Artificial intelligence, this is how Alphabet (Google) will find new drugs
Will artificial intelligence replace humans in the discovery of new drugs? Alphabet tries with Isomorphic Labs. All the details
Alphabet, the company Google's parent company, announced that it will use artificial intelligence to develop tools that can identify new drugs and to get it launched a new company , the Isomorphic Labs.
THE LINK WITH DEEPMIND
Isomorphic Labs will use the computer platforms of DeepMind , another subsidiary of Alphabet acquired by Google in 2014, which has done groundbreaking work using artificial intelligence to predict the structure of proteins – the AlphaFold2 program.
DeepMind CEO Demis Hassabis will also be CEO of Isomorphic Labs, but the two companies, a spokesperson said, will remain separate and collaborate on occasion.
WHAT WILL ALPHABET'S ISOMORPHIC LABS DO?
Building on the work initiated by DeepMind, the new company will build tools that can help identify new drugs, but also models that can predict how drugs will interact with the body.
Isomorphic Labs could leverage DeepMind's successes in protein structure to understand how multiple proteins interact with each other.
The company, according to a spokesperson to The Verge , may not develop its own drugs, and instead sell its models, focusing on developing partnerships with pharmaceutical companies.
POSSIBLE LIMITS AND DIFFICULTY
Despite the positive premises, the road is all uphill. In fact, developing and testing drugs is an even more ambitious challenge than understanding the structure of proteins.
"Over 90% of drugs that make it to a clinical trial end up failing," chemist Derek Lowe told Science . And most of the problems aren't due to something wrong at the molecular level.
Artificial intelligence can certainly be of support to humans but as Helen Walden, professor of structural biology at the University of Glasgow pointed out in an interview with The Verge , it cannot be a quick solution to the countless challenges of drug development. .
Several large companies, Techcrunch notes, have been formed and funded with hundreds of millions of dollars to pursue very similar goals over the past five years, but there has been no visible revolution or a famous wonder drug discovered by artificial intelligence for a disease. previously incurable.
HOW THEN CAN AI HELP SCIENCE?
For years, The Verge writes, experts have argued that artificial intelligence can make finding new drugs faster and cheaper.
As proof of this, according to the latest study by the US National Center for Biotechnology Information, cited by Ansa , the price to create a new drug is, on average, 1.3 billion dollars. As the report explains, researchers currently physically synthesize each individual component and then test it in the laboratory under conditions similar to humans.
However, AI could, for example, help scan databases of potential molecules to find some that best suit a particular biological target or fine-tune compounds.
"Biology is probably too complex and messy to be encapsulated in a simple set of ordered mathematical equations," said Hassabis. "But just as math has turned out to be the right descriptive language for physics, biology may turn out to be the perfect kind of language for applying AI."
This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/innovazione/intelligenza-artificiale-ecco-come-alphabet-google-scovera-nuovi-farmaci/ on Sun, 07 Nov 2021 06:35:22 +0000.