Výsledky 25. kola Veřejné grantové soutěže
Všem žadatelům o výpočetní čas v rámci 25. Veřejné grantové soutěže děkujeme za podání žádostí.
Uchazeči o výpočetní zdroje požádali v rámci této výzvy, do následujícího alokačního období o 1 374 890 uzlohodin. S ohledem na vysokou poptávku po výpočetních zdrojích Barbora GPU a Karolina GPU a k zajištění účelnosti využití alokovaných zdrojů, přistoupila alokační komise k mírnému snížení alokací některých z hodnocených projektů.
Všechny projekty prošly technickým hodnocením a hodnocením poměru počtu registrovaných publikací na projekt.
O alokacích komise rozhodla na základě výsledků hodnocení vědecké excelence, výpočetní připravenosti a společenské užitečnosti. Váhou 50 % přihlédla také k poměru počtu registrovaných publikací na projekt. Poměr počtu registrovaných publikací na projekt je definován jako počet publikací registrovaných v posledních 3 letech dělený počtem projektů ukončených minimálně 1 rok a nejvíce 4 roky. V průměru komise očekává 1.0 publikace na projekt. Projekty v rozsahu do 10 % požadovaných zdrojů nebyly předmětem peer-review nezávislými hodnotiteli. V těchto případech komise předpokládala nejvyšší možné bodové hodnocení vědecké excelence, výpočetní připravenosti a společenské užitečnosti. U projektů nových uživatelů komise předpokládala nejvyšší možné hodnocení poměru počtu registrovaných publikací na projekt.
Nízký poměr počtu registrovaných publikací na projekt byl hlavní příčinou krácení alokací.
Komise konstatovala obecně velmi dobrou vědeckou a technickou úroveň hodnocených projektů, kde z maximálního dosažitelného počtu 30 bodů bylo průměrně dosaženo 24 bodů, minimálně 18 bodů. Jedenáct z hodnocených projektů dosáhlo 27 bodů a více. Komise zároveň konstatovala prozatím nízký zájem o výpočetní zdroje LUMI-G v poměru k jejich kapacitě.
Komise zvlášť projednala žádosti č.1429 ,1439, 1444, 1445, 1448, 1449, 1453, 1456, 1457, 1458, 1459, 1460, 1466, 1470, 1471, 1478, 1479, 1481 a 1482. Tyto projekty byly subjektem krácení alokovaných výpočetních zdrojů z důvodu nízkého počtu publikačních výstupů navazujících na předchozí projekty.
Mezi 51 projektů, z toho 6 multi-year, bylo rozděleno celkem 1 329 142 uzlohodin. Projekty, které nebyly podrobeny peer-review, zastupují celkem 5,6 % z celkového počtu alokovaných uzlohodin.
ALOKAČNÍ KOMISE VE 25. KOLE VEŘEJNÉ GRANTOVÉ SOUTĚŽE IT4INNOVATIONS ROZDĚLILA VÝPOČETNÍ ZDROJE TAKTO:
Řešitel: Pavel Plevka
Phage infection of bacterial biofilm 2
Karolina CPU 19531, Karolina FAT 1300, Karolina GPU 20000, Karolina VIZ 40
In 2017, the World Health Organization declared Staphylococcus aureus to be an antibiotic-resistant pathogen for which new therapeutics are urgently needed. Upon infection, S. aureus forms biofilms that can only be treated by the long-term application of several antibiotics in high doses or the surgical removal of the infected tissues. An alternative approach, phage therapy, has not been approved for clinical use, because the effects of phage infection on a biofilm are not sufficiently characterized. We will study the dynamics of the propagation of phages in a S. aureus biofilm and molecular details of phi812 replication in a cell. We will determine how sub-populations of metabolically dormant or phage-resistant cells in a biofilm provide herd immunity against phi812 infection. We will use focused ion beam milling together with cryo-electron microscopy and tomography to determine high-resolution structures of previously uncharacterized phi812 replication and assembly intermediates in S. aureus cells. We will study the function of bacterial membranes and macromolecular complexes in the initiation and completion of phage genome delivery, the assembly of phage portal complexes and heads, and the mechanisms of genome packaging and head-tail attachment. This proposal’s biological significance lies in its focus on the as-yet uncharacterized interactions of phages and bacteria under biologically and clinically relevant conditions.
Řešitel: Martin Friak
Theoretical analysis of materials mysteries related to medieval organ pipes
Karolina CPU 43200, Karolina VIZ 40
Tin was historically often used and is still employed nowadays, e.g., in soldering. Interestingly, some aspects of the beta-Sn to alpha-Sn phase transformation (known as tin pest), that turns bulk Sn into a powder, are still not fully understood. Tin has still some mysteries even today. Published transformation-related data are contradictory regarding both the mechanism of transformation and the influence of solutes. This problem is studied for instance in historical organ pipes. They were made of Sn-Pb alloy with various ratios of components from almost pure Sn to almost pure Pb. Surprisingly massive degradation due to undesirable phase transition was observed in organ pipes with content of Pb low enough to be attacked by atmospheric corrosion and high enough to prevent the organ pipe from the tin pest. In this project, modern atomic-scale theoretical modelling, in particular quantum-mechanical calculations, may shed a new light on atomistic details of this mysterious phase transition.
Řešitel: Raman Samusevich
Enzymes function prediction using machine learning
Karolina CPU 700, Karolina GPU 4200, Karolina VIZ 40
Terpene synthases (TPSs) are enzymes (catalytic proteins) responsible for the biosynthesis of the largest class of natural products, including widely used flavors, fragrances, and first-line medicines. Although the amount of available TPS protein sequences is increasing exponentially, characterizing the function of each TPS requires challenging and time-consuming experiments as well as significant domain expertise. The objective of this project is to develop predictive models for characterizing the function of TPSs directly from their protein sequences. Such a model will have a multitude of applications in drug discovery and synthetic biology and will provide an important precedent towards computational characterization of the catalytic function of enzymes directly from their protein sequences.
Řešitel: Oldřich Plchot
Multichannel Speaker Recognition in Far-field and Noisy Data
Karolina GPU 3700
The project's main goal is to substantially advance state of the art in robust multi-channel Speaker Recognition (SR). We will perform basic research in the domain of trainable multi-channel front-end that can be interconnected with arbitrary discriminatively trained speech application supporting training from single-channel raw signal or standard acoustic features (Mel-Frequency Cepstral Coefficients (MFCCs), Mel filter bank outputs). As a result of joint training of the signal processing part (beamformer) and the speech application, we expect better performance than what can be achieved with fixed beamforming tuned often only for perceptual speech quality. We will also compare our results with the latest state-of-the-art beamforming methods designed to process the multi-channel audio data independently of the target application. We will focus on the speaker embedding extractor, which is the main building block for Speaker Verification and speaker diarization systems.
