Novel AI Can Create New Proteins That Can Cure Diseases, Develop New Materials

Researchers have developed a new AI tool that could help them unravel previously undiscovered proteins while also helping them design new ones, reveals a report by MIT Technology Review.

Unsplash

The report highlights that the AI could help unlock the development of more efficient vaccines, expedite research for curing diseases such as cancer or help in the discovery of materials not known to man before.

In 2020, DeepMind opened the eyes of the world with AlphaFold -- an AI tool that harnesses deep learning to solve gargantuan challenges in the fields of biology by accurately predicting the shapes of proteins known to science.

Now, a new tool dubbed Protein MPNN, revealed by a group of researchers from the University of Washington offers a powerful tool to that technology.

Usually, researchers engineer proteins by tweaking those that are already present in nature, however, Protein MPNN will open an entire new universe of possible proteins that researchers can design from scratch.

Basically, if researchers already have a particular protein structure in mind, they will help them find the amino acid sequence that folds into that shape. It works based on a neural network that¡¯s trained on a considerable number of examples of amino acids which fold into three-dimensional structures.

However, researchers also need to solve another issue -- to design proteins that are useful for real-world applications like an enzyme that could devour and eliminate plastic, and for this, they¡¯d have to figure out what protein backbone would have that function.

For this, researchers made use of two ML methods dubbed ¡®constrained hallucination¡¯ and ¡®in painting¡¯. The former allows users to do a random search among all possible protein sequences with certain functions, allowing them to explore the space of all possible protein structures thanks to ML¡¯s ability to crunch a ton of data sets. Only 20 amino acids exist which can be fused into a massive number of possible sequences.

UNIVERSITY OF WASHINGTON

In the latter, it works like autocomplete on your phone¡¯s keyboard, but for protein structures and sequences. Based on these methods, researchers are able to create a completely new protein that hasn¡¯t been in nature before, like a massive ring-like structure.

David Baker, one of the scientists behind the paper and director of the Institute for Protein Design at the University of Washington, said in a statement, ¡°In nature, proteins solve basically all the problems of life, ranging from harvesting energy from sunlight to making molecules. Everything in biology happens from proteins.¡±

¡°They evolved over the course of evolution to solve the problems that organisms faced during evolution. But we face new problems today, like covid. If we could design proteins that were as good at solving new problems as the ones that evolved during evolution are at solving old problems, it would be really, really powerful.¡±

ProteinMPNN, is now available for free on the open-source repository GitHub.

For more in the world of technology and science, keep reading Indiatimes.com