The artificial intelligence of Google into one of the riddles of life

molecular biology to understand how a protein will fold in on itself is a leap that can be compared to that of the first man on the Moon. The final form that it

02 January 2021 Saturday 04:34
1091 Reads
The artificial intelligence of Google into one of the riddles of life

molecular biology to understand how a protein will fold in on itself is a leap that can be compared to that of the first man on the Moon. The final form that it assumes, in fact, affects the properties of the protein under the aspect of physiological and pathological In 1969, the same year Apollo 11, a molecular biologist from the american said that would not have been sufficient for the duration of the universe to find the correct one among all the possible combinations in which a protein can fold. "We thought that it would have taken decades to arrive at this result," said Venki Ramakrishnan, 2009 Nobel for chemistry.

3D

Thanks to the artificial intelligence (AI), instead, it is already today possible to model in 3D with an accuracy of more than 90% of the folding of a protein from a sequence of one dimensional array of amino acids, the "building blocks" that constitute it. This will facilitate the creation of new drugs and the understanding of details that still elude us in the biology and in the structure of the essential components for life such as proteins. "Over the years we have developed various methods of bioinformatics to predict as is the bending, but had large margins of error," explains Daniela Corda, director of the Department of biomedical Sciences of Cnr. "This new medium is important for the possibility to accelerate the development of drugs, bypassing a phase long and laborious, which so far passed through the purification and crystallization of the protein to study."

DeepMind

DeepMind, the london society tied to Google via the holding company, Alphabet, has announced to be coming to a head the dilemma that lasted for more than 50 years, thanks to AlphaFold, his AI system that has participated to a competition specification, Casp (Critical Assessment of Structure Prediction), in which various computational models from a quarter of a century, confront and challenge each other to be able to solve the mystery of how to fold proteins. But why is it so important to understand the way in which you fold the proteins? "The Dna that contains the genome of all living beings to be able to work, it must be translated into long strands of amino acids which then form proteins, which become active fold more times in the special structures of various shapes," continues the String. "By the way of bending depends on both the interaction with other proteins and the formation of molecular complexes, which are essential for the creation of cellular structures and complex".

System neural network

AlphaFold is not a regular computer program, but is structured with the connections that simulate the human brain. "In order to arrive at the remarkable result obtained with a margin of error infinitesimal, equal to the diameter of an atom," explains Alessio Bechini, professor of Bioinformatics in the course of biomedical Engineering at the University of Pisa, "has been instrumental in the large increase of computing power that has been achieved in recent years. A tool such as AlphaFold requires huge computing power, unimaginable only ten years ago."

A weapon

it Is still uncertain if DeepMind will share the technology with the scientific community. Demis Hassabis, co-founder and chief executive of the company, has reported that it is planning the publication of the details, but not before 2021. Maybe you will not make it in time to use this new weapon to fight the Sars-CoV-2, but it is certainly a tool that will help you to combat with rapid and effective responses to possible future pandemics that may surprise as the Covid-19. "When will it be available I will use it without doubt," concludes the researcher at the Cnr. "It is a huge step forward to find new drugs."

Updated: 02.01.2021 04:34