Řešitel: Radim Špetlík
Weak Signal Analysis in RGB Images
DGX-2 500, Karolina GPU 4500
Glass-reflection removal, or glass-glare removal, is a problem of significant practical importance with applications ranging from license plate reading  to digital cleaning of camera optics . Given a single photo, the task is to remove reflections, or glares, without affecting the background. Much research in recent years has focused on instances of reflection removal constrained by specific qualitative attributes, such as requirement of full-screen reflection [6,7,8]. Glass-glare removal research has assumed constraints by development environment [3,1], or requirements of additional specialized hardware [4,5,2].
We address the general problem of reflection, or glare, removal aiming at reducing heavy constrains required by previous work with focus on the whole spectra of the problem – we consider reflections of all sizes and strengths and we only require a single RGB image as an input.
Řešitel: Sergiu Arapan
The effect of the 3d ferromagnetic metal interface on the topological spin states of Bi(1-x)Sb(x)
Karolina CPU 30000, Karolina GPU 3000, Karolina VIZ 40
Bi(1-x)Sb(x) is a topological insulator with protected metallic surface states. The carriers in the topologically protected surface states (TSS) have their spin and momentum locked at 90 degrees due to strong spin-orbit coupling (SOC). This property has been proposed as the foundation of applications based on magnetization switching by electric current. One switching mechanism is the spin-orbit torque (SOT) that occurs when at the interface of a ferromagnetic and material with strong SOC. This phenomenon demonstrates the conversion between charge and spin currents. The spin-to-charge conversion involving TSS is considered a promising route for generation of terahertz (THz) emission in spintronic devices. In this work we investigate the electronic structure of the interface between the ferromagnetic layers of Co and Fe and Bi(1-x)Sb(x) films and the effect of the interface on the TSS.
Řešitel: Michal Kolar
Conformational properties of low-structured fragments of ribosomal proteins
Karolina GPU 3200
Proteins are ubiquitous biomolecules that play a role in almost all processes in living organisms. Proteins are synthesized on ribosomes as nonbranched chains of amino acids in the process called translation, where the ribosome acts as a highly efficient catalyst. The ribosome has evolved into a complex biomachine, in bacteria containing about 50 proteins and 3 strands of the ribosomal RNA. Early ribosomes were likely smaller and contained proportionally less proteins and more RNA then modern ribosomes. During evolution the interface of the two components has increased. It is unknown, what drove the selection of ribosomal proteins and how the rRNA-protein interactions evolved. In the proposed project, we will use atomistic molecular dynamics simulations to investigate the conformational dynamics of ribosomal protein fragments. Our results will enhance the understanding of how ribosome evolved at molecular level.
Řešitel: Ctirad Červinka
Interpretation of anisotropy of molecular crystals from first-principles
Barbora FAT 1000, Karolina CPU 32000, Karolina VIZ 40
Macroscopic material properties are always an imprint of the character of the microscopic structure and interactions at the atomic scale. Molecular crystals are built from (in)organic molecules, often exhibiting significant asymmetry, which are regularly packed in the crystal lattice. Inequivalency of local configurations of the neighboring molecules in three Cartesian directions, imposed by their packing, gives birth to appreciable anisotropy of such molecular materials. This phenomenon has serious consequences in design of pharmaceutical ingredients or organic semiconductors. Anisotropy can lead to preferential crystallization of (un)desired crystal modifications, affecting the large-scale production of medicaments. Conductivity of electric charge in molecular semiconductors is then dramatically different in all three direction, which has impact on design and manufacturing of thin-film optoelectronic devices, requiring a proper oreientation of the molecules in the thin layer. This proposal aims at using quantum-chemical calculations to interpret and to quantify the anisotropy in terms of molecular interaction, and their individual mechanistic components.
Řešitel: Martin Srejber
Towards the first realistic lipid nanoparticle: a multiscale simulation study
Karolina CPU 39000, Karolina VIZ 40
In the past few decades, liposomal delivery systems (such as lipid nanoparticles, LNPs) became one of the most promising delivery options in fields like cancer chemotherapy, gene therapy, liposome-entrapped drug delivery or vaccines. Recent advances in RNA-based medicine have provided new opportunities for facilitating messenger RNA (mRNA) as a potent genetically engineered information carrier. In the best-known recent example, Covid-19 vaccines, the immunogenic messenger RNA coding SARS-CoV-2 spike protein is enveloped by a liposome consisting of a mixture of natural and engineered lipids – ionizable lipids (ILs) PEGylated lipids, phosphatidylcholines, and cholesterol. Tunable properties of these lipids e.g., the modulation of charge of ILs depending on the pH environment or the length of the polyethylene glycol chain of PEGylated lipids make those lipids prominent candidates in liposomal drug delivery. The main aim of this project is to investigate and rationalize the complex and multi-step mechanism underlying vaccine delivery facilitated by lipid nanoparticles. First, we plan to derive novel coarse grain (CG) parameters for engineered lipids allowing us to study LNPs close to experimental sizes and time scales. We would like to investigate the preparation of LNPs by enhanced computational techniques to create a reliable atomistic model of LNP. Furthermore, we plan to study one of the key steps in drug delivery: the mechanism of LNP interaction with cell membrane.
Řešitel: Pablo Nieves
Computational design of spin-based terahertz detectors
Barbora CPU 30000, Barbora VIZ 40
Nowadays, terahertz waves are used in many technological applications such as food inspection and quality control, medical diagnostics, drug development, security screening and environmental monitoring. The EU-funded s-NEBULA project aims to create a platform of room-temperature spin-based terahertz building blocks based on novel magnetism and optics concepts. Here, in the context of the s-Nebula project, we explore and model different types of spin-based terahertz detectors by means of atomistic spin dynamics simulations.
Řešitel: Karel Carva
Magnetic ordering in two-dimensional van der Waals halides
Karolina CPU 18000, Karolina VIZ 40
Systems with magnetically ordered two-dimensional (2D) layers bound by van der Waals (vdW) interaction are getting increasingly interesting for high-tech magneto-electric and magneto-optic applications in nanostructures. Critical microscopic properties controlling formation of magnetism in low dimensions are magnetic anisotropy and exchange interactions . Here we calculate these properties by first principles methods and examine general features of finite temperature magnetic order in this regime. The method is applied to several interesting systems from this class, in particular vanadium trihalides. We study how is the Curie temperature affected by interlayer coupling in these systems, how it could be modified by pressure and how much these results differs from the mean-field model results. We also calculate phonon spectra, whose comparison with experiment could reveal more details about the structure. The improved understanding of 2D magnets obtained here could help in finding their representatives operable at room temperature and thus bring these intriguing systems closer to real world applications.
Řešitel: Pavel Hobza
Unusual stability of covalent dative boned, H-bonded, and charge transfer complexes with increasing solvent polarity
Barbora CPU 10000, Barbora FAT 200, Barbora GPU 800, Barbora VIZ 40, Karolina CPU 80000, Karolina FAT 200, Karolina GPU 29000, Karolina VIZ 40
The project aims to improve understanding of the role of solvent polarity on the stability of covalent dative and non-covalent complexes through computational approaches. These studies will be done in close collaboration with experimental partners who use state-of-the-art experimental techniques. This collaboration has the potential to contribute to understanding of the solvent effect on complex stability with great applicability in various fields. The DFT calculations, employing selected DFT functionals, will be performed to calculate the electronic, and optical properties of various complexes. These studies will be carried out using implicit as well as explicit solvents environment.
