Issue #03
June 2017

Message from the co-ordinator

VI-SEEM project is now at its mid-point, and the infrastructure, data, services and user communities are now all coming together. The project has successfully completed its mid-term review carried out by European Commission and the independent experts, held in Brussels in May.

The services of the Virtual Research Environment are provided through a compact service catalogue available at https://services.vi-seem.eu/ui/catalogue/services/, while the VRE portal at https://vre.vi-seem.eu/ provides a comprehensive environment for collaborative research for the communities of Southeast Europe and the Eastern Mediterranean regions.

VI-SEEM second Call for Proposals for projects accessing the VI-SEEM services and associated infrastructure is open, available at https://vi-seem.eu/2nd-call/. The call is addressed to scientists and researchers that work in academic and research institutions in the region in the fields of Life Sciences, Climate research, and Digital Cultural Heritage, and we are excited to welcome new applications and user groups in joining our community.

This newsletter also showcases new applications and services provided by the project, as well as project key regional and national training and dissemination events.

VI-SEEM joined forces with PRACE to organize Spring School in Cyprus

System Administration and Data/Computational Services for Scientific Communities

Members of the PRACE, VI-SEEM and European communities got insights into topics of high interest during a 3-day training event that took place at the Computation-based Science and Technology Research Center (CaSToRC) of The Cyprus Institute on 25-27 April 2017. The event brought together researchers from Europe and the Eastern Mediterranean to train scientists and technical personnel in utilizing HPC for their research and in operating such facilities.

Two parallel workshops were held, one to train researchers in developing codes (Developer track) and one to train technical personnel to operate HPC infrastructure (Systems Administration track).

The Developer Track brought together trainers from The Cyprus Institute, the National Center for Supercomputing Applications (NCSA) and the Juelich Supercomputing Center who delivered various subjects such as “Parallel and Scalable Machine Learning and Deep Learning”, “Accelerator Programming” and “Data Visualisation”. In total 29 participants followed this track from Bulgaria, Cyprus, Greece, Hungary, Romania, USA and other regional countries.

The Systems Administration Track was initiated by CaSToRC and it was the first of its kind in a PRACE seasonal school. It was attended by 35 participants from Cyprus, Czech Republic, Germany, Greece, Hungary, Israel, UK, Egypt and other South Eastern European countries. Trainers from Ghent University, NCSA, Juelich Supercomputing Center and industry delivered various subject such as “Modern Scientific Software Management”, “Security Aspects of HPC Centers”, “System Monitoring”, “Data Services” and “Cloud Services”. The track also included a discussion on the “Future of HPC Administration” where participants and trainers gave their opinions, comments and advice on the future trends they believed would shape future trends in HPC System Administration.

The taining material, including YouTube streams of the lectures, is available at the event website: https://events.prace-ri.eu/event/601/page/4

VI-SEEM 2nd call for accessing resources and services:
submissions deadline extented to 26 June

VI-SEEM has extended the deadline for submitting project proposals for accessing the VI-SEEM services and associated infrastructure in the context of its second call, until 26 June 2017. The call is addressed to scientists and researchers that work in academic and research institutions in the region of South Eastern Europe and Eastern Mediterranean - specifically Albania, Armenia, Azerbaijan, Bosnia and Herzegovina, Bulgaria, Cyprus, Egypt, Former Yugoslav Republic of Macedonia, Georgia, Greece, Hungary, Israel, Jordan, Lebanon, Moldova, Montenegro, Romania, Serbia, and Turkey.

The project proposals should address open research topics in the fields of Life Sciences, Climate research, and Digital Cultural Heritage.

Via this call VI-SEEM opens possibilities for regional scientists from the selected scientific fields to gain access to the advanced resources and services that it offers.

The list of services and resources offered by the VI-SEEM VRE can be found at the VI-SEEM service catalogue and in the VI-SEEM VRE portal.

Access to underlying computational resources will be awarded for a maximum period of 12 months, while access to underlying storage resources may be provided for up to 2 years.

