Proteins are the building blocks of life on Earth and they play a vital role in almost all biological processes. In recent years, protein research and engineering have become fundamental fields of study, and AlphaFold and AlphaFold 2, a machine learning tool created by DeepMind, have been major breakthroughs in protein research. AlphaFold can accurately predict the three-dimensional structures of proteins using only their amino acids.
However, despite these advancements, it is still extremely challenging to analyze the folding and transport of biological proteins experimentally. To resolve this issue, DeepMind researchers created AlphaFold 2 Complex (AF2Complex), a deep learning technique that can predict the physical interactions between several proteins. With an unprecedented level of detail, AF2Complex is able to determine which proteins interact with one another to form functional complexes, which parts of each structure are most likely to interact, and which protein complexes are most likely to combine to generate supercomplexes.
In order to demonstrate the tool’s effectiveness and the potential impact it can have on biomedical research, the researchers applied AF2Complex to a pathway in Escherichia coli (E. coli). The team looked at the production and movement of outer membrane proteins (OMPs) and compared a few proteins crucial for synthesizing and transporting OMPs to roughly 1,500 other proteins. By comparing the tool’s predictions to previously published experimental data, most of the formerly known interacting pairs, and several unknown ones, were accurately predicted by AF2Complex.
The AF2Complex model has the potential to provide new targets for developing antibiotics and treatments, as well as a foundation for leveraging AF2Complex to speed up biomedical research computationally. With the help of AF2Complex, researchers can now use various protein combinations to perform their experiments faster and more efficiently.