HPC-Multiscale Modelling

This pole aims at sharing computational techniques, skills and tools in order to develop new materials and predict their final properties.

It also aims at improving the modelling techniques and computer codes to account for most of the chemistry and physics of structured matter.

The pole uses the university resources such as the Technological Platform of Intensive Calculation, where high performance simulations and modelling are realized. High performance computing combines the computing power of a large number of processors, memory and storage devices in HPC clusters to solve problems that are too large for a single computer.

Yoann OLIVIER yoann.olivier@unamur.be +32 (0) 81 724534

Non-permanent members representative:
Charlotte BOUQUIAUX charlotte.bouquiaux@student.unamur.be

Pole Members

Juan Cabrera Jamoulle juan.cabrera@unamur.be                   Academic      
Benoît Champagne benoit.champagne@unamur.be               Academic      
Luc Henrard luc.henrard@unamur.be                    Academic      
Laurence Leherte laurence.leherte@unamur.be               Academic      
Vincent Liégeois vincent.liegeois@unamur.be               Academic      
Alexandre Mayer alexandre.mayer@unamur.be                Academic      
Yoann Olivier yoann.olivier@unamur.be                  Academic      
Eric Perpète eric.perpete@unamur.be                   Academic      
Frédéric Wautelet frederic.wautelet@unamur.be              ATG           
Freddy Zutterman freddy.zutterman@unamur.be               Collaborator  
Joseph Fripiat joseph.fripiat@unamur.be                 Emeritus      
Daniel Vercauteren daniel.vercauteren@unamur.be             Emeritus      
Pierre Beaujean pierre.beaujean@unamur.be                PhD           
Charlotte Bouquiaux charlotte.bouquiaux@unamur.be            PhD           
Pauline Castenetto pauline.castenetto@unamur.be             PhD           
Tanguy Colleu-Banse tanguy.colleubanse@unamur.be             PhD           
Alban De Gary alban.degary@unamur.be                   PhD           
Antoine Fauroux antoine.fauroux@unamur.be                PhD           
Mathias Fraiponts mathias.fraiponts@unamur.be              PhD           
Emerick Guillaume emerick.guillaume@unamur.be              PhD           
Antoine Honet antoine.honet@unamur.be                  PhD           
François Mairesse francois.mairesse@unamur.be              PhD           
Bruno Majerus bruno.majerus@unamur.be                  PhD           
Christ Daniel Matatu Mbengo christ.matatumbengo@unamur.be            PhD           
Manon Mirgaux manon.mirgaux@unamur.be                  PhD           
Virgile Neyman virgile.neyman@unamur.be                 PhD           
Gaetano Ricci gaetano.ricci@unamur.be                  PhD           
Nicolas Roy nicolas.roy@unamur.be                    PhD           
Mégane Van Gysel megane.vangysel@unamur.be                PhD           
Tom Cardeynaels tom.cardeynaels@unamur.be                Post-doc      
Laura Le Bras laura.lebras@unamur.be                   Post-doc      
Tarcius Nascimento Ramos tarcius.nascimento@unamur.be             Post-doc      

Highlighted publications

  • "Second harmonic generation responses of ion pairs forming dimeric aggregates", Nascimento Ramos, T., Castet, F. & Champagne, B., 26 mars 2021, Journal of physical chemistry B. 125, 13, p. 3386-3397
  • "Negative Singlet-Triplet Excitation Energy Gap in Triangle-Shaped Molecular Emitters for Efficient Triplet Harvesting", Sanz-Rodrigo, J., Ricci, G., Olivier, Y. & Sancho-García, J. C., 21 janv. 2021, Journal of physical chemistry A. 125, 2, p. 513-522 10 p.
  • "Multiresolution non-covalent interaction analysis for ligand-protein promolecular electron density distributions", Leherte, L., 2021, Theoretical Chemistry Accounts. 140, 1, 13 p., 9.

Higlighted videos

- Multiscale theoretical investigation of the second harmonic generation of the di-ANEPPS dye in a collection of lipid membranes


Most biomolecules possess few natural moieties with exploitable optical properties. The use of exogenous dyes can improve the contrast in tissues for being detected by commercially available microscopes. Usually, ANEP-like fluorescent dyes are used, but in the last few years, these compounds have also gained interest in the field of Second Harmonic Generation (SHG). This technique possesses the advantage, with respect to fluorescence, that only non-centrosymmetric region produces a SHG signal.

On this basis, the objective of this work is the elaboration of a multiscale theoretical chemistry method to predict and interpret the SHG responses of dyed lipid membranes and applying it to a hierarchy of lipid membranes of increasing complexity.