Important Dates:

Opening date: 4th May 2017
Closing Date: 26th June 2017
Allocation decision: by September 2017

All the details of the call and the application form are available at:
https://vi-seem.eu/2nd-call/

VI-SEEM Life Sciences applications: NANO-Crystal

Nanoparticles (NPs) as drug delivery systems have shown significant promise in cancer treatment, where they are used to improve the biodistribution of cancer drugs. Thus, nanoparticles need to be designed with optimal size and surface characteristics in order to decrease side effects and drug toxicity while maximizing treatment impact. Computational approaches assist researchers in this design by modeling nanoparticles to systematize how MNP attributes affect their interaction with cell components as well as with their drug loading. A limiting factor in such modeling studies by the wider scientific community is the absence of a tool that constructs the morphology of nanoparticles. NANO-Crystal is a web-based tool, which constructs spherical nanoparticles of a given radius defined by the user. Apart from the creation of the spherical nanoparticles, code that constructs crystal nanoparticles has been already developed and it is going to be released as an update in NANO-Crystal within 2017. This computational toolbox computes the macroscopic morphology of any periodic crystal by forming different shapes based on Miller indices and it is able to make a link between macroscopic morphology and atomistic structure for a periodic crystal, which is a valuable tool for scientists.

NANO-Crystal is a web-based tool, implemented for the construction of nanoparticles of a given radius. In this version, the spherical coordinates tool has been implemented. This tool creates coordinates for spherical nanoparticles. The algorithm works by calculating the number and the Cartesian coordinates of smaller spheres that fit on the surface of the nanoparticle and then produces the output morphology. The home page menu allows two selections for the user: (i) the radius of the nanosphere (nm), and (ii) the radius of smaller spheres (nm), that will cover the surface of the nanoparticle. The program computes the number of smaller spheres that fit on the bigger surface and the user can download their Cartesian coordinates (output format .xyz). The program code is implemented using PHP server-side scripting language, which is embedded into the HTML and CSS code. JQuery, a cross-platform JavaScript library, is also used. For local host of the webpage tool, the Wamp server is used. Moreover, a crystal computational morphology toolbox for constructing and modeling different crystal nanoparticle shapes has been developed. Computational approaches are used for computing the macroscopic morphology of any periodic crystal by forming different shapes based on Miller indices and the distance measure from the center of the crystal and visualizing the resulting crystal. That crystal is a polyhedron that is created as the intersection of multiple polyhedra and individual planes via the steps that follows. This tool is planned to be imported in the NANO-Crystal webserver in Q3 of 2017.

Server: http://nanocrystal.vi-seem.eu
Manual: http://www.drugdesign.gr/uploads/7/6/0/2/7602318/nano-crystal.pdf

Digital Cultural Heritage

Life Sciences

Climate

VI-SEEM climate applications

Through VI-SEEM, scientists at the Cyprus Institute (CyI) have developed an application of GPU accelerators in Earth system modelling. They developed software that generates CUDA kernels for numerical integration in the global climate model EMAC, used to study climate change and air quality. The focus is on atmospheric chemical kinetics, the most computationally intensive task in climate-chemistry simulations. This approach can serve as the basis for hardware acceleration of numerous geoscientific models that rely on KPP for chemical kinetics applications. A source-to-source compiler outputs a CUDA compatible kernel, by parsing the FORTRAN code generated by the Kinetic Pre-Processor (KPP) general analysis tool. Performance evaluation, using Fermi and Pascal CUDA-enabled GPU accelerators shows achieved speedups of 4.5× and 22.4× respectively, of the kernel execution time. A node-to-node real-world production performance comparison shows a 1.75× speed-up over the non-accelerated application. The accuracy and correctness of the accelerated implementation are evaluated by comparing to the CPU-only version of the application. The approach followed, including the computational workload division and the developed GPU solver code can potentially be used as the basis for hardware acceleration of numerous geoscientific models that rely on KPP for atmospheric chemical kinetics applications.

Citation: Alvanos, M. and Christoudias, T.: GPU accelerated atmospheric chemical kinetics in the ECHAM/MESSy (EMAC) Earth system model (version 2.52), Geosci. Model Dev. Discuss., doi:10.5194/gmd-2017-63, in review, 2017.

