CERN and the LHC experiments’ computing resources in the global research effort against COVID-19

CERN is the hub of vast global computing resources and collaborations, representing a considerable potential in the fight against COVID-19, with applications ranging from the support of therapy and vaccine research to the deployment of the data-sharing platform Zenodo, and from online educational platform tools to epidemic modelling.

As a first response to the pandemic, the particle physics community mobilised its number-crunching capabilities, allocating processors from the data centres of CERN, the LHC experiments and the Worldwide LHC Computing Gridto support volunteer computing initiatives such as Rosetta@home and Folding@home, which simulate protein dynamics to help understand the SARS-CoV-2 virus. With the pandemic worsening, the number of voluntary contributions to the distributed computing project rapidly increased and the joint processing capacity of the initiatives grew to exceed several exaFLOPS: a world first. In such a supportive environment, CERN and the particle physics community realised that the time had come to make a gradual transition from dedicating computer cores to such initiatives to contributing to them in more specific ways, through data management and data analytics expertise, as well as through software resources.

Open source technologies, which are key to particle physics data management, are also particularly well suited for optimising the transfer and management of COVID-19 research data. Examples include FTS, the File Transfer System developed by CERN, and Rucio, the scientific data management system initially developed by and for the ATLAS experiment and now adopted by many scientific communities. Developments to adapt them to the needs of Folding@home have been successfully completed. Folding@home, in a recently published paper, thanked CERN and the particle physics community for this collaboration on data management.

Over recent months, Zenodo, the open-data repository, developed by CERN with co-funding from the European Commission and available to all sciences in support of global Open Science, has been extended with additional storage and a dedicated community for research against COVID-19. As of today, popular data analysis execution frameworks like binder reference almost a thousand different data flows based on data stored in Zenodo. Examples of research projects and datasets published in Zenodo range from medical research, including lung infection data, to virology, including raw diffraction data of the SARS-CoV-2 structure, as well as economic and demographic impact assessment studies of the COVID-19 pandemic. 

Other ways in which CERN computing resources are being deployed in the context of the pandemic include distance learning in the Open Up2U initiative, coordinated by the GÉANT partnership of European national research and education networking organisations. CERN is contributing through two services based on its SWAN, CERNBox and EOS technologies. Institutions from 21 countries   have expressed interest in using these services for the upcoming school year.

CERN is also involved via CERN openlab, a public-private partnership, in various projects that are adapting their findings and models to support the fight against COVID-19. Through Circular Health, an international initiative started by the One Health Center of Excellence at the University of Florida, CERN is collaborating with the Centre for Research on Health and Social Care Management (CERGAS) at Università Bocconi in Italy to publish open access datasets in Zenodo and analyse epidemiologic, demographic and environmental data to look for possible correlations between the propagation and effects of COVID-19 and factors such as age, gender, comorbidities and air pollution.

CERN openlab, University College London (UCL) and the Italian Institute of Technology (IIT) are collaborating on one of the current projects of CompBioMed, the European Commission H2020-funded centre of excellence focusing on the use and development of computational methods for biomedical applications. The project aims to accelerate the development of antiviral drugs by modelling proteins that play a critical role in the virus's life cycle. Such simulations of proteins of the COVID-19 virus provide information that is essential to the drug discovery process. The researchers are investigating the use of deep neural networks and plan to apply the findings of recent work done at CERN to deep generative models (Generative-Adversarial Networks, or GANs, used for fast simulation of radiation transport in calorimeters).

BioDynaMo, a biology development simulation framework based on an innovative agent-based engine, which benefited from funds from the CERN Medical Applications budget and was developed by researchers in CERN openlab, the University of Newcastle and other institutes, was initially created to model the growth of cellular structures under specified environmental conditions, for example to understand tumour growth or perform in-silico treatment trials. However, it also proved to be very efficient in modelling other types of dynamic systems, such as virus propagation. The BioDynaMo team in CERN openlab has teamed up with the Global Health Institute at the University of Geneva to implement a COVID-19 localised spreading model using the BioDynaMo modeller. A grant to develop the model further has recently been awarded to the team by the European Open Science Cloud (EOSC).

Thanks to support from the CERN Medical Applications budget and initial financing, in addition to BioDynaMo, two other important initiatives have been developed to the wider benefit of society: the CAFEIN and MARCHESE research projects. The former, the Computer-Aided deFEcts detection, IdeNtification and classificatioN (CAFEIN) research project, started with the funding of a PhD student. Its initial goal was to support clinicians in prognosis, prevention and the proposal of personalised therapies in relation to brain lesions, providing a patient-centred medical diagnosis and treatment programme. With the development of the pandemic, the decision was taken also to use CAFEIN to help distinguish COVID-19 pneumonia from other types of viral and bacterial pneumonia.

To boost the project, the CERN Engineering department, together with the CERN Knowledge Transfer group, submitted a funding request under Horizon 2020, the EU Framework Programme for Research and Innovation, for the “Digital Transformation in Health and Care” call in June. If the project is selected, fully anonymised data and images will be collected from the National and Kapodistrian University of Athens and Masaryk University in Brno, the Association pour la Recherche et le Développement des Méthodes et Processus Industriels in Paris, and the Politecnico di Milano, together with the start-up ARAMIS, which would be involved in testing and validating the robustness and adaptability of the platform.

Through this research project, data science and artificial intelligence would support doctors in their decision-making regarding the best therapeutic procedure to follow for both brain lesions and COVID-19 treatment. The versatility of CAFEIN could allow it to be used to help treat other diseases in the future.

MARCHESE, the Machine-learning-based human ReCognition and HEalth monitoring SystEm, is the third research project to benefit from initial funding from the CERN Medical Applications budget. MARCHESE is now at the heart of a collaboration between the digital Experimental Cancer Medicine Team (digitalECMT) at the University of Manchester and the Survey Mechatronics and Measurements (SMM) group in the Engineering department at CERN, which has been collaborating on applying CERN’s know-how in the domain of machine learning for robotics to help tackle cancer. In the context of the COVID-19 pandemic, digitalECMT repurposed its systems and started a collaboration with the University Hospital in Southampton to assist their teams in collating and analysing key clinical, laboratory and scientific data.

The SMM group will be involved in the development of artificial intelligence predictive algorithms, which will integrate and work in tandem with real-time analytics for clinical trials to identify patients at risk and effective new treatments, in order to improve clinical outcomes. A grant proposal has been submitted to the Innovation Funding Service of UKRI (UK Research and Innovation).

In all of these projects, CERN is in close contact with the medical community through, for example, the Organization’s collaboration agreement with the World Health Organization.

This article was first published on 31 July by CERN.