Řešitel: Mauricio Maldonado Dominguez
Transduction of chemical energy into motion. The case of pericyclic reactions.
Karolina CPU 31758, Karolina VIZ 40
The translation of motion from chemical reactions into useful energy is becoming increasingly recognized as fertile ground to increase the efficiency of such transformations, yet it remains vastly unexplored. In this project, we will investigate a series of pericyclic reactions in solution where recent studies show nonequilibrium energy redistribution. We will study to which extent this excess is funneled into the translational velocity of the nascent products and neighboring solvent molecules. We recently demonstrated that this energy excess yields insight into C-H cleavage (J. Am. Chem. Soc. 2020, 142, 3947, project OPEN 18-2) and bifurcating reactions (Chem. Sci., 2021, 12, 12682, project OPEN 22-17). The projected results will pave the way to the design of processes where the energy that is typically wasted during chemical reactions can be utilized in subsequent processes, towards maximization of their energy efficiency.
Řešitel: Jiri Brabec
Theoretical modeling of the CO2 capture in MOFs materials
Karolina CPU 30000, Karolina VIZ 40
Over the past few decades our improved understanding of the risks of climate change have pushed most of the world to commit to a reduction of CO2 emissions to 2°C over pre-industrial levels. Meeting this target will be increasingly difficult given that most models predict that, to limit warming, CO2 emissions would have to stop increasing by the 2nd half of the 21st century. The development of the CO2 capturing materials and detailed understanding of the process is very important for achieving this task. In this work, we focus on the theoretical modeling of the CO2 capture by metal organic frameworks (MOFs) based on copper paddle-wheel building unit. The theoretical study will be performed at DMRG, DMRG-AC and later at the DMRG-AC-in-DFT level, in order to properly describe different type of interactions including strong correlation effects, dynamical correlations and dispersion.
Řešitel: Prashant Dwivedi
INvestigating mechanical propertieS of high enTropY aLloys through molEcular dynamics simulations. (INSTYLE)
Barbora CPU 29223, Barbora VIZ 40, Karolina CPU 7200, Karolina VIZ 40
High-entropy alloys (HEAs)  are metallic materials that can crystallize as a single phase with substantially high entropy by mixing equally or near equally five or more principal elements. HEAs have given rise to intensive research activities in materials science and engineering because of their new type of multi-principal metallic structures that exhibit excellent mechanical properties such as high fracture resistance, tensile strength , good thermal stability, high corrosion resistance , and ductility at ambient and high temperatures , etc. Thus, in this project, we aim to study the mechanical properties of quaternary (four-component) Fe25Ni25Cr25Co25, and some selected quinary (five-component) (Fe20Ni20Cr20Co20Cu20, & Fe-rich ones like (Fe30Ni19Cr21Co17Cu13)), high-entropy alloys under uniaxial tensile loading by means of classical molecular dynamics (MD). The goal of the project is to investigate the effect of phase separation  on the mechanical properties such as elastic modulus and tensile strength, as well as the evolution of dislocations, stacking faults, and deformation twinning of HEAs with Cu and without Cu. The outcome of this research is especially important for improving the mechanical properties of HEAs which will help to expand the practical application of these alloys such as in aerospace and nuclear technologies, and most in general, in all kinds of challenging industrial applications.
Řešitel: Jiri Klimes
Accuracy and precision for extended systems IX
Barbora CPU 15000, Barbora VIZ 40, Karolina CPU 16000, Karolina VIZ 40
Computer simulations have become indispensable for understanding the majority of experimental observations and thus understanding our world. This is especially true at the atomic and molecular level where experimental observations can become very difficult to interpret. However, if simulations are to be useful, they need to model the system reliably. For example, if there are different pathways for a chemical reaction, the simulation needs to be able to correctly predict their relative importance so as to increase our understanding of the system. The world at the atomic level is governed by the complex laws of quantum mechanics. In practice, this means that increasing the reliability of simulations, i.e., using less approximations, increases the computational cost substantially. While we understand well the impact of some of the approximations, there are some for which our knowledge is limited. In some cases, these can lead to unexpectedly large loss of the reliability of calculations. In our project we analyse some of the problematic cases occurring for calculations of intermolecular interactions, either in small clusters, or in molecular crystals.
Řešitel: Georg Zitzlsberger
Applications of Deep Neural Network based Urban Change Detection using Remote Sensing
Barbora GPU 200, Karolina CPU 200, Karolina GPU 1000, Karolina VIZ 40
This work is a continuation of the HS BLENDED project (funded by ESA in 2020-2021) and OPEN-21-31 (ending July 2022). We developed and trained a set of neural networks to monitor urban changes using a combination of SAR and optical observations with mission pairs ERS-1/2 & Landsat 5 TM (1991-2011), and Sentinel 1 & 2 (2017-now). Originally, the three sites Rotterdam, Liege and Limassol were used. We would like to continue finding further uses and tailor the trained networks and involved methods. To extend the use, we consider a documentation of war related changes in Ukraine (Kyjiw and Mariupol) over 2022. In addition, we also analyze the utility of identifying urban changes due to floods (e.g. flood in 2021 in Ahrtal/Western Germany).
Řešitel: Jiří Jaroš
Modeling of Low Intensity Focused Ultrasound Using Convolutional Networks
Barbora CPU 5000, Barbora FAT 10, Barbora GPU 400, Barbora VIZ 40, Karolina CPU 1000, Karolina GPU 500, Karolina VIZ 40
Transcranial low-intensity focused ultrasound (LIFU) therapy is increasingly used for the non-invasive treatment of brain disorders. However, conventional numerical wave solvers are currently too computationally expensive to be used online during treatments to predict the acoustic field passing through the skull.
As a step towards real-time predictions, we developed a fast iterative solver for the heterogeneous Helmholtz equation in 2D using a fully-learned optimizer. The lightweight network architecture is based on a modified UNet that includes a learned hidden state. The network is trained using a physics-based loss function and a set of idealized sound speed distributions with fully unsupervised training (no knowledge of the true solution is required). The learned optimizer shows excellent performance on the test set, and is capable of generalization well outside the training examples, including to much larger computational domains, and more complex source and sound speed distributions, for example, those derived from x-ray computed tomography images of the skull.
In the previous project we managed to extend our 2D solver to a 3D one, however, the computational requirements for training showed to be extremely high, almost 10 days on a single Karolina GPU node. In this project, we would like to further prune the UNet structure, implement more robust evaluation of the loss function, and improve training and inference procedure.
