AlphaFold or how Google preserves its health

“Enthusiasm is the basis of all progress”, defended Henry Ford. So, let’s get excited for the following announcement at the end of July from DeepMind, a 100% subsidiary of Alphabet-Google: its artificial intelligence has gone much faster than expected to model in 3D all of the 200 million proteins that exist. , and AlphaFold, the database published with EMBL’s European Bioinformatics Institute (EMBL-EBI), is already used by more than 500,000 researchers in 190 countries.

As the researchers themselves tell very well, on a storytelling blog set up by Google, this technology makes it possible to accelerate the fight against antibiotic resistance, the fight against plastic pollution, to protect bees, to to focus on the creation of the world or to overcome orphan or tropical diseases, too often deprived of research budgets.

Where researchers spent 10 years trying to reconstruct the form and interest of such and such a protein in such a context, with marketing success rates regularly below 10% and for 2.5 billion dollars, the machine only takes a few minutes. Very few voices are raised to criticize the Google project, apart from the very relayed Derek Lowe, for whom it is only one more important step, but not the panacea. The daring defender especially forgets to specify that he is the director of the biomedical research institute of Novartis, which leads a competing project with IBM Watson Health…

Free, a way to “kill” the competition

And to say that is to recall that beyond the collective astonishment, Google is not a philanthropic enterprise, even one endowed with tens of billions of dollars that it could spend as lost funds to advance research in offering its database to the first comer.

Faced with Amazon Neptune, a kind of Google Search for scientists who want to compile publications on a subject, or Microsoft, launched in another variant, AlphaFold is in direct competition with private platforms of software solutions for the development of new drugs. which are already generating turnover, such as Schrodinger, Simulation Plus (6.6 million euros out of 10 last year) or BenevolentAI (5.5 million euros). From Roche to Bristol Myers Squibb, via Bayer, Sanofi or Merck, everyone is on the same quest for the drugs of tomorrow.

Google is further. In mid-May, she entrusted the CEO of DeepMind, Demis Hassabis, with managing her new nugget, Isomorphic Labs, born last November to discover the drugs of tomorrow. This new layer is added to Verily, which promises precision medicine. Its partners include GSK, L’Oréal and Procter & Gamble. But it’s best to stop for a minute on three of these technologies: the Immune Profiler, to analyze the human genome and discover new genetic therapies; the Virtual Stainer, for much more precise biopsies than those passed under current microscopes, and Terra, with MIT, Harvard and Microsoft, a cloud-based collaboration platform to strengthen data analysis.

The cloud, the true Holy Grail of the group

Verily’s joint ventures with Primasun focus on sleep and its importance for good health, or with OneFifteen. Its other subsidiary, Calico Labs, focuses on diseases related to age and aging, including cancer.

For the experts, it is sure, 10 years after having abandoned the conquest of connected objects and four years after the troubles linked to the partnership with Ascension, one of the largest health players in the United States, which operates 2,600 sites of care, including 150 hospitals and 50 homes for the elderly, Google has “ONE” objective: to bring all these beautiful people to the cloud. To his cloud. The most cost-effective way to take advantage of all this data and all these services in an environment that is certainly competitive, but where not everyone can offer so much.

Leave a Comment