One of the greatest biological mysteries “mostly solved” with the help of artificial intelligence



[ad_1]

  • Helen Briggs
  • BBC Science Correspondent

Caption below photo,

Deepmind model of Legionnaires protein (Kasp-14)

One of the biggest mysteries in biology has been largely solved with the help of artificial intelligence.

Predicting how proteins fold into unique three-dimensional shapes has angered scientists for half a century.

Deep Mind, a London laboratory that uses artificial intelligence, has largely solved this problem, say the organizers of an exceptional scientific competition.

A better understanding of the form of protein may play a key role in the development of new drugs for treatment of various diseases.

The discovery of Deepmind is expected to accelerate the study of a large number of diseases, including Cowid-19.

Their program determines the form of protein at an accuracy level that can be compared to expensive and time-consuming laboratory methods.

Dr. Andrei Kristafovic, of the University of California, one of the councils of scientific arbitrators, described the achievement as “truly exceptional.”

“Being able to research protein forms quickly and accurately has the potential to revolutionize scientific life,” he says.

What are proteins?

Proteins are present in all living things, in which they play a key role in chemical processes that are essential for life.

Made from a series of amino acids, they bend in countless possible ways into complex shapes that carry with them the key to the way they perform vital functions.

Caption below photo,

Receptor-bound protein displayed on a computer.

Many diseases are associated with the role of proteins in catalysis of chemical reactions (enzymes), fight against disease (antibodies) or the role of chemical messengers (hormones such as insulin).

“Even tiny rearrangements of these vital molecules can have catastrophic consequences for our health, so one of the most effective ways to understand a disease and find new ways to treat it is to study the proteins involved,” says Dr. John Molt of University Of Maryland, USA, Chair of the Scientific Arbitrators Council.

“There are tens of thousands of human proteins and many billions of them in other species, including bacteria and viruses, but the switch form in just one requires expensive equipment and can take years.”

How does the competition work?

Christian Anfinsen received the Nobel Prize in 1972 for his work, which showed that it is possible to determine the form of a protein based on the sequence of amino acids in which it is composed.

Every two years, a large number of teams from more than 20 countries use computers to blindly try to predict the shapes of sets of about 100 proteins based on their amino acid sequences.

Signature under the video,

Male nipples and other parts of the body that we do not need

At the same time, biologists plant 3D structures in the laboratory using traditional techniques like x-crystallography and NMR spectroscopy, which determine the location of each atom in each other in their protein molecule.

A team of scientists from Caspa (a community-based experiment to critically evaluate techniques for predicting protein structure) compares these predictions with 3D structures using experimental methods.

Casp uses a measurement method, known as the Global Distance test, to estimate accuracy from 0 to 100. The result of about 90, which was achieved by the Defind Alphafold program, is considered as comparable to laboratory techniques.

What happened this year?

In the last round of the competition, CASP-14, AlfaFold determined the shape of about two-thirds of the protein with an accuracy compared to laboratory experiments.

The arbitrators have said that the shape accuracy of most other proteins is too high, though not quite at this level.

Signature under the video,

Small organisms, insects and fungi live in our households together with us.

AlfaFold is based on a concept called deep learning.

In this process, the structure of folded proteins is represented by a spatial graph.

The program “Learn” using information about 3D forms of known proteins stored in the Public Protein Database.

The artificial intelligence program managed to do in just a few days what it takes years in the laboratory.

How will the information be used?

Knowing the 3D structure of proteins is important to make medicines and understand human diseases, including cancer, dementia and infectious diseases.

Caption below photo,

The virus that causes COSID-19 has a recognizable spiky protein (marked in red)

One example is Cowid-19, where scientists have studied how a spiky protein on the surface of the SARS-Cow-2 virus interacts with receptors in human cells.

Professor Andrew Martin of University College London (UCL), a former Caps participant and current arbitrator, told the BBC. That “understanding how a protein sequence falls into three dimensions is indeed one of the major questions of biology.”

“The whole way a protein works depends on its three-dimensional structure, and the functioning of a protein is important for everything related to health and disease.”

“By knowing the three-dimensional structure of proteins, we can help to make drugs and intervene in health problems, whether it be an infection or a hereditary disease.”

Professor Dame Janet Thornton of EMBL’s European Institute of Bioinformatics in Hinkston, UK, said that how proteins bend to create “completely unique three-dimensional forms” is one of the greatest mysteries in biology.

“A better understanding of the structure of proteins and the possibility of predictions with computers means a better understanding of life and, of course, human health and disease,” she explains.

Signature under the video,

Virus mutations are common during a pandemic.

What will happen next?

Other scientists will also study the data to determine how accurate this method of artificial intelligence is and how well it works on the level of detail.

There is still a large gap in knowledge, including the understanding of how multiple proteins fit into each other and how proteins interact with other molecules such as DNA and RNA.

“Now that the problem is largely solved for individual proteins, the way is open for the development of new methods for determining the shape of protein complexes – a collection of proteins that work together to form most of the live machinery and other applications,” says Dr. Kristafovic.

Follow us on Facebook I Twitter. If you have a suggested topic for us, please contact [email protected]



[ad_2]
Source link