Řešitel: Jana Precechtelova
Design of reliable structural ensembles for accurate and cost-effective fragment-based calculations of NMR chemical shift sequence trends in intrinsically disordered proteins
Barbora CPU 18806, Barbora VIZ 40, Karolina CPU 9400, Karolina VIZ 40
In the project, we design a computer-assisted approach that facilitates the interpretation of experimental NMR data obtained for intrinsically disordered proteins. In this way, it enables the understanding of disorder-function relationships. IDPs have a significant role in the regulation of molecular mechanisms that cause neurodegenerative diseases such as the Alzheimer and Parkinson’s. The flexibility of IDPS hamper their structural characterization solely by the experiment. Computational methods such as the ones to be applied in the current proposal alleviate the problem substantially. In the proposed project, we combine the sampling of IDP structures by molecular dynamics with fragmentation techniques for the automatic construction of model systems as well as with quantum mechanics (QM) calculations of NMR spectroscopy chemical shifts (CSs). Our objective is to perform density functional theory calculations of CS sequence trends in IDPs using a conventional and machine-learning generated structural ensembles. The calculations will demonstrate if small to modestly-sized ensembles can be designed that provide reliable CS sequence trends at a reduced computational cost. The project contributes to the development of approaches for computer-aided structural characterization of IDPs.
Řešitel: Vladimír Ulman
Fiji Bioimage Informatics on HPC - Path to Exascale
Barbora CPU 1700, Barbora VIZ 40, Karolina CPU 4200, Karolina GPU 400, Karolina VIZ 40
Bioimage Informatics on HPC allows IT4Innovations to be involved in research on new topics in the area of big biological image data processing on HPC.
In this project, we are developing a Fiji plugin called “HPC Workflow Manager”. Fiji, a popular open source image processing package, offers wealth of functionality but it targets mainly desktop systems and workstations. The growing volumes of image data nevertheless represent computational loads that are well beyond of what the workstations can handle. We aim, with our HPC Workflow Manager, to enable biological power-users to adopt their Fiji procesing pipelines to introduce parallelism into them and to be able to smoothly and comfortably execute them at HPC installations.
To date, there is no widely-adopted parallelization platform in Fiji for non-programmers. For example, the KNIME software is indeed friendly and flexible software also for biology users. It comes with its own seamless cluster execution of its workflows, including workflows that use nodes from the KNIME Image Processing community extension. Yet, surprisingly, this community is rather small. CellProfiler offers an AWS-based extension for distributed image processing, currently only as a set of command line-operated scripts. In Fiji, the BigDataProcessor2 plugins offer TB-sized image processing but only for a limited set of operations. Our HPC Workflow Manager offers, on the other hand, remote distributed execution of any Fiji pipeline.
Řešitel: Ilia Ponomarev
Reactive Molecular Dynamics Simulations of Mo-S-C Tribological Coatings (REMEDY)
Barbora CPU 2700, Barbora VIZ 40, Karolina CPU 18300, Karolina VIZ 40,
Friction and wear account for about a quarter of world’s energy consumption, which is why providing lubrication is essential for sustainable engineering and green technologies. Solid lubrication in the form of a thin protective coating is one of the most promising approaches in solving the problem via reducing friction and wear and extending lifetime of the mechanisms.
The most common solid lubricants are diamond-like carbon (DLC) and transition metal dichalcogenides (TMDs). Both, however, have some shortcomings. While showing high hardness and very low friction in humid air, DLC films with low hydrogen content tend to exhibit high friction coefficient (up to 1) in vacuum or inert gas atmosphere. Increasing hydrogen content helps to overcome the problem, but also reduces the hardness of the film. Pure bulk TMD coatings, while showing superlubricity in vacuum, exhibit low corrosion/oxidation resistance, low load-bearing capacity and high wear[6,7].
Combining DLC and TMD is a promising approach to get the best of the two: low friction in both vacuum and humid air, corrosion resistance, low wear and high hardness[8,9]. We aim to model the structures and the tribological properties of the DLC-MoS2 coatings on the atomistic level using reactive molecular dynamics within the ReaxFF approach. We are going to provide insights into structure of such films and the mechanisms of the tribological action in order to guide the experimental researchers in the field.
Řešitel: Alberto Marmodoro
Barbora CPU 15000, Barbora VIZ 40, Karolina CPU 5000, Karolina VIZ 40
A key property of magnetic materials is their ground state ordering. The simplest scenario of ferromagnets (FM) involves parallel atomic magnetic moments, which produce a macroscopic magnetization. In colinear antiferromagnets (AF) these are instead anti-parallel and produce zero net field, but still a variety of measurable effects. More complex, non-colinear arrangements are also increasingly in the spotlight.
In addition to the pattern of magnetic ordering, the new research line of “altermagnetism” (AM) is presently calling attention to the role of lattice geometry, in controlling the alternation of spin-polarization not only in terms of swapped majority and minority states across e.g. anti-parallel magnetic sublattices, but also as a function of specific directions in reciprocal space. An example is the colinear room temperature AM RuO2, for which the rotated oxygen octahedra around antiparallel Ru1 and Ru2 atoms produce a correspondingly rotated spin-polarized Fermi surface.
While experimental techniques like angular resolved photoemission spectroscopy (ARPES) provide direct insight on the electronic structure, the underlying physics can be expected to manifest itself also in other forms of spectroscopies. In particular, magnons are collective magnetic excitations that can be probed e.g. via inelastic neutron scattering. While the key features of these excitations, such as the dispersion slope at the Gamma point or the Landau-damped lifetime due to Stoner competition, are well-known in the case of FM or AF ground state, the case of altermagnetism is only starting to be explored. This project will examine the problem via ab-initio density functional theory (DFT) calculations.
Řešitel: Vojtech Mlynsky
Investigating folding landscapes of simple nucleic acid motifs to fine-tune the AMBER nucleic acids force fields.
Karolina CPU 3906, Karolina GPU 17800, Karolina VIZ 40
Atomistic molecular dynamics (MD) simulations represent an established technique complementary to experiments for investigating structural-dynamics of nucleic acids. However, contemporary MD methods still suffer from limited accuracy of empirical potentials (force fields), including imbalances in the non-bonded force-field terms. We have recently introduced several different approaches for improvement of the state-of-the-art AMBER RNA force field such as addition of a new term for modification of hydrogen bonds (general H-bond fix, gHBfix), separate adjustment of interactions formed by terminal residues via terminal HBfix, update of the standard RESP charge derivation model (introduction of WRESP-EP charges) and modification of targeted pairwise van der Waals parameters via non-bonded fix (NBfix). Here, we aim to probe simultaneous effect of all those recently introduced force-field modifications by employing extensive set of MD simulations of small model structural motifs. We plan to use established enhanced sampling methods to reconstruct their folding energy landscapes. Estimation of folding free energies will serve as an indication of to what extend is the stability of folded (native-like) conformations affected by the introduced force-field modifications. Besides RNA, we also plan to test the above-mentioned adjustments on small non-canonical DNA motifs. Our ultimate goal is to link together those modifications and introduce universally applicable AMBER RNA force field version.