Previous work[1] highlighted the huge surrounding effects when considering the influence of the environment on the SHG response of the di-8-ANEPPS (Figure 1) dye [from the isolated gas phase to the 16:0-16:0 PC (phosphatidylcholine) bilayer]. This called for further investigation of the role of the nature of the lipid bilayer on the di-8-ANEPPS SHG properties. In this work, the di-8-ANEPPS NLOphore is embedded in a collection of model lipid membranes. They are composed of: (i) lipids of the same category as 16:0-16:0 PC (zwitterionic glycerophospholipid), 16:0-18:1 PC, and 16:1-16:1 phosphoethanolamine (PE) to study the effect of the unsaturation(s) of the alkyl chains and of the nature of the polar head; (ii) another type of lipid, sphingomyelin (SM); and 18:1-16:0 SM, (iii) various concentrations of cholesterol in 16:0-16:0 PC membranes to obtain the first mixed systems of this study. The systems are currently being modeled via MD simulations, before being validated by comparing structural parameters with experimental data. This poster presents preliminary results of the NLO response of the dye from extracted snapshots and compares the results from one environment to the next.

- Near-field and electrodynamic modes in plasmonic systems

Tanguy Colleu-Banse (1(, Vincent Liégeois (2), Luc Henrard (1),

Laboratoire de Physique du Solide (LPS), Department of Physics (1),
Laboratory of theoretical Chemistry (LTC), Department of Chemistry (2),
University of Namur, Namur
E-mail: tanguy.colleubanse@unamur.be

The SURFASCOPE project aims to simulate Surface Enhanced Raman Spectroscopy (SERS).  The very high cross section of the Raman process in SERS relies on the intensity and spatial variation of the near-field response at the position of the probed molecules. In this context, we have simulated plasmonic systems with Discrete-Dipole Approximation. This flash presentation will show the electromagnetic response of several plasmonic devices with a focus on the differences between the far-field response and the near-field response.
The studied particles are a disk, two disks close to each other and a donut-like shape (hollow disk). With those examples, we exemplify the difference between the far-field (scattering cross-section), the near-field inside the particles (absorption cross-section) and the near-field outside the particles (related to the SERS enhancement). The figure below shows the electric field around two disks of gold with a diameter of 10 nm at a plasmonic resonance.  

- Electronic and optical properties of molecules and nanoparticles : the role of electronic correlation

Antoine HONET (1), Luc HENRARD (1), Vincent MEUNIER (2)
(1) Department of Physics and NISM, University of Namur, Namur, Belgium
(2) Department of Physics, Applied Physics and Astronomy, Rensselaer Polytechnic

The optical response of molecules and nanoparticles is at the origin of many static and dynamical properties of matter. As an example, electrochromism is directly related to the modification of the optical absorption with the charge (number of electrons) of the systems.
The modelization and the simulation of such an effect relies of an accurate description of the electronic correlation. In brief, this electron correlation can be defined as the missing part in an independent electron (or mean field) description and is particularly important in collective excitations, such as plasmons.
Tight-binding (TB) is an often-used method to compute electronic properties of molecules and solids systems in the independent electrons picture. It is less frequently used to compute optical absorption of an incident electromagnetic radiation because of the time needed to calculate the atomic polarizability [1, 2].
A way to overcome this limitation is to work in Fourier space [3] which can be also formulated in a Green’s function formalism [4]. We then developed the Green’s function method to access the polarizability more easily. This formalism applies for TB as well as for a mean field treatment of a Hubbard Hamiltonian, a model that takes into account the electron correlation.
We have developed a self-consistent algorithm that takes into account the electron-electron correlation that is not included in tight-binding or in mean field Hubbard by a so-called GW correction. The reason to use GW correction is to include correlation effects at a small computation cost, in comparison with exact diagonalization.
These developed techniques are now being used to describe PAH and nano-graphene, including electron-doped or nitrogen-doped nano-graphene in order to explore the effect of the dopants and of the correlation between electrons.

- Paving the way for the computational design of supramolecular drugs

Laura LE BRAS (a), Yves DORY (b), Benoît CHAMPAGNE (a)

(a) Laboratoire de Chimie Théorique, Namur Institute of Structured Matter (NISM), Université de Namur – Namur, Belgium
(b) Laboratoire de Synthèse Supramoléculaire, Institut de Pharmacologie, Université de Sherbrooke – Sherbrooke, Canada
E-mail: laura.lebras@unamur.be

It has been established that the Parkinson’s disease (PD) finds its origin in the aggregative behavior of the α-synuclein (α-SYN) protein[1-3]. The aggregates formed are characterized by a stacking of the proteins and a dimerization of those stackings (Figure 1)[4]. Considering and understanding the aggregation process of α-SYN thus appears to be a pre-requisite for the development of a drug. One of the strategies to treat PD could be to avoid the formation of the aggregates. It can be achieved through the design of small organic molecules, that will be able to (1) self-assemble as organic nanotubes (ON) via hydrogen bond network and (2) interact with the large supramolecular assembly that is α-SYN. In this sense, the in silico design and thus computational chemistry, is a valuable tool. Thanks to a multi-scale approach, combining classical molecular dynamics simulations and quantum chemistry calculations, it will be possible to fully characterize the ON. In particular we will be able to provide answers to some fundamental questions such as: Are the identified molecules forming ONs? Are ONs stable enough? Are they interacting with α-SYN? A schematic representation of the concept of the project is provided in Figure 1.

