Understanding how proteins essential for life are made has been one of the "big challenges" in biology, and scientists have spent decades trying to understand how they are made.

Starting today, determining the 3D shape of any protein known to science will be simple, as a new tool has identified the structures of about 200 million proteins, after the company "DeepMind" revealed a revolutionary artificial intelligence network to predict the structures of 3D proteins, covering almost all proteins. known to all living organisms, and this will enable scientists to have immediate access to in-depth information about the basic building blocks of life.

The ability to understand the structure of a protein allows determining how it works and how effective drugs are when interacting with it (DeepMind)

Alphafold and DeepMind

According to the British newspaper “Daily Mail”, before the artificial intelligence program known as “Alpha Fold”, scientists used to spend months or years understanding the structure of proteins, and researchers often used tools such as X-rays, but the “Alpha Fold” program Developed by Google's "Deep Mind" company, it is capable of performing deep learning with the aim of predicting the structure of proteins.

The first version of this program was published in 2018, and the second version of it was published in late 2020, and it is available with open source software to search within databases on the protein complex (proteome) of species and organisms.

More than 500,000 researchers in the world from 190 countries have used the “Alphafold” database to display more than two million structures of proteins, and this complex information is now available at the same speed as the search on Google, and the program now predicts the structure of almost all proteins known to science, whether in animals or plants. Or humans, bacteria, or other organisms.

Thanks to the Alphafold tool, researchers can access a much larger number of protein structures (Alphafold).

Deep Mind is a British artificial intelligence company, founded in 2010, and renamed after it was acquired by Google in 2014. The company has created “neural network software” that can learn how to play video games in a similar way to a human, and a neural network It may be able to access an external memory, and so the computer is able to simulate the short-term memory of the human brain.

Important resource for scientists

The ability to quickly see the structure of a protein in three dimensions is valuable to scientists seeking to treat diseases and researchers who want immediate access to in-depth information about the building blocks of life.

Since its launch in 2020, researchers have already used Alphafold to understand the proteins that affect honeybee health, and to develop an effective malaria vaccine.

An expanded database could serve as an important resource for scientists to better understand diseases, and could also accelerate innovation in drug discovery and biology.

The latest update includes the predicted structures of animals, plants, bacteria, fungi, etc. (DeepMind)

Demis Hassabis, DeepMind's founder and CEO, says the database allows researchers to search for 3D structures of proteins "as easily as doing a Google search by keywords."

In an article on the company's website, he explained that the latest version of the data gives a strong push to the database, and the update includes structures for "plants, bacteria, animals and many other living organisms", and this opens up huge opportunities for the "Alphafold" program to influence important issues such as "sustainability, fuel, food insecurity and neglected diseases.

Just a starting point

“Alphafold is perhaps the largest contribution from the AI ​​community to society,” Jian Ping, a professor of computer science at the University of Illinois Urbana-Champaign who specializes in computational biology, told MIT Technology Review. Scientific".

"It could also help scientists reassess previous research to better understand how diseases occur," he added.

Predicting the structures of proteins takes a long time, and having a tool with 200 million protein structures readily available, Muhammad al-Quraishi, a systems biologist at Columbia University who is not involved in the DeepMind research, told MIT Technology Review. It would save researchers a lot of time.” However, for many proteins, "we are interested in understanding how their structure is altered by mutations and natural allelic variation, and it will not be going through this database."

Others use protein structure predictions to develop vaccines and investigate basic biology questions, such as examining the evolution of proteins when life first evolved.

However, in a Science article, the researchers caution that launching the database is just a starting point, "and it's clear that there is still a lot of biology, and a lot of chemistry, to be done."