Řešitel: Karel Sindelka
Computer simulations of colloidal solutions and organic crystals
Barbora CPU 15278, Barbora VIZ 40, Karolina CPU 4297, Karolina VIZ 40
Aqueous solutions are omnipresent in nature, industrial processes, and daily life. Understanding their behaviour in inhomogeneous environments (self-assembled or confined systems) in equilibrium aand non-equilibrium (shear flow) conditions is important in many applications from medicine to environmental protection. The first part of this project focuses on interactions of surfacant monolayers with soft surfaces that play important role in industrial or household products (e.g., fabric softeners, cosmetic products). We use mesoscopic simulations to provide molecular-level insights into chemical and physical behaviour of these systems in and out of equilibrium.
The second part of the project focuses on polycyclic aromatic hydrocarbons (PAHs) and their nitrogen derivatives. PAHs are useful functional materials in organic electronic devices (such as OLEDs or solar cells), but they are also of interest to widely different fields, such as astrochemistry (components of interstellar dust), so understanding their properties and behaviour is important. Incorporating nitrogen into these compounds (azaPAHs) allows for tuning their properties to find even more use in modern organic electronics materials. However, the accessibility of these compounds is very limited largely because the research is scarce. We use all-atom simulations to investigate the PAHs and azaPAHS with stress on their structural properties under wide range of temperatures above and below their melting points."
Řešitel: Michael Komm
Particle-in-cell simulations for development of an advanced Langmuir probe model
Karolina CPU 11400, Karolina VIZ 40
Plasma detachment is the foreseen regime in ITER and future thermonuclear reactors as it allows to protect the plasma-facing components from excessive heat loads. It is typically associated with low divertor plasma temperatures (≤ 5 eV) – conditions at which diagnostics such as Langmuir probes struggle to deliver reliable results. The aim of this proposal is to develop an advanced model of the swept Langmuir probe by means of particle-in-cell simulations. The advanced model should be able to improve the analysis of experimental measurements and improve understanding of physics processes related to detachment.
Řešitel: Ondřej Souček
EIS - Evolving Ice Shells
Karolina CPU 3500, Karolina VIZ 40
The recent discoveries of liquid water oceans in the interiors of Jovian and Saturnian satellites Europa and Enceladus have made these icy moons some of the most intensively studied Solar System objects with a great astrobiological potential. However, the deep ocean habitats, possibly harbouring life, remain hidden below the outer ice shells - a few to tens of kilometers thick barriers that make any direct contact with the ocean extremely challenging. This highlights the importance of the search for links between surface observations and interior structure and dynamics. Such a goal can be achieved by numerical modelling of the ice shell dynamics and by identifying its surface manifestations. This goal will be pursued by studying: 1) the elements of planetary ice shell evolution, i.e. processes on various spatial and temporal scales governing the shell deformation, stress development, shape evolution and rotational dynamics, and 2) the surface signatures of internal processes and assessing their relevance as potential observables by spacecraft missions.
Řešitel: Ales Vitek
Classical and quantum Monte Carlo simulations of complex microsystems
Karolina CPU 10000, Karolina VIZ 40
This project will be focused on computationally demanding classical and quantum Monte Carlo simulations of atomic clusters.
Particularly, classical Monte Carlo simulations will be performed on mercury clusters. Recently, we have published a few papers about mercury clusters, we computed their photo absorption spectra or thermodynamic properties. Small mercury clusters are bonded by van der Walls forces, while greater one are metal binded. We have found, that for Hg13 cluster, higher pressures deformate the global minimum structure into similar, but smaller geometry, where higher electronic quantum states affect the ground potential energy surface. We would like now to investigate, how the higher electronic states will affect the thermodynamics properties of larger mercury clusters.
The path integral Monte Carlo approach will be focussed on the theoretical modelling of photoabsorbtion spectra of cold helium clusters, where light helium nucleis together with low temperatures requires the quantum simulation approach.
Řešitel: Hermann Detz
Boron-incorporation into III-V alloys
Barbora CPU 8333, Barbora VIZ 40, Karolina CPU 4688, Karolina VIZ 40
Sensing systems based on mid-infrared absorption spectroscopy that can be integrated on a chip are relevant for a wide range of potential applications, including bio-chemical and medical diagnostics or monitoring of food quality. The hereby needed lasers and detectors rely on layered structures of semiconductor materials, which require precision down to single atomic layers. In order to fabricate the active material for these optoelectronic devices, it is necessary to fully understand the chemical reactions during material growth. This project puts particular emphasis on the addition of boron, which allows to compensate strain and therefore to use a wider range of possible materials. Until now, there are only sparse experimental reports on BGaAs, which shall be complemented by ab-initio density functional theory simulations to gain a deeper insight into the growth process and resulting material properties to pave the way for the next device generation.
Řešitel: Valeria Butera
Gallium Nitride-based materials as promising photocatalysts for CO2 reduction: a DFT study.
Barbora CPU 5556, Barbora VIZ 40, Karolina CPU 6250, Karolina VIZ 40
Photocatalytic CO2 conversion to clean fuels and chemicals is important for mitigating the climate change and reducing the dependence on fossil energy resources. In order to achieve this goal, effective photocatalysts need to be developed. Among them, gallium nitride (GaN) has been proven as efficient materials for the reduction of highly stable CO2 molecule. Its photochemical activity can be notably improved by the incorporation of Mg doping or substitutional alloying with In. Preliminary studies have been performed to explore the photocatalytic properties of bare and mono Mg-doped GaN, and the different routes leading to the production of methanol. Within this project, we further extend our density functional theory (DFT) investigation performing periodic boundary condition (PBC) calculations, which allows us to employ a more extended surface for a detailed analysis of the CO2 coverage, by considering the adsorption of one to nine CO2 molecules on the pristine GaN(100). We address also doping and alloying considering up to eight different incorporation sites for both Mg and In. Eventually, we evaluate the CO2 adsorption on doped Mg:GaN and substitutional InxGa1-xN alloys. As second step of our study, we will evaluate the water adsorption on two different non-polar GaN surfaces, namely (100) and (110), and on the corresponding defective GaN surfaces. We will then calculate the phase stability diagrams to evaluate which surface is more stable under selected conditions.
Řešitel: Ales Prachar
Barbora CPU 11500, Barbora VIZ 40
The project is focused on the study of the flow field behind the propeller and its impact on other parts of the UAM aircraft by advanced CFD and optimization methods. Considering this flow field affecting by the propeller during the design process of individual aircraft parts, it seems to be a very important parameter that will allow aerodynamic design of aircraft with higher efficiency and lower emissions. The resulting shape and position of the individual parts of the aircraft will probably be different from the one that would be proposed by the current procedure without the influence of the propeller wake. An example is the design of a winglet in a propeller wake, which should have the highest efficiency in the cruise, but at the same time should not degrade the performance and properties in other flight regimes and the different loading of the propeller. The design of the aircraft with lower emissions is fully in line with the regulation of European Commision.
