An artificial intelligence system developed by DeepMind (owned by Google) has made a giant leap in solving one of the greatest mysteries of biology, mapping the three-dimensional structure of proteins. The AlphaFold program defeated the “competitors” in the international scientific competition CASP (Critical Assessment of protein Structure Prediction) and managed to outline the structure of a protein with a unique guide to its amino acid sequence.
AlphaFold correctly predicted two-thirds of the protein structure. Within days, the algorithm made progress that required scientists years of costly research in their laboratories. “I was really shocked when I saw it. “I never expected to experience that,” said University of Maryland structural biologist Dr. John Mult, one of the organizers of the international competition that started in 1994 and takes place every two years. The ability to predict the structure of a protein would give unprecedented impetus to the biological sciences and the science of medicine, accelerating efforts to understand the structural elements of cells and allowing the rapid development of advanced drugs. “Data changes forever,” said Andrei Loupas, an evolutionary biologist at the Max Planck Institute for Developmental Biology.
The AlphaFold program has already helped Lupas find the structure of a protein that has plagued his lab for a decade. “It simply came to our notice then. It will change the research. It will change bio-engineering. It will change everything. ” The “structural problem” Proteins are the building blocks of life, responsible for almost everything that happens inside cells. The function of a protein is determined by its three-dimensional structure – “structure is function” is, after all, one of the axioms of molecular biology.
Proteins adopt a shape without external influences, guided by the laws of physics. Tens of thousands of different proteins play a vital role in the health of organisms and diseases – for example the coronavirus Sars Cov 2, which causes Covid-19 disease, manages to penetrate human cells thanks to a protruding spike – a protein that structurally “unlocks” receptor protein in human cells. Unlike the shape that is difficult to “guess”, biologists can easily analyze the sequence of amino acids that determine the structure of proteins.
Lupas predicts that the real revolution will be achieved when scientists can now only use computers to predict how proteins interact with other molecules. For decades, laboratory experiments aimed at mapping proteins, often with zero results. Since the 1950s, scientists have used X-rays to target crystalline proteins. From the diffraction of light, they were able to guess the “atomic coordinates” and X-ray crystallography has to date had the lion’s share in the sketching of proteins, but without particularly satisfactory results. Apart from being inaccurate, these experiments are very time consuming and costly.
AlphaFold prediction was in some cases identical to those of successful crystallographic methods and cryo-EM (cryo-EM) approaches. Scientists hope that Artificial Intelligence now makes it possible to study life in completely groundbreaking ways. And they are not wrong since the AlphaFold algorithm has already correctly predicted the shapes of several SARS-CoV-2 coronavirus proteins.
In the future it is likely to predict which of the thousands of drugs bind properly to critical proteins and can work therapeutically. DeepMind has promised to ensure broad access for scientists to its “rescue” programs.
NOTE-COMMENT: Of course this technology could be used for other purposes that the human mind does not think about ………..
