Proteins are among the most important molecules in any living organisms. They are designed to perform particular functions in the body like transporting nutrients, blocking pathogens etc. They are essential for life and are made up of smaller units called amino acids arranged in a string structure. There are 20 possible amino acids to choose from. Full diversity of proteins from insulin to collagen to antibodies that are needed for immunity, all of these proteins are made from sequences of these 20 amino acids.
To know the 3D structure of the protein is the holy grail of biology. Everything from positioning of different amino acids, the structure of the bends and the folds, the complexity of long amino acid chains that forms the 3D structure of the protein and knowing this essential to allow the protein to function.
The scientific community was stunned by a major breakthrough just a few days ago. DeepMind, a London based artificial intelligence company owned by Google announced that it had predicted the structure of nearly every protein to known to science. Normally understanding a structure of one protein can take from 3 to 5 years. Researchers might spend their entire PHDs to figure out how a single protein folds. DeepMind’s AI system called the Alphafold trained on existing data and went far beyond its earlier trial accomplishments. With its clever algorithm, Alphafold can accurately predict the 3D structure of over 200 million proteins. This is an astonishing achievement and it is also open source. Anyone in the research community can use this information for free which promises to unlock future innovations.
This breakthrough can now predict protein structures for plants, bacteria, animals and other organisms. This will lead to opening of many opportunities for Alphafold to have impact on important issues such as sustainability, food insecurity and neglected diseases. Many experts are calling this event a new era in digital biology where AI and computational methods can help us understand and model biological processes.