Proteins are crucial to almost every fundamental biological process necessary for life. They do everything from create and maintain the shape of cells to serving as both signal and receiver for cellular communications. Proteins are composed on long chains of amino acids and they perform their varied tasks by folding themselves into precise 3D structures that determine how they function and interact with other molecules.
Because their exact shape is so crucial to their function research into uncovering the exact shape is a central task to molecular biology. This task is especially important for the development of lifesaving and life-altering medicines. Predicting how proteins fold themselves based on their amino acid sequence has been advanced in recent years by computational methods.
AI opens doors to super fast prediction
But even with huge steps forward, the methods to are limited in the scale and scope of the proteins that can be predicted. However new research from Harvard may change that A scientist from the Harvard medical school as used a form of artificial intelligence (AI) to predict the structure of theoretically any protein based on its amino acid sequence.
The AI methods is known as deep learning and could improve the current predictive methods by making them a million times faster with the same accuracy. systems biologist Mohammed AlQuraishi has published his research about the use of AI in protein shape prediction in Cell Systems on April 17. "Protein folding has been one of the most important problems for biochemists over the last half century, and this approach represents a fundamentally new way of tackling that challenge," said AlQuraishi, instructor in systems biology in the Blavatnik Institute at HMS and a fellow in the Laboratory of Systems Pharmacology.
Long way to go
"We now have a whole new vista from which to explore protein folding, and I think we've just begun to scratch the surface." Proteins are built from a library of 20 different amino acids. These different amino acids can be imagined like letters in an alphabet that can be combined into words, sentences, paragraphs and larger texts Unlike a flat page though, amino acids are physical objects positioned in space with its chains forming loops, spirals, sheets and twists.
Despite an intense effort by scientists for more than four decades a fast and cost-effective way to predict these complex shapes hasn’t been achieved. The new AI method could open the doors to understanding disease and designing ways to fight them.
"What's compelling about the problem is that it's fairly easy to state: take a sequence and figure out the shape," AlQuraishi said. "A protein starts off as an unstructured string that has to take on a 3D shape, and the possible sets of shapes that a string can fold into is huge. Many proteins are thousands of amino acids long, and the complexity quickly exceeds the capacity of human intuition or even the most powerful computers."
The new method shows huge promise. In its current form it isn’t ready to be used for drug discovery or design, but will continue to be optimized by AlQuraishi and others in his lab.