Fair warning to the protein complexes behind some of planet Earth’s most nefarious diseases, which have hid from mainstream researchers through decades of molecular science: Stony Brook University is coming for you.
An SBU-led research team has created an automated computer server that calculates complex computations for modeling protein interactions. If that sounds like thick science, it is: Aside from water, proteins are the most abundant substance in living cells, and modeling their interactions with cellular functions is critical to understanding – and ultimately defeating – an enormous range of diseases.
Understanding the rules by which proteins interact with cellular functions could enable researchers to design new interactions and lead to the development of new pharmacological treatments for major-league diseases ranging from cancer to HIV.
Unfortunately, the exact structure of protein interactions has baffled scientists for decades, for a very simple reason: Proteins are really, really small.
Now, researchers from SBU’s Laufer Center for Physical and Quantitative Biology, working with partners at Boston University, have developed a user-friendly, globally accessible server that simplifies those complex computations, allowing scientists to model protein interactions “with a handful of clicks from a home computer,” according to the university.
That’s the thrust of the ClusPro server, says scientist Dima Kozakov, an assistant professor in SBU’s Department of Applied Mathematics and Statistics and an affiliate member of both the Laufer Center and the university’s Institute for Advanced Computational Science.
Defining the atomic structure of protein complexes in three dimensions ain’t easy, Kozakov noted, but is key to helping researchers unlock the molecular mechanisms of numerous diseases, including some that have historically trumped science’s best minds.
The main problem is that proteins are smaller than the wavelength of visible light, meaning they can’t even be seen by even the best microscopes. So, studying protein complex structures requires the use of super-advanced microscopic visualization technology, such as X-ray crystallography.
Some of those “expensive and time-consuming experimental techniques” do an OK job revealing the structures of individual proteins, according to Kozakov, but the “delicate, short-lived protein complexes pose a much more difficult problem.”
“Obtaining 3D structures of such complexes is not a trivial matter,” he noted.
To bridge the gap between individual protein structures and protein complexes, researchers have traditionally relied on special computational algorithms that predict a complex’s structure based on the structures of individual proteins. In essence, computers calculate billions of possible molecular arrangements – a method known as “protein docking.”
For the last decade, scientists have developed unique protein-docking algorithms to predict disease behaviors, but those highly sophisticated computational programs were accessible only to specially trained scientists and programmers.
Now, the ClusPro server is availing the protein-docking science to all researchers.
ClusPro allows its users to model not only protein-protein complexes, but also interactions of proteins