Řešitel: Rene Kalus
Ternary recombination processes in cold argon plasmas
Karolina CPU 6600, Karolina VIZ 40
Molecular ions play an important role in cold plasmas of heavier rare gases and crucially influence their interactions with environment, production of secondary ions, and potential benefits of their use, e.g., in biomedicine. Argon is by far the most often considered buffer gas in rare-gas plasma generators due to its cheapness. Specifically, two-atomic ions are the most important composite ions of argon plasmas and ternary recombination processes (involving a simultaneous interaction of an atomic ion with two carrier gas neutrals) represent a very important reaction channel leading to their formation. Importantly, both electronically ground-state atomic ions and fine-structure excited ones play a role and advanced, inherently non-adiabtaic dynamics (NAD) methods must be used to realistically model the recombination processes. In a series of previous calculations, we studied collisions of atomic and diatomic ions of various rare gases using a home-made NAD code and showed robustness and reliability of employed approaches by reproducing available experimental data and predicting and explaining phenomena not directly accessible by measurements. The main intention of the present project is to extend the previous work to a question how molecular ions are formed in cold rare-gas plasmas and what are the characteristics of their nascent populations. The project is part of a broader research aim solved in a tight collaboration with several groups at the Université Toulouse III.
Řešitel: Pavlo Polishchuk
Karolina CPU 6200, Karolina VIZ 40
CACHE challenges focus on specific protein targets of biological or pharmaceutical relevance. Participants should predict hits and CACHE will validate these hits experimentally. Each competition includes a hit-finding and a hit expansion round of prediction and experimental testing after which all data, including chemical structures, will be made publicly available without restrictions on use.
The target of the first CACHE Challenge is Leucine-rich repeat serine/threonine-protein kinase 2 (LRRK2), the most commonly mutated gene in familial Parkinson's Disease (PD). PD-associated LRRK2 mutations tend to promote LRRK2 filament formation and enhance LRRK2 interaction with microtubules. Recent structural data reveals that only compounds stabilizing the open form of LRRK2 antagonize the pathogenic formation of LRRK2 filaments in cells, but most kinase inhibitors stabilize the closed form of LRRK2. An alternative and so far overlooked strategy is to pharmacologically target the WDR domain of LRRK2, which is juxtaposed to the kinase domain. The WDR domain in LRRK2 may be important for recruiting LRRK2 signalling partners or for binding to tubulin. Identifying chemical starting points binding to the WDR domain of LRRK2 is a novel approach to target this protein.
Řešitel: Martin Matys
Laser-driven ion acceleration using structured targets III
Karolina CPU 4440, Karolina VIZ 40
Laser-plasma ion accelerators are currently receiving particular scientific attention as promising source of accelerated charged particles, since they are able to generate much stronger electric fields in comparison with conventional accelerators and can possibly replace them in future in several impressive applications, including proton therapy for the treatment of the cancer cells, production of PET (positron emission tomography) medical isotopes, generation of ultrashort neutron pulses, radio isotope source, etc. To improve the maximal energy and quality of laser-accelerated ions a novel scheme for ion acceleration is investigated in this project. Interaction of high-intensity laser pulse with thin overdense double layer targets with initial corrugation on the interface, results into controlled rupturing of the foil. The remaining bunches are then accelerated as whole well-collimated structures, exhibiting monoenergetic behavior. The increase of the laser pulse power and intensity also brings questions about the effects of quantum electrodynamics on the ion acceleration in this unexplored multi-10 PW physics, which will be addressed in this project.
Řešitel: Frantisek Mihok
Simulated thermoelectric properties of alloys
Karolina CPU 317200, Karolina VIZ 40
Energy consumption is steadily growing with a growing population. Energy needs of modern societies are on the rise as well. Renewable resources and energy management efficiency are proving essential in maintaining energy suficiency while not destroying the planet environment further. Thermoelectric generator provide unique opportunity to improve energy management. These modules made from specialized materials with suitable thermoelectric properties offer cost effective, simple, maintenance free and reversible solution for managing energy from waste heat. Modules consist of multiple N and P semiconductor pairs connected in series. These devices offer flexibility in their use as heating or cooling device or electricity generating device because of Peltier and Seebeck effects respectively. The main downside of current thermoelectric modules is their low efficiency but new materials with marginally improved conversion efficiency sparked renewed scientific and even commercial interest in thermoelectric modules.
This project looks to find and evaulate new materials which exhibit significant thermoelectric properties. Specific focus will be placed on promising SnSe alloys. Materials will be modeled and optimized using DFT methods of NWChem. Afterwards their theoretical thermoelectric properties will be determined using Quantum Espresso and LAMMPS molecular dynamics. Lastly, interaction of materials with their surroundings and their properties under different temperature gradients will be investigated using molecular dynamics methods as well.
Řešitel: Jan Roman
AI-assisted development of medical image segmentation algorithms
Barbora GPU 100, Barbora VIZ 180, Karolina GPU 150, Karolina VIZ 180
Medical image processing can help to find and understand irregularities in the human body. It can detect or even predict diseases. As a data source, Magnetic Resonance Imaging (MRI) or Computed Tomography (CT) is often used. This project is a joint cooperation between University Hospital Ostrava and IT4Innovations and it aims to develop a solution for medical doctors that would allow them to automatically segment selected tissues from medical images. In this way, time of the process from medical image analysis to diagnosis can be reduced significantly and it can improve the level of personalized medicine.
Řešitel: Vladislav Pokorný
Thermopower in superconducting nanohybrids
Barbora CPU 7000, Barbora VIZ 40
Superconducting nanohybrids are nanoscopic devices in which active elements such as carbon nanotubes or semiconducting nanowires are connected to a combination of superconducting and metallic electrodes. Such systems gained a lot of interest recently for their applications in quantum computing, rapid single flux quantum electronics and sensor technology. Understanding the complex interplay among the various phenomena is a necessary step in developing a new generation of such devices. Thermoelectric effects that connect the flow of electric current with temperature difference between the parts of the device are one of the most interesting phenomena. Supercomputers are now a necessary tool that helps us to build our theoretical understanding and explain the available experimental results before these systems can become the tools to extend the abilities of the current silicon-based electronics.
Řešitel: Miroslav Pospíšil
Interactions of cadmium sulphide subnanoparticles.
Barbora CPU 6000, Barbora VIZ 40
Nanoparticle based structures play significant role in recently developed progressive materials. Its applications are found in catalytic processes, construction materials improvement, drug delivery and many other industrial aspects. New catalytic processes were discovered and these new photochemical, photoelectric and photon conversion processes offer actual application potential. Photocatalytic reactions of CdS and other chalcogenide nanoclusters are usable for the carbon dioxide photocatalytic reduction coupled with the transformation of carbon dioxide into simple organic molecules calculated by the authors presently. Photocatalysts offer effective reaction processing ways suitable for eco-friendly chemical manufacturing and waste recycling. Many technological industrial processes requires combustion heating and cannot be replaced effectively by another energy transfer. Increasing interest in pure hydrogen and hydrogen rich fuels as replacement of fossil ones, faces the difficulties of hydrogen gas storage and transport and also difficult usage in eg. metallurgical processes and hot material processing, where hydrogen dissolving degrades quality of steel or chemical reaction changes material composition. Semiconductor catalysts are significantly cheaper than commonly used platinum catalysts and can be engineered and designed for required process activities.
