Understanding protein interactions is key to innumerable fields – including, notably, drug design. Now, researchers from the Georgia Institute of Technology have developed a machine learning tool to predict interactions between multiple proteins, paving the way for easier identification of drug targets for antibiotics and therapeutics.
The open-source, publicly available tool is called AF2Complex – short for AlphaFold 2 Complex, since the tool is built on top of DeepMind’s AlphaFold 2 protein structure prediction program. Like AlphaFold 2, AF2Complex predicts protein structures using amino acids; but AF2Complex is then also able to predict the likelihood of the modeled proteins forming complexes – or complexes of complexes (supercomplexes) – and identify other relevant variables (like interaction sites). “We essentially conduct computational experiments that try to figure out the atomic details of supercomplexes (large interacting groups of proteins) important to biological functions,” said Jeffrey Skolnick, a professor at Georgia Tech and one of the authors of the study, in an interview with Georgia Tech’s Audra Davidson. Skolnick compared AF2Complex to a “computational microscope powered by deep learning and supercomputing.”
“Unlike predicting structures of a single protein sequence, predicting the structural model of a supercomplex can be very complicated, especially when the components or stoichiometry of the complex is unknown,” said Mu Gao, another of the authors on the paper and a senior research scientist at Georgia Tech. “In this regard, AF2Complex could be a new computational tool for biologists to conduct trial experiments of different combinations of proteins.”
“The successful development of AF2Complex earlier this year makes us believe that this approach has tremendous potential in identifying and characterizing the set of protein-protein interactions important to life,” Gao added. “To further convince the broad molecular biology community, we [had to] demonstrate it with a more convincing, high impact application.”
The application in question: an essential protein transport pathway in E. coli that moves outer membrane proteins from where they are created (inside the cells) to their homes on the outer membranes of the bacteria. Leveraging the Summit supercomputer at Oak Ridge National Laboratory – which ranked fifth on the most recent Top500 list – the researchers compared all known proteins in E. coli’s cell envelope to several proteins already known to be key to the pathway in question.
“Encouragingly, among the top hits from computational screening, we found previously known interacting partners,” Skolnick said, adding that there the tool identified several previously unknown pairs. “Since the outer membrane pathway is both vital and unique to gram-negative bacteria [on the outer membrane], the key proteins involved in this pathway could be novel targets for new antibiotics. As such, our work that provides molecular insights about these new drug targets might be valuable to new therapeutic design.”
To learn more about this research, read the article from Georgia Tech’s Audra Davidson here or read the paper, “Deep learning-driven insights into super protein complexes for outer membrane protein biogenesis in bacteria,” here.