After the success of DeepMinda British company of Google, in the prediction of the 3D structure of beyond 200 million proteins thanks to the algorithm AlphaFoldnow it’s on Half. The company announced that it has planned the facility outside 600 million proteins of microbial origin that have not yet been characterized. This was possible thanks to the algorithm ESMfold60 times faster yet less accurate than AlphaFold.
The result can be found on the site bioRxiv, a site that collects all scientific articles that have not yet been researched by the scientific community. The researchers used a model of artificial intelligence that has been used in several language activities, because this is able to predict a text from a few words. This algorithm is trained using known protein sequences and formed by the combination of 20 amino acidseach of which are represented by a letter.
So the system has learned to auto-complete the sequence of proteins as if it were a text and determine the 3D structure of the protein. The test used DNA sequences recovered from sources such as soil, seawater, human gut and other microorganism habitats. Most of these sequences were from organisms that had not been cultured in the lab and therefore significant unknown to scientists.
More was provided in just two weeks 617 million protein structures, and more than a third of these forecasts are expected to be of high quality and therefore highly reliable. The results and code behind this artificial intelligence model are freely accessible.
Source: Lega Nerd
I am Bret Jackson, a professional journalist and author for Gadget Onus, where I specialize in writing about the gaming industry. With over 6 years of experience in my field, I have built up an extensive portfolio that ranges from reviews to interviews with top figures within the industry. My work has been featured on various news sites, providing readers with insightful analysis regarding the current state of gaming culture.