PerMedCoE is working towards the creation of digital twins that can virtually represent cell-level systems to computationally predict the response of potential treatments.
- These cell-level simulations can be personalised with patient-specific data to capture individuals´ genetic and environmental influences to provide a deeper understanding of the biological context of the diseases under study.
Under the coordination of the Barcelona Supercomputing Center, PerMedCoE (European Centre of Excellence for Personalised Medicine) exploits partners´ expertise to provide exascale-ready tools and methodologies that help to further understand cancer and potential appropriate treatment.
Alfonso Valencia, ICREA Professor, Life Science Director at the Barcelona Supercomputing Center and PerMedCoE coordinator, highlights: In its two years of existence the centre has made substantial progress in the optimisation of cell-level simulations and their integration in HPC environments. These tools and workflows, now openly available to the community, open new avenues for understanding cancer cell biology, which may help in the optimisation of personalised treatments in the long run.
Through these three cancer case studies, the PerMedCoe project uses real public data to combine core simulation tools (MaBoSS, CellN0pt, COBREXA, CARNIVAL, and PhysiCell) into computational workflows running in HPC facilities.
Drug synergies of cell lines in cancer treatments
To identify the consequences of different drug synergies used in cancer treatments, PerMedCoE HPC- ready cell-level simulation software models combined interactions between targeted cancer therapies. The project has proposed effective drug combinations for gastric, colon and prostate cancers using MaBoSS, CellNOpt and COBREXA to simulate thousands of cell lines and to enable the browsing and clustering of data so that patterns can be identified. Capturing patient heterogeneity allows for next-generation modelling to further personalise drug treatments. A workflow for the prediction of personalised targeted drug combinations, focusing on early cancer metastasis on colon cancer, was developed and successfully run on the Barcelona Supercomputing Center’s MareNostrum 4 supercomputer.
Cancer diagnosis based on omics information
In this use case, cell-level simulation tools are used to identify the different clinical course of individual patients based on molecular and clinical information. An ongoing analysis examines the case of 551 patients suffering from Chronic Lymphocytic Leukaemia at Hospital Clinic in Barcelona, Spain. The tools MaBoSS, COBREXA and PhysiBoSS produce metabolic models based on omics information and knowledge from the literature. They help to predict clinical outcomes and test different initial personal conditions and potential interventions.
Tumour evolution based on single-cell omics and imaging
PerMedCoE tools have been used to model tumour evolution by using HPC simulations to help answer questions such as ‘What would have happened if the tumour had not been removed from the patient?’ This approach deepens our understanding of tumour evolution and provides opportunities for finding personalised treatments in the case of a relapse. Unlike classical mathematical approaches, the use of HPC allows for the simulations of billions of cells incorporating genetic and environmental perturbations and paves the way to simulate real-sized tumours, a step towards testing drugs and treatments in digital twins (virtual models designed to reflect a physical object).
BSC leads the implementation and benchmark of the PerMedCoE use cases through a joint effort across the Life Science, Operations and Computer Science Departments. The use cases are critical for demonstrating the use of the project’s core applications, building blocks and workflows. Moreover, PerMedCoE’s developments are complemented by the community benchmarking efforts. BSC spearheads these efforts by bringing together an international community of tool developers and users to analyse and test similar external tools. In addition to fostering exchange between experts in this field, benchmarking encourages agreements on standards to ensure the interoperability of the tools used in the workflows so the right tool is found for the right problem.
About PerMedCoE
PerMedCoE is the HPC/Exascale Centre of Excellence for Personalised Medicine in Europe and aims to provide an efficient and sustainable entry point to the HPC/Exascale-upgraded methodology to translate omics information into actionable models of cellular functions of medical relevance. Coordinated by the Barcelona Supercomputing Center (BSC), this CoE has been awarded €5 million in funding from the European Commission and will run from 1 October 2020 to 30 September 2023. A range of 12 world-class academic and industry partners from across Europe participate in this CoE: Barcelona Supercomputing Center, CSC – IT Center for Science, University of Luxembourg, Institut Curie, University Hospital Heidelberg, Atos Spain, KTH Royal Institute of Technology, European Molecular Biology Laboratory-outstation European Bioinformatics Institute, Centre for Genomic Regulation, Max Delbrück Center for Molecular Medicine, the University of Ljubljana and ELEM Biotech.
- Caption: Training AI models with massive amounts of experimental and clinical data will revolutionise molecular biology and medicine, providing new insights and novel methodologies for improving disease prognosis and treatment. Credit: This composite image was generated using materials from Paul Macklin (digital twin image), the COVID19 Disease Map community (molecular pathway image), and the BSC.