculate its position relatively to the measurement devices. In our case, this allows the user to control the borehole's trajectory in order to achieve, for instance, the interception of the second well. The work entails developping the graphical user interface for PathControl's PMR software. It also involve detailing the method used as well as the software layout and how to use it.

Internship Research

Investigation of Condensation Mode on PTFE-based Coating.


The enhancement of heat transfers during condensation is at the core of numerous researches as it will lead to increased efficiencies and productivities for various fields especially for power generation. One of the main solutions explored at the current time is to engineer a surface coating in order to promote dropwise condensation, in opposition of the filmwise condensation generally observed in actual condensing systems. With this goal in mind, we investigate the use of polytetrafluoroethylene-coated (PTFE) surfaces to improve heat transfers.

EDF Internship

Nanotechnologies in Energy Production.
(Report confidential)


We present the interest of superhydrophobic and superhydrophilic surfaces for the enhancement of heat transfers. For this purpose, we review the literature on the subject and investigate the impact of these surfaces for two applications: improvement of energy production and improvement of cooling methods. Furthermore, we evaluate different hydrophobic products proposed on the market and try to make out which ones would suit EDF best. Finally, we conduct and expose the results of a study on the effect of a magnetite deposit on a surface's wettability.

Research Project

Shape Optimisation of Vehicle Components using the Level-Set Method.
(Report in french)


We investigate shape and topologic optimisation based on the Level-Set method developed by the mathematicians J.A. Sethian adn S. Osher. We start by detailing the theory and the optimisation algorithm. We then illustrate how the method works on a simple 2D case. The algorithm is implemented combining Matlab and Code Aster.

Group project

Creation of a Smart Multimedia Files Manager.
(Report in french)


This study project has for goal the development of a multimedia software, that we baptised Tag'n'Link. The project was proposed by the laboratory LIRIS of Ecole Centrale Lyon. LIRIS stands for Computer Science Laboratory for Image Processing and Information Systems (Laboratoire d'Informatique en Image et Systèmes d'Information). LIRIS was developing algorithms to detect concepts and emotions in pictures or music files. An emotion being defined by a degree of positivity (joy/sadness) and a level of activity. A concept is for instance a word or an object, for instance a car in a picture. The idea of the projects was to use these algorithms as a black box and to develop a software around them in order to exploit them. The main noveltie was to use both the algorithm and user tags associated to the media files in order to easily generate playlists, slideshows, and combinations of the two according to a certain degree of similarity.

Get in touch