The Artificial intelligence has proven to be an unmatched ally for the medicine. So much so that its applications reach the diagnosishe treatment and the follow-up of many diseases. A new study of the University of North Carolina School of Medicine has also shown promising signs for the use of this technology in obtaining new drugs.
Not in vain, with AI it is possible to analyze medical images and optimize the execution of clinical trials. This greatly facilitates drug discovery, as explained by this entity. Specifically, it is a study by the scientist Bryan Rothprofessor of pharmacology at the entity.
Specifically, the AlphaFold2 AI system has been studied. As reported, this system predict protein structures with artificial intelligence. This technology “has made it possible for scientists to identify and evoke an almost infinite number of candidate drugs for the treatment of neuropsychiatric disorders.
“It has made it possible for scientists to identify and conjure up an almost infinite number of drug candidates”
Thus, “our results suggest that AF2 structures may be useful for drug discovery“, points out the lead author. This would provide “an almost infinite number of possibilities to create medications that achieve the intended goal to treat a disease; this type of AI tool can be invaluable.”
As has been announced, this system works by extracting data from a massive base of information about known proteins. With this information, the technology predicts protein structure models with the help of all this data.
Once you establish a model, you can simulate how different molecular compounds (as drug candidates) fit and produce the desired effects. In this way, researchers can use the resulting combinations to better understand protein interactions and create new drug candidates.
“Our results suggest that AF2 structures may be useful for drug discovery”
After creating prediction models, researchers can study physical models of proteins through complex techniques of microscopy and X-ray crystallography. “With the push of a button, up to 1.6 billion potential drugs were directed to the experimental models and the AlphaFold2 models. Interestingly, each model had a different result for the drug candidate.”
“Although the models have different results, they are very promising for the discovery of drugs“, is explained from the university. In this sense, “the researchers determined that the proportion of compounds that really altered the activity of the proteins for each of the models was around 50 and 20%,” they comment.
“Of the hundreds of millions of potential combinations, 54% of the drug-protein interactions that used AlphaFold2 sigma-2 protein models were successfully activated through a bound drug candidate,” they emphasize, so the results offer a 51% success rate.
