AlphaFold

Terms from Artificial Intelligence: humans at the heart of algorithms

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

AlphaFold and its successor AlphaFold2 applied some of the technology originally designed for AlphaGo and applied this to protein folding. While some of the behaviour of proteins is given by their chemical and molecular structure, the physical shape of a protein, the way it naturally folds given the order of molecular units, is also critical. For example, this folded shape deterimines which parts of the molecular structure are exposed and which hidden inside and thus the extent to which they are free to interact with other molecules and engage in biological processs. Determing the folded shape from the chemcal structure is highly complex, and often relied on experimentally observing the structure using electron microscopy and other methods. AlphaFold revolutionised this field allowing new molecules to be tested for efficacy purely by computer simulation (in silco). This means only a relatively small number need progress to full lab-based synthesis and testing.

Used on Chap. 14: page 332; Chap. 23: page 562

Also known as AlphaFold2