Of the many tropical diseases that have been plaguing the poverty-stricken population of underdeveloped and developing countries, Leishmaniasis is one. Its fatal form– Visceral Leishmaniasis, affects more than three lakh people every year. However, drug discovery and development for the disease has been slow for reasons including lack of capital, the complexity of the procedure, and the time taken to determine the protein structure of pathogens.
Role of protein structure in drug discovery
Protein is a complex chain of amino acids linked together in a unique pattern. Determining the chain of amino acids that form the protein is difficult. But what’s more challenging is determining the pattern in which the protein folds over itself.
Figuring this out involves understanding the interatomic forces in the structure. It’s like solving a puzzle that takes just microseconds to form. The three-dimensional protein structure has active sites where substrates attach and complete the cellular processes like a lock and key.
Subscribe to our Newsletter
Join our editors every weekday evening as they steer you through the most significant news of the day, introduce you to fresh perspectives, and provide unexpected moments of joy
Your newsletter subscriptions are subject to AIM Privacy Policy and Terms and Conditions.
AlphaFold revolutionises structural biology
Commonly used scientific procedures such as X-ray crystallography, cryo-electron microscopy, and nuclear magnetic resonance are expensive and time-consuming. Out of the 200 million proteins known, scientists have discovered the structure of only 1,70,000 due to these tedious procedures.
The DeepMind lab in the UK has developed an alternative to such techniques. Their proprietary algorithm AlphaFold can predict the structure of previously unsolvable proteins in a time-efficient manner and can prove to be a game-changer.
DeepMind’s team pushes AI beyond basic research. For instance, these experts have been applying technology to predict the 3D structure of proteins based solely on their genetic sequence. As a result, the algorithm swept the floor at competitions such as CASP, determining the protein structure in challenges accurately.
AlphaFold is trained on 170,000 protein structures. It determined the structure of SARS CoV-2 proteins accurately in January 2020. The structure was later deemed accurate by scientists and used extensively by researchers to understand the COVID-19 virus.
DeepMind shakes hand with DNDi
DeepMind has partnered with a Geneva-based Drugs for Neglected Diseases initiative(DNDi). The non-profit pharmaceutical company researches ignored diseases(mostly tropical diseases) such as Chagas disease and Leishmaniasis.
The company had an almost miraculous breakthrough in treating Sleeping Sickness when it developed Fexindazole, a safer drug than its predecessor Melarsoprol (a drug with a five percent mortality rate). However, its similar strategy of developing a drug of Chagas failed as the protozoan causing it has been highly resilient.
DNDi, in collaboration with researchers from the University of Washington, University of Dundee, and GlaxoSmithKline, has identified a unique molecule that binds with the protozoan Trypanosoma Cruzi, shuts it down and kills it. With AlphaFold, they can determine the protein structure of that molecule that can kill the pathogen not in years but days, speeding up the process of drug design and discovery.
Drug research can now be concluded in a couple of years. Knowledge of protein structure can allow us to devise proteins that denature the ones in pathogens, rendering them ineffective on humans.
Cautions and concerns
Applying AI to structural biology is not welcomed by all the members of the scientific community. For instance, Steven Finkbeiner, Professor of Neurology at the University of California, has concerns about the accuracy of the structure generated by AlphaFold. He cautions that because bacteria, protozoans, and other single-celled organisms are simple, it is easy to determine their protein structures. Still, the human body is complex, and solely depending on an algorithm would be a mistake. He further has concerns regarding the implications of the algorithm in the drug discovery process– it is “too soon” to make declarations about any such role in the future.
Despite the concerns and cautions, this development at DeepMind Labs could be a milestone in structural biology and reduce the time science takes to understand the biology of proteins. Furthermore, with DeepMind open-sourcing the algorithm for global use, we might be better equipped to combat the next pandemic.