VI-SEEM Digital Cultural Heritage applications: the CLOWDER data management system

Clowder (http://dchrepo.vi-seem.eu) is a research data management system designed to support any data format and multiple research domains. It contains three major extension points: preprocessing, processing and previewing. When new data is added to the system, preprocessing is off-loaded to extraction services for extracting appropriate data and metadata. The extraction services attempt to extract information and run preprocessing steps based on the type of the data, for example to create previews. This raw metadata is presented to the user in the Clowder web interface. Users can upload, download, search, visualize and get various information about these data. Data, in the case of the VI-SEEM project and more specifically in the field of Digital Cultural Heritage, can be of very diverse types and formats. More specifically users can upload massively (zipped) or individual files of:

3D Models: where extractors clean up and prepare for visualization on the platform itself.
Scanned books and their metadata: OCR algorithms will be used to extract the text in the documents so that users can find books using both metadata information and the book’s contents.
Image, text and sound files and their metadata, organised in collections.
Advanced documentation data, such as Reflectance Transformation Imaging, and analysis of material properties of structures, works of art and artefacts.

Example of interactive data visualization: screengrab of Clowder page illustrates online interactive preview of a model documented with Reflectance Transformation Imaging technique in WebGL, while its linked metadata displayed on the right side of the screen. PTM data uploaded by user can be accessed, interacted with and downloaded (provided it is permitted by its author) by other users of the platform. Researchers from the Centre for the Study of Ancient Documents, of Oxford University, holding one of the largest epigraphical archives in the world, will be relying on this feature to upload Ptolemaic inscriptions and offer a wealth of archive materials that were previously unavailable online.

VI-SEEM team ranked in the top 6 teams in worldwide computer-aided drug design competition

The VI-SEEM project “Predicting the structure of the FXR-ligand inhibitor complexes and affinities using computer-aided drug design within the D3R challenge” (Acronym: “D3R”) ranked in the top 6 teams (together with Eli Lily, Merck, University of Edinburgh, University of Strasburg, National Research Council Canada) in a worldwide computer-aided drug design competition (https://drugdesigndata.org/about/grand-challenge-2).

Computational tools, such as docking and scoring, for modeling the interactions of proteins with drug-like compounds (ligands) hold great promise to speed the discovery of new, safer medications and to reduce the cost of the drug discovery process. However, there is still a need for improved methods of predicting ligand poses and affinities or relative affinities.

The Drug Design Data Resource (D3R) initiative organizes blind challenges for computer-aided drug design (CADD), where participants are asked to predict how candidate drugs will bind to pharmaceutical drug targets such as proteins (unknown binding modes) or predict how well these candidate drugs will bind to these proteins (affinity predictions). The goal of D3R is to engage the CADD community through blind prediction challenges, training opportunities and hosting an annual workshop. The blind evaluation of computational drug design methodologies by using datasets of structure-related activity of co-crystalized inhibitors is of great importance. Such evaluations promote the application of computational methods without the use of any bias in their setup such that the results are a true indicator of their limitations and possible pitfalls in the actual drug design process. This can further indicate new areas of investigation for the improvement of computer-aided drug design methodologies.

In this drug design contest, named “D3R Grand Challenge 2”, Roche Pharmaceuticals donated a blinded unpublished dataset containing 36 high quality crystal structures of the Farnesoid X receptor (FXR) target and inhibition data (IC50s) for 102 compounds. About 50 teams from around the world participated in the contest aiming to predict the real crystallographic poses and the affinities of the compounds.

The D3R VI-SEEM project was awarded 25,500 core hours on the GPU island of the Greek National Supercomputer ARIS throught the 1st VI-SEEM call. The participating team included Christina Athanasiou, Dr. Sofia Vasilakaki and Dr. Zoe Cournia from the Biomedical Research Foundation of the Academy of Athens, and Dr. Dimitris Dellis from the Greek Research and Technology Network (GRNET).

On 27 March 2017 the results of the study were presented at a webinar hosted by Roche and the University of California San Diego, who organized the competition. All the presentations can be viewed here:
https://drugdesigndata.org//about/d3r-webinar-march-27-2017.

The true crystallographic binding poses of an inhibitor of the FXR protein, bound onto FXR is shown in orange and the predicted pose is shown in green. Their difference, which is very small, is expressed in Root Mean Squared Deviation (RMSD) in Angstrom. In ribbons, parts of the protein can be seen.

