THE ALPHAFOLD NETWORK

By comparing and analysing protein structures, it is possible to get ideas about biological evolution, diseases, defence mechanisms, etc. This explains the human quest for finding the structures of proteins. 

In 1972, Christian B. Anfinsen won the Nobel Prize in Chemistry for his experiments that showed that a protein could fold into its structure based on the information contained in the sequence of amino acids. 

Max Perutz and others experimentally determined the first protein structures of myoglobin and haemoglobin. They did this through a method called X-ray crystallography that uses protein crystals and X-rays. 

India has had a legacy of being a top player in the field of protein structural work, both experimental and computational. The Ramachandran Plot devised nearly 60 years ago by G.N. Ramachandran and others from the University of Madras is used even today the world over to validate protein structures.

In 1994, John Moult and his colleagues started an exercise, to bring fun and rigour into structure prediction, called Critical Assessment of Protein Structure Prediction (CASP).

A Google company , named ‘DeepMind , based in London, recently declared that it had predicted the three-dimensional structures of more than 200 million proteins using  an AI-based protein structure prediction tool, called as AlphaFold DeepMind became a subsidiary of Google after a 2014 acquisition and is best known for its gamer AI 

This whole system is based on a computer system called deep neural network. The deep neural network is based on  the concept of human brain, neural networks.  This neural network uses the large amount of data   and provides desired output after  the data analysis , the way human brain works . 

The data analysis and processing is done at the junction point between the input & output network layers  by the black box called as ‘the hidden network’. When the protein sequence is given in the input commands of AlphaFold system, it provides the three dimensional  structures of these protein sequences. It works like human brain on the concept of deep learning which includes ‘training, learning, retraining and relearning’. 

It is using  the available structures of 1,70,000 proteins in the Protein Data Bank (PDB) to train the deep neural network computer model. Then it  works to learn the structural predictions of predict the proteins   not ithe PDB . Then by the method of relearning  it  improves the accuracy of prediction of  protein structures.

AlphaFold has so far  predicted the structures of about  214 million  protein sequences deposited in the Universal Protein Resource (UniProt) database. Proteins are called as the building blocks of the living cells .

It carries out all the important functions of the living cell. Evaluating the  protein structure and function will help to understand the human diseases.

The prevalent techniques to predict the  protein structures are x-ray crystallography, nuclear magnetic resonance spectroscopy, or cryogenic electron microscopy. But these  techniques are not  really accurate  .

 AlphaFold is not the infallible technique for predictions of protein structures . RoseTTaFold, developed by David Baker at the University of Washington in Seattle, U.S., is another tool although it is not as accurate as AlphaFold

 DeepMind’s AlphaFold team led by Demis Hassabis and John Jumper changed tack and switched to ‘attention-based’ deep learning which has been successful in image and speech recognitionAlphaFold s able to produce highly accurate side chains , when the backbone is highly accurate and considerably improves over template-based methods even when strong templates are available.

The methodology that we have taken in designing AlphaFold is a combination of the bioinformatics and physical approaches: we use a physical and geometric inductive bias to build components that learn from PDB data with minimal imposition of handcrafted features 

 In particular, AlphaFold is able to handle missing the physical context and produce accurate models in challenging cases such as intertwined homomers or proteins that only fold in the presence of an unknown haem group. AlphaFold generates predictions about individual protein structures, but it sheds little light on multiprotein complexes, protein-DNA interactions, protein-small molecule interactions, and the like—dynamics that are essential to understand for many biomedical use cases. Like any AI system AlphaFold has learned to make predictions based on its training data, it may struggle to accurately predict the shapes of unusual new proteins. 

With AlphaFold open-source, an entire ecosystem of biotechnology research and startups will spring up around it in the years ahead.