Periodic Reporting for period 2 - PerMedCoE (HPC/Exascale Centre of Excellence in Personalised Medicine - PerMedCoE)
Reporting period: 2022-04-01 to 2023-12-31
The primary aim of the HPC/Exascale Centre of Excellence in Personalised Medicine (PerMedCoE) is to develop a robust and sustainable roadmap for adopting upcoming European HPC/Exascale technologies in this field. To do so, PerMedCoE is optimizing key software tools in computational biology to enhance their computational capabilities. Through this effort, the PerMedCoE initiative seeks to empower the research community to harness supercomputing resources effectively, enabling the transformation of complex biomedical data analysis into actionable models of cellular functions with medical significance. PerMedCoE roadmap is divided into four steps: 1) optimising core applications for mathematical modelling and cell-level simulations to the new pre-exascale platforms; 2) integrating personalised medicine into the new European HPC/Exascale ecosystem, by offering access to HPC/Exascale-adapted and optimised software; 3) running a comprehensive set of biomedical use cases; & 4) building the basis for the sustainability of PerMedCoE by coordinating personalised medicine initiatives and HPC communities, and reaching out to industrial and academic end-users, with use cases, training, expertise, and best practices. In this context, the combination of simulations and HPC computing is expected to have a great impact on society, health and business. In addition, the European healthcare systems are under great economic and social pressure, demanding better, faster and cheaper solutions. These challenges represent a unique opportunity for cellular simulations, which offer sophisticated tools to accelerate diagnostics and personalise treatments for patients. In doing so, they uphold the basic principle of personalised medicine: saving invaluable time and resources for healthcare systems and, more importantly, alleviating patient suffering.
PerMedCoE project has made significant progress in advancing the integration of high-performance computing (HPC) with personalized medicine tools. The PerMedCoE roadmap has involved the optimisation of key software tools for mathematical modelling and cellular simulations on current European computing platforms. To do so, PerMedCoE 's approach has involved scaling up core tools to achieve several objectives: i) enhancing their utilization of computational resources on individual machines; ii) integrating distributed computing protocols to enable concurrent simulations across different machines; and iii) optimizing algorithms to be GPU-compatible, and therefore, adaptable to future exascale technologies. These efforts were complemented by rigorous benchmarking activities on several European supercomputers, ensuring the reliability and scalability of the updated tools. Additionally, PerMedCoE aimed to integrate its infrastructure seamlessly into the broader European HPC/Exascale ecosystem, facilitating accessibility and usability for researchers across diverse domains of personalized medicine. A pivotal aspect of this integration was the development and dissemination of a comprehensive catalogue of building blocks integrating selected core applications with a set of third-party tools of great relevance in biomedical research. This approach allows biomedical researchers to easily combine HPC-ready biomedical tools and represent relevant biomedical problems as computational workflows. To illustrate this concept, the project implemented a set of practical use cases focusing on (i) improving cancer diagnosis, (ii) predicting drug synergies to design more effective treatments, (iii) studying tumour evolution and metastasis and (iv) characterizing COVID-19 infection in lung epithelial cells. All these developments have been shared via the project's website and official social media platforms. Both technical and scientific findings have been also presented at various international conferences and workshops. Moreover, PerMedCoE has dedicated significant efforts to crafting training programs, including courses tailored to introduce life science researchers to the realm of HPC. These initiatives encompass webinars, openly accessible study materials, and tutorials. Through these endeavours, PerMedCoE has fostered a dynamic community of users engaged in leveraging HPC for personalized medicine applications.
PerMedCoE has equipped the Personalized Medicine research community with a suite of readily deployable computational workflows, seamlessly integrating core tools into pre-exascale HPC environments to tackle real-world use cases on an unprecedented scale. These achievements stem from the enhanced computational capabilities of the selected core tools, facilitating the effortless creation of computational workflows capable of modelling intricate biological scenarios involving millions of cells. This approach has facilitated the development of significant use cases, yielding scientific insights crucial for understanding tumour evolution and COVID-19 infection dynamics. Moreover, PerMedCoE has established an effective and reproducible framework for scaling analytical tools in biomedicine that will guide future biomedical researchers in harnessing HPC technologies for their own investigations. These advancements are poised to profoundly impact the landscape of personalized medicine at different levels. First, the computational demands for realistically simulating biological systems, such as tissues and organs, are anticipated to surge exponentially in the coming years, demanding more powerful tools to process and leverage the upcoming increase of biomedical data. Second, cellular simulations will empower clinicians to assess thousands of potential treatments customized for individual patients in a virtual environment. This transformative approach will not only facilitate the creation of more effective therapies but also lead to substantial reductions in hospital care expenses within healthcare systems, thereby enabling public administrations to allocate their financial resources more effectively.