VI-SEEM DCH regional training, 6-8 February 2017, Alexandria, Egypt

As digitization of cultural heritage artifacts progresses by the museums of Europe, and access to their digital archives is provided to an ever-growing number of people from all around the globe, operating Digital Libraries and facilitating data-mining technologies for large repositories requires excessive amounts of computing power that only HPC can offer. Responding to this pressing need the VI-SEEM Digital Cultural Heritage regional training was organized on 6-8 February 2017 with the objective of introducing participants to the VI-SEEM project e-Infrastructure and services, with special focus on advanced techniques of 3D scanning and imaging for digital cultural heritage applications. The VI-SEEM project responds to the need for international collaborations and shared activities for the sustainable creation of multicultural inclusive environments and the untapped potential of digital technologies and ICT, which create the right conditions for developing and testing transregional strategies that enable the use of cross-media content and the creation of a polythematic agenda.

The event received strong interest from members of regional and local communities with more than 25 attendees. Participants were introduced to laser scanning and photogrammetric techniques for the documentation and 3D reconstruction of cultural heritage artefacts, and many more technical details of heritage sciences pipelines and workflows, including the use of HPC for parallel processing of big data in cultural heritage.

Τhe agenda covered the following topics:

Introduction to 3D Scanning Techniques and theoretical aspects
Hands-on session on 3D Artifact scanning
Scanning Data post processing using Scanner Software
Extra software tools to optimize scans for Real-time software and web applications
Hands-on Session on 3D photogrammetry photo-shooting
Hands-on: Accessing the BA-HPC and running applications on the BA-HPC

CyI provided software regarding the use of HPC for structure-from-motion photogrammetric techniques of 3D reconstruction and the CyI team presented the DCH training portal of VI-SEEM and offered training on the use of the CLOWDER data management system for cultural heritage and its advanced features, such as online viewers and OCR.

VI-SEEM national events

VI-SEEM national dissemination event in Tirana, Albania
On 13 April the Polytechnic University of Tirana, Faculty of Information Technology, hosted a national VI-SEEM event, to present the VI-SEEM project, local infrastructure and its usage procedures, as well as to present applications in meteorology, life sciences and cultural heritage, and training for concepts of parallel programming. The event was attended by more than 40 participants.

VI-SEEM national dissemination event in Žabljak, Montenegro
The VI-SEEM national dissemination event in Montenegro was organized on 27 February, during XXII International Scientific Information Technology Conference 2017. The conference covered topics and round table discussions on development trends in the field of information and communication technologies, including current activities in this field in Montenegro.
On 27 February, during the plenary part of the conference, the VI-SEEM project was presented, including information on the available computing and storage resources, research areas and their applications. Special focus was given to the Montenegrin contribution i.e. molecular dynamics simulations performed by the relevant team. PSOMI application was presented, achieved results and planned activities in future work. Moreover the importance of molecular dynamics simulations in drug design, life science and medicine was emphasized. The event was attended by 20 participants from three Montenegrin universities, Institute for marine biology, the first Montenegrin center of excellence and other participants from technical faculties from region. The possibility of including other researchers from Montenegro in the project next call was considered. The meeting agenda and presentations is available at: https://events.hpc.grnet.gr/event/43/

Upcoming events

VI-SEEM related scientific conferences:

VI-SEEM related events

About VI-SEEM

VI-SEEM is a three-year project that aims at creating a unique Virtual Research Environment in Southeast Europe and the Eastern Mediterranean, in order to facilitate regional interdisciplinary collaboration, with special focus on the scientific communities of Life Sciences, Climatology and Digital Cultural Heritage.
VI-SEEM unifies existing e-Infrastructures into an integrated platform to better utilize synergies, for an improved service provision that will leverage strengthen the research capacities of user communities, thus improving research productivity and competitiveness on the pan-European level.
The project kicked-off in October 2015 and the consortium consists of 16 partners: lead institutes from the SEEM region, specializing in provision of scientific computing and storage resources, and scientific user support.

VI-SEEM project receives funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 675121.

newsletter