1er May, the Joint European Disruption Initiative (JEDI) will launch the first phase of a major new challenge in Europe, dedicated to the search for molecules to fight against Covid-19. Called ” Billion molecules against Covid-19 “(One billion molecules against Covid-19), the initiative will reward a team of researchers with an amount of 2 million euros that will have achieved an ambitious objective: to locate in a library of more than a billion molecules those capable of blocking the mechanisms of the SARS-CoV-2 virus and preventing it from entering human cells. Participants in this challenge will be able to rely on strategic partners who will provide access to equipment in the field of high-frequency computing.
” This challenge is a way to think outside the box to find a fast path to therapeutic treatment “Explains André Loesekrug-Pietri, executive director of the foundation, which claims to be the European equivalent of the Defense Advanced Research Project Agency (DARPA) in the United States. ” It is open, and therefore does not compete with current projects on the search for a treatment ” The initiative, which is the first challenge officially launched by JEDI, is financially supported by the AXA research fund and the pharmaceutical group Merck.
A multidisciplinary approach
To accelerate the search for drugs against the epidemic, JEDI is using a multidisciplinary approach, open to laboratories, companies and start-ups. ” We set the objectives to be achieved. The means are free “, Supports André Loesekrug-Pietri,” We want to be surprised by the solutions provided ” A total of around 100 teams, bringing together skills in dynamic modeling, algorithms, molecular biology and virology, are expected to kick off in the first phase.
This will take place over four weeks and will consist of molecular screening, carried out on a “scale never before known” according to the JEDI. This “in-silico” work promises to be dizzying: the teams will have to use 3 different calculation methods, left to the researchers’ discretion, in order to sift through the library of a billion molecules.
They will then have to establish a list, consensual to the 3 methods deployed, of 10,000 molecules classified by affinities on at least 3 targets of the SARS-CoV-2 virus (such as for example the famous Spike proteins, whose arrangement in a crown is specific to the the coronavirus family), among the twenty already identified. The desired affinity must be less than 100 nanomoles. ” It’s a very strong level of affinity with the molecule “, Emphasizes André Loesekrug-Pietri.
Between 10 and 50 petaflops of available power
Strategic partners will provide the basic tools: Merck will provide participants with the molecule library of its subsidiary Sigma Aldrich, the American project Folding @ home, specializing in the simulation of protein dynamics, will bring its catalog of targets, while the high performance computing specialists Atos and the French organization GENCI (Large national equipment for intensive computing) will provide access to their supercomputers, including the supercomputer Jean Zay, already mobilized in the fight against the Covid-19. Deutsche Telekom will also make available CPU and GPU power to support this project.
” A recently published article in the journal Nature believed that the equivalent 160,000 cores to test tens of thousands of molecules in one day “, Point André Loesekrug-Pietri” We wanted to secure colossal computing power for this project ” A ticket system will be set up to allow teams to benefit from calculation time. JEDI Billion Molecules GrandChallenge aims at least between 10 and 50 petaflops of power available for the participating teams.
” In molecular screening, it is often the brute force technique that is used. It is in this area that multidisciplinarity will prove to be interesting, because it will make it possible to develop algorithms that will be able to increase by 50, even 100, the computing capacity. It’s very close to what’s done in climate models “, Continues André Loesekrug-Pietri.
A prestigious scientific committee
A scientific committee, also multidisciplinary and composed of big names in science, such as Sir Peter Ratcliffe, 2019 Nobel Prize winner in medicine, Bernhard Schölkopf of the Max Planck Institute for Intelligent Systems, Marion Guillou former CEO of INRA, Stephane Requena director GENCI technical team or Daniel Verwaerde, former CEA general administrator, will be responsible for evaluating the submissions and establishing an “ultimate” list of molecules. It will be the starting point for the second phase of the challenge. The team whose list of molecules has the highest affinity for the virus will receive a prize of 250,000 euros.
Reduce viral load by 99%
This will also last 4 weeks and should end at the end of June, will consist of an in-vitro evaluation of the molecules to allow a reduction of the viral load by 99%. The teams may be the same as in the first phase or joined by new ones. Again, an envelope of 250,000 euros will be awarded to the team with the best results.
If the first two phases are sequential, phase 3 has an additional objective: it consists in finding cocktails of molecules “on the shelf”, approved by the American health agency FDA and which can quickly be evaluated clinically, which shows a level of very high interaction with the virus. A true grail, the team that succeeds in identifying this molecule will receive a “special price” of 1.5 million euros. This third phase can be activated at any time and should last for at least 12 weeks. However, its industrialization could take longer.