The Limited Times

Now you can see non-English news...

When AI is used to create new drugs

2021-11-10T17:02:02.948Z


It will do so thanks to AlphaFold2, a software "agent" by Isomorphic Labs, a new startup of Alphabet born from an offshoot of DeepMind (ANSA)


By Alessio Jacona *

First chess, then the game of Go and now the drugs.

The race of DeepMind, a company specialized in the development of algorithms founded by Demis Hassabis and acquired in 2014 by Google (now Alphabet), now seems unstoppable: its artificial intelligence technologies will in fact also power a new "sister" startup, called Isomorphic Labs, which was born with the aim of using AI to develop new medical treatments and new drugs (also against Covid-19).

It will do this using AlphaFold2, a more advanced version of a software “agent” that makes predictions on the structure of proteins: based on a deep learning system, it is able to predict the 3D structure of a protein directly from its amino acid sequence.

The technology had already made headlines for the first time at the end of 2020, when the first AlphaFold version had won the international Casp (Critical Assessment of Protein Structure Prediction) competition, proving that it could predict the structure of proteins based on the sequence of amino acids that compose them.

It is then in the middle of this year that AlphaFold returns to the headlines for having managed to predict the 3D structure of the 20,000 proteins expressed by the human genome, as well as those of 20 other organisms (from the mouse to the malaria parasite) crucial for the research. . Another important success whose results, published in the journal Nature, have also been collected in a database that - thanks to a collaboration with the European Molecular Biology Laboratory (EMBL) - has been made freely accessible to scientists all over the world.

To give an idea of ​​the step forward, just think that usually the study of the structure of proteins takes place with techniques such as X-ray crystallography and cryelectronic microscopy, it takes a very long time (even decades) and often fails. . Which also explains why the structures of just 170,000 out of over 200 million known proteins have so far been resolved.

In short, it seems like yesterday when in 2016 the AlphaGo software ("ancestor" of that AlphaFold2 on which the activities of the new startup Isomorphic Labs will be based) literally beat Lee Sedol, at the time one of the strongest players in the world in the game of Go, prompting many to question the very future of human intelligence. Today that dismay is overcome, replaced by the awareness that AI will not serve to replace human beings, but to "increase" their abilities, leading them to new scientific successes.

"We believe that using computational and artificial intelligence methods we can help scientists take their work to the next level, accelerating the process of drug discovery," said Demis Hassabis, who in addition to DeepMind will now also lead Isomorphic Labs, and which will have to contend with the competition of over 160 companies engaged in the same sector, including the Canadian Atomwise, the South Korean Standigm or the German Genome Biologics.

According to a recent study by the US National Center for Biotechnology Information, the price to create a new drug is on average $ 1.3 billion.

As the report explains, researchers currently physically synthesize each individual component and then test it in the laboratory under conditions similar to humans.

Using AI would allow for extraordinarily faster and safer testing as well as much less expensive.

---

* Journalist, innovation expert and curator of the Artificial Intelligence Observatory ANSA.it

Source: ansa

All news articles on 2021-11-10

You may like

News/Politics 2024-03-22T23:04:48.727Z

Trends 24h

News/Politics 2024-03-27T16:45:54.081Z
News/Politics 2024-03-28T06:04:53.137Z

Latest

© Communities 2019 - Privacy

The information on this site is from external sources that are not under our control.
The inclusion of any links does not necessarily imply a recommendation or endorse the views expressed within them.