Since the late 1970s, research has been conducted to understand how a protein’s structure could be determined solely from its amino acid sequence, and this past year Google may have helped propel that understanding forward through AlphaFold (2).
Currently, researchers use laboratory techniques like x-ray crystallography and cryo-electron microscopy when determining a protein’s structure. However, the significant amount of time, effort, and financial investment spent, and low resolution these methods can produce are considerations when running these experimental tests.
CASP, Critical Assessment of (protein) Structure Prediction, is used to assess the quality of protein structure with a score range from 0 to 100 GDT (Global Distance Test). A score of 90 GDT is considered competitive in predicting structure and since 1994 has been considered one of the best methods to measure protein structure prediction.
From the latest CASP assessment, the AlphaFold system obtained a median score of 94.2 GDT, indicating an average error of approximately 1.6 Angstroms. If AlphaFold is able to maintain its accuracy in predicting protein structure without needing to manipulate the physical protein itself, then the implications of its use in the field of biotechnology, medicine, and healthcare may be most likely highly significant. For instance, one direct benefit of this new method is the ability to study the protein structures of proteins typically difficult to analyze, such as membrane proteins, through traditional methods.
Broadening to more therapeutic approaches, improving protein structure predictions would allow groups to understand specific diseases on the molecular level more easily and understand how mutations within the amino acid sequence output would affect interactions such as binding and signaling. By understanding causal effects typically difficult to determine protein mutations, precision drug development could be utilized to complement or replace other forms of less individualized or effective treatment options.