Řešitel: Marta Cudova
Offloading of Workflows Executions to Remote Computational Resources II.
Barbora CPU 3500, Barbora VIZ 40, Karolina CPU 700, Karolina VIZ 40
In recent years, the therapeutic ultrasound has grown in a number of applications such as tumor ablation, targeted drug delivery or neurostimulation. Precise preoperative treatment planning tailored individually to each patient is crucial for the maximalization of the treatment’s outcome.
The fundamental challenge shared by all applications of therapeutic ultrasound is that the ultrasound energy must be delivered accurately, safely, and noninvasively to the target region within the body identified by a medical doctor. The estimation of treatment outcome heavily depends on the computationally very intensive ultrasound, thermal and tissue models which are only realizable with the use of HPC facilities.
Since there is a lack in expertise of the clinical end-users to use HPC resources efficiently, we developed the k-Plan software simplifying the everyday use of the HPC resources without a need to specify task execution parameters, dependencies, and their monitoring. k-Plan also reacts to the problem of task execution parameters selection, e.g., number of nodes, estimated execution time, storage space, etc. The end-users have no knowledge about the strong and weak scaling of the software being used, yet these characteristics have great impact on the calculation cost and overall execution time.
The goals of this project are to (1) deploy the k-Plan software and offer it to a small set of pilot end users from clinical environment to execute realistic treatment plans, (2) to tailor the task submission planning logic to IT4Innovations clusters, (3) and investigate methods that allow k-Plan to automatically tune execution parameters for individual tasks.
Řešitel: Karel Tůma
Numerical study of formation and e volution of omega phase in Ti alloys
Karolina CPU 3000, Karolina VIZ 40
Modern aircraft belong to the most sophisticated machines. Their efficiency and reliability are based on decades of multidisciplinary research. The same applies to surgical fixation of fractures or surgical replacement of the hip or knee joint, which are among the greatest advances in medicine.
Metastable β-titanium alloys are known for their optimal combination of strength and ductility, which is obtained by complex thermomechanical treatment accompanied by several phase transformations. One of the most studied is the β → ω transformation during which small nano-particles of the ω phase are formed. This process is a result of two major mechanisms: elasticity and diffusion. However, it is rather impossible to design experiments that would separate them and enable us to fully understand the formation of the ω phase.
We aim to develop a phase-field model to describe the formation and evolution of the ω phase and study how the individual mechanisms influence this process.
Řešitel: Pavlo Polishchuk
De novo design of synthetically feasible compounds
Karolina CPU 3000, Karolina VIZ 40
Development of new drugs is the search of small molecules which interact with a particular target and produce a desired response. The complexity of this task is greatly determined by the vastness of the drug-like chemical space which size is estimated as ~1033 compounds. In the nearest future, it will be impossible to enumerate this space or perform any kind of exhaustive search. One of the popular strategies to explore and navigate chemical space is de novo design - model-driven generation of new chemical structures with promising predicted properties. It can go far beyond of the currently available libraries and can discover new chemical entities. The key feature of de novo approaches is structure generation. However, synthetic accessibility of generated structures remains the main issue. We combined the previously developed approach to structure generation CReM, which allows to control synthetic feasibility of generated molecules, with molecular docking in order to generate compounds fitting to a binding site of a chosen protein. The current study is a continuation of the previous project and will be focused on improvement of developed in silico modeling tools and their application to design molecules for a clinically relevant target - zinc metalloprotease Zmp1 of Mycobacterium tuberculosis.
Řešitel: Tomas Panek
Phylogenomics of Heterolobosea
Karolina CPU 2440, Karolina VIZ 40
Phylogenomics has proven to be indispensable for inferring the eukaryotic tree of life. However, there is considerable variation in dataset construction approaches that can be a source of various artefacts (e.g. due to contaminations or insufficient identification and removal of paralogs). We will use Phylofisher software for proper ortholog detection and contamination removal and we will reconstruct the phylogeny of Heterolobosea using transcriptomic and genomic data from all main lineages of the group. As Theodosius Dobzhansky said: “Nothing in biology makes sense except in the light of evolution.” We will use results of this project as part of the bigger initiative aiming to illuminate genetic background of extraordinary ecological diversity of Heterolobosea. It will provide a phylogenetic framework to test our hypotheses on the evolution of Heterolobosea.
Řešitel: Petra Sukova
Stellar transits through accretion flow onto supermassive black holes
Karolina CPU 2200, Karolina VIZ 40
We will study how the repetitive transits of a star in the close neighbourhood of the supermassive black holes in the centers of Galaxies can influence the process of accretion of plasma. We will use general-relativistic magneto-hydrodynamical code to follow the motion of the gas, while it is being pushed by the star along its trajectory. Some of the gas may be expelled from the accretion disc into the polar region, where a strong magnetic field ordered along the axis of rotation of the black hole accelerates the gas away from the center up to relativistic speed. On the other hand, the density waves caused by the motion of the star in the medium are spreading both downwards and outwards and hence influence the accretion rate on the black hole and the density distribution in the disc. To see the complicated structure of the outflow and inflow, robust 3D simulations are needed with resolution high enough to capture the details of the flow close to the black hole horizon while stretching far enough to see the outflow on larger scales. We will find the observational traces of the presence of the star close to the nucleus, which will help to find promising targets for the upcoming European space gravitational observatory LISA (Laser Interferometer Space Antenna).
Řešitel: Ales Prachar
Barbora CPU 900, Barbora VIZ 40, Karolina CPU 800, Karolina VIZ 40
The participation on the AIAA CFD Drag Prediction Workshop is a good way how to assess quality of the CFD simulations. The high quality validation data, which are used as a reference, allows studying advanced flow phenomena, like shock induced flow separation. The study cases focused on force/moment and pressure predictions for the NASA Common Research Model wing–body configuration. During the previous Drag Prediction Workshops, the aeroelastic effects were studied, leading to the static aeroelastic twist and deflection at each specified condition. Various grid topologies are used for the grid refinement studies, the angle of attack sweep and a Reynolds number study is the part of the work.
Řešitel: Ivan Kološ
Numerical modeling of load of structures in quasi-static effect of wind
Karolina CPU 1200, Karolina VIZ 40
The project is focused on numerical modeling of flow around objects in the atmospheric boundary layer. This issue is complicated mainly due to the atmospheric turbulence, which requires the use of advanced numerical models of the flow coupled with detailed computational mesh of the domain. This research will contribute to bigger efficiency in design of building structures.
