Actualité - Innovation

PerMedCoE: a European center of excellence providing AI to assist with personalized medicine

Sabine D'Andrea
The Bioinformatics and Computational Systems Biology of Cancer team, led by Emmanuel Barillot, has just won large-scale European project funding as part of the H2020 program - Research Infrastructures.
Emmanuel Barillot

With close to 5 million euros over three years - €441,000 of which is just for the Institut Curie team.The PerMedCoE project aims to take full advantage of the power of supercomputers active today in Europe in order to optimize the analysis of biological and medical data from patients, and offer them a personalized course of treatment.

Despite complex IT simulations and considerable progress in the molecular characterization of tumors, personalized medicine still has a long way to go before it can truly provide an individualized treatment offer to each patient suffering from serious forms of cancer. Today, by precisely analyzing the tumor, physicians are able to identify a patient’s clinical profile, classify him/her within a population with the same molecular specificities (DNA mutations or their consequences) and prescribe a treatment adapted to his/her disease. This model, which distributes patients within groups, has helped offer patients better treatment by preventing rejection of certain treatments, but is nonetheless reaching its limits.

Scientists are now seeking to model all interactions between the various components of an organ, at the genomic, proteomic and metabolic level, in the hopes of being able to unscramble the complex data, understand the development of the disease and offer a specific individualized treatment. The approaches range from simulations of these mathematical models to artificial intelligence. The performance of current software programs is still not adequate for handling this information generated in volumes soon to reach one Exabyte (1,000 Petabytes, or a billion Gigabytes).

Europe therefore decided to launch the PerMedCoE program. PerMedCoE stands for HPC/Exascale Center oExcellence in Personalized Medicine. Its goal is to improve the performance of this diagnosis assistance software, bearing in mind that this data from these supercomputers - high-performing computers - must remain accessible to end users (physicians), within an environment that is suited to the specific and confidential aspects of personalized medicine.

The program aims to integrate personalized medicine into the new HPC/Exascale European ecosystem by developing the accessibility of its offer through the following:

1. Optimization of four mathematical applications for translating omics data (genomic, transcriptomic, proteomic and metabolomic) into usable medical actions.

2. The development of complete kits describing concrete scenarios of use in personalized medicine, and including:

  • tumor diagnosis
  • the revelation of medication interactions in some cancers
  • tumor prediction and development
  • within the context of rare diseases, the modeling of each patient’s individual profile

3. The coordination of the two expert communities, that of personalized medicine and that of artificial intelligence, but also the involvement of industrial players. Sharing training and best practices, in addition to concrete scenarios of use.

The Institut Curie team will provide its expertise in the analysis of molecular tumor data via high-performance calculation with Philippe Hupé, as well as mathematical modeling and simulation of tumor development, with Laurence Calzone and Andrei Zinovyev.

This PerMedCoE project is essential for furthering research in cancer and adapting treatment options,” explains Emmanuel Barillot, manager of the Bioinformatics and Computational Systems Biology of Cancer team“It will help to fully exploit the potential of omics tools, and bring this potential in line with imaging tools, which nowadays are offered to all patients. In what we hope is the very near future, the PerMedCoE center of excellence will allow the physician to fine-tune the diagnosis and the therapeutic approach by incorporating all data gathered on the patient in question.


Pour aller plus loin