In this presentation we will develop the computational approach and the strategy that has been designed to take into account the large assemblies that are involved (α-SYN and ON). Then, different families of organic molecules will be considered and their properties as ON, as long as their interactions with α-SYN, will be investigated.

- A structural study by crystallography and Molecular Dynamics of hIDO1 to assist design of original inhibitors

Manon Mirgaux, Laurence Leherte, Johan Wouters

Laboratoire de Chimie Biologique Structurale (CBS), Namur Institute of Structural Matter (NISM), Namur Research Institute for Life Sciences (NARILIS), University of Namur (UNamur), Belgium, Email: manon.mirgaux@unamur.be.

Due to its degradation into numerous bioactive metabolites, L-Tryptophan (L-Trp) is the less abundant but essential amino acid in mammals. [1] The main way to transform L-Trp is the kynurenine pathway (KP, Figure 1), allowing the transformation of 95\% of the L-Tryptophan available from food. [2] Through KP, L-Trp is converted into L-kynurenine to produce essential metabolites such as redox cofactors, neuroprotectors and neurotoxics. [3] As a result, the KP pathway confers to L-Trp a central role in many diseases. [2, 4] Therefore, enzymes of the kynurenine pathway can be considered as a set of therapeutic targets. Particularly, the first step of this road, catalyzed by hIDO1 or hTDO, raised an interest in the cancer research in cancer immune escape and in the resistance to immunotherapy. [5] Over the years, hIDO1 inhibitors have been developed (Indoximod, Epacadostat, PF-06844003, Navoximod and Linrodostat) but, recently, the most advanced of them (Epacadostat) failed in the clinical trial. [6]

Since 2006, several structures of hIDO1 are deposited on the PDB. However, a loop of the enzyme, the JK-loop, has never been resolved. As the JK-loop is very likely involved in the closure of the active site, this loop plays a key role in the mechanism and the inhibition. [7] Recent fails in clinical trials of hIDO1 inhibitors trigger a revision of the enzyme functioning involving the JK-loop. [6] It is recognized that there is a lack of structural information on hIDO1 to understand the key role of the JK-loop.

In the present work, the refinement of the JK-loop is obtained for the first time by X-ray diffraction experiment, thanks to its crystal packing mode. To support the X-ray observation, Molecular Dynamics trajectories are also carried out to provide a dynamical information about the loop in the presence of the cofactor. Such new structural and dynamical information highlights the importance of the JK-loop in confining the labile heme cofactor into the active site

- Aggregation effects on nonlinear optical properties of stilbazolium ion pairs

Tárcius N. Ramos, Benoît Champagne

Theoretical Chemistry Lab, Namur Institute of Structured Matter, University of Namur, rue de Bruxelles, 61, B-5000 Namur, Belgium
E-mail: tarcius.nascimento@unamur.be

The electric field induced second harmonic generation (EFISHG) technique has been widely explored to understand the nonlinear optical (NLO) responses in liquid environments. One requirement for such measurements is the charge neutrality of the targeted compounds, leading therefore to additional challenges on neutral ionic pair complexes. Besides, a wide range of NLO responses values is expected when these ionic compounds aggregate or dissociate. The EFISHG response on ionic pair was firstly measured for amphiphilic polyenic push-pull[1] and recently for pH-triggered NLO switches.[2] Moreover, large concentration dependence on the NLO responses was observed for stilbazolium ion pair derivatives.[3] Theoretical investigations can be used to rationalize these results. However, this requires a multiscale computational approach, where the thermodynamics is included by using molecular dynamics with classical force fields while the electronic structure is accounted for by quantum chemistry methods on selected snapshots. Due to the large computational needs and the complexity of the approach, few theoretical studies have been done to understand this phenomenon.[4,5]
Following the investigation of stilbazolium ion pairs,[3] we have theoretically shown that the relative position between the cation and anion plays a crucial role in tuning the EFISHG response on monomer complexes.[5] Now, we are expanding such studies on dimer aggregates to better understand their large concentration dependence. Our results indicate several aggregation structures since stacked or head-to-head structures change the symmetry of the complexes. In comparison to the values obtained on monomer complexes, the second-order EFISHG contribution (μβ_(\/\/)) is reduced by ~50% in dimers due to reductions of μ and β_(\/\/) around 70%. On the other hand, the third-order EFISHG contribution (γ_(\/\/)) is not strongly impacted by the aggregation. One of the studied complexes showed predominantly aggregates in stacked structures. For this case, the value of the μβ_(\/\/) response is ~90% and of the γ_(\/\/) is ~70% of the values obtained for monomers. Correlation analyzes indicate that the intermolecular interactions, represented by the relative distances among the ionic species, play a central role in changing the β tensor elements values contrary to the bond length alternation and intramolecular torsion angles.