Řešitel: Andrzej Kadzielawa
Designing effective interatomic potentials for Tungsten-based alloys
Barbora CPU 33000, Barbora VIZ 40, Karolina CPU 24000, Karolina GPU 11800, Karolina VIZ 40
The progress of fusion technology incentivizes the development of the materials applicable in reactor vessels as the so-called first wall. This material is required to fulfill a set of rigorous conditions, e.g., high thermal conductivity, hardness, or melting temperature. Currently, the best performing materials are Tungsten-based alloys. However, since they are produced as saturated/unstable solid solutions, the fail-safe performance depends on preserving the unstable state. While the primary goal of the research (this project is part of) is the designing of a Tungsten-Chromium alloy doped with a third transition metal, the question of the melting temperature is the one in need of answering after some candidates are proposed.
Our central hypothesis is that by using the density functional theory, we are able to design a set of effective interatomic potentials to model the solid-to-liquid transition of such a structure realistically.
Řešitel: David Zihala
Resistance and pathogenesis of extramedullary multiple myeloma
Barbora CPU 1111, Barbora VIZ 40, Karolina CPU 278, Karolina VIZ 40
Extramedullary Disease (EMD) is a more aggressive manifestation of the second most common blood cancer, Multiple Myeloma (MM). In a standard MM, malignant plasma cells (PCs; terminally differentiated B cells) clonally proliferate in a bone marrow which ultimately leads to anaemia, myelosuppression, bone lesions, renal failure and other clinical consequences resulting from paraproteinemia. EMD is characterised by malignant PCs independent on a bone marrow microenvironment. This critical acquired feature enables tumour PCs to infiltrate other tissues and organs and means worse prognosis for patients. The incidence of both MM and EMD is increasing worldwide. While MM remains incurable, extensive research has led to significant improvement in treatment and consequent extension of survival time for patients. In contrast, the knowledge about MM to EMD transition and further development of EMD is limited to only a few publications and molecular therapeutic targets are still missing. In our study, we focus on molecular characteristics and tumor microenvironment of EMD, using bulk and single-cell RNA sequencing to shed more light on the resistant nature and pathogenesis of this disease.
Řešitel: Gabriela Necasova
Parallel Numerical Solution of Differential Equations Using Modern Taylor Series Method
Barbora CPU 1200, Barbora VIZ 40
The partial differential equations (PDEs) and ordinary differential equations (ODEs) are widely used in many real-life problems. Differential equations arise from problems in physics, engineering, mechanics, and other sciences. The analytic solution of differential equations, particularly the solution of initial value problems (IVPs) in higher-order equations, is often difficult and, in most cases, practically impossible. Several numerical methods can be used for the numerical solution of PDEs. One of them is Method of Lines (MOL) which transforms the PDE into a system of ODEs. The project deals with the numerical solution of IVPs described by the system of ODEs. The goal of this project is to show positive properties of the newly proposed parallel version of Modern Taylor Series Method (MTSM) that offers extremely accurate, stable, and fast numerical solution of systems of ODEs. The numerical solution obtained by MTSM will be compared to the other state-of-the-art ODE solvers.
Řešitel: Urszula Wdowik
Phase transition in TiZrNbHfTa(N) high entropy composite upon pressure and temperature
Barbora CPU 76032, Barbora VIZ 40, Karolina CPU 33600, Karolina GPU 4900, Karolina VIZ 40
High entropy alloys (HEA) represent a class of novel engineering materials with outstanding mechanical properties, including high-temperature strength, fracture toughness, high hardness, good corrosion and wear resistance, low creep or high ductility, that are essential demands on functional materials for highly challenging applications. Here, we propose ab initio research of phase transformation of newly synthesized TiZrNbHfTa HEA doped with nitrogen, which is induced by external pressure and high temperature. This theoretical study is intended to support and explain results of the ongoing experimental investigations performed within the research project Development of novel high entropy composites with nitrogen synthesized under high pressure and temperature, Grant No. 2021/41/B/ST8/03758, National Science Center (NCN) Poland.
Řešitel: Martin Zelený
Ab initio calculation of novel 3rd generation data for thermodynamic modelling
Barbora CPU 8500, Barbora VIZ 40, Karolina CPU 43200, Karolina VIZ 40
The thermodynamic modeling has recently become extremely useful to metallurgists, engineers and materials scientists in development of new alloys for specific applications or design and optimization of materials synthesis and processing processes. In all these areas, the use of phase diagrams allows research, development, and production to be done more efficiently and cost effectively. While phase diagrams have proved invaluable for materials science, their prediction within so-called CALPHAD (CALculation of PHAse Diagram) approach critically depends on either experimental data or, when experiments are difficult/impossible, materials parameters determined by theoretical calculations. The proposed project aims at computing these data using modern quantum-mechanical calculations.
Řešitel: Amina Gaffour
Calculation of NMR spin-spin couplings for intrinsically disordered proteins: a prospective tool to facilitate experimental NMR studies
Barbora GPU 16300, Karolina CPU 9375, Karolina VIZ 40a
The phosphorylation and de-phosphorylation of intrinsically disordered proteins (IDPs) regulate a vast range of molecular processes that are responsible for the development of neurodegenerative diseases. The structure characterization of IDPs by nuclear magnetic resonance (NMR) spectroscopy has to be facilitated by NMR parameter predictions in order to construct structural ensembles that best represent the measured NMR data. The prediction of NMR spin-spin couplings typically builds on empirically parametrized Karplus equations . Alternatively, quantum mechanics (QM) calculations can be applied to guide the construction of structural ensembles. The principal objective of the proposed project is to develop an accurate computational approach for the calculation of NMR spin-spin couplings that combines the molecular dynamics (MD) calculations with density functional (DFT) calculations . The design of the MD/DFT computational protocol will be facilitated by the validation against experimental datasets as well as empirical predictions. The proposed project can contribute to the understanding of molecular processes that lead to neurodegenerative diseases and thus also to the development of new medical treatments.
Řešitel: Oldřich Plchot
Training Speaker Embedding Extractors Using Multi-Speaker Audio with Unknown Speaker Boundaries
This research aims to develop methods for training speaker embedding extractors using data with weak annotations or even without any annotations. More specifically, we are laying bases for utilizing vast amounts of readily available data online but which come without careful annotations. However, we can often obtain some metadata along with it. We will demonstrate the developed methods on the full (uncut) VoxCeleb recordings for which we know the names of the celebrities appearing on each video without knowledge of the time intervals the celebrities appear in the video. We intend to keep the whole process fully independent of any systems trained on outside data, including an initial speaker diarization necessary for rough pre-segmentation. We will introduce modified objective functions with aggregation over segments that will be used to train a competitive ResNet-based embedding extractor, but without the need of a large-scale, fully annotated dataset. We intend to release a codebase that allows training on data that could not be so-far efficiently utilized. It will help researchers and industry to address under-resourced (in terms of supervised training data) domains such as the speech of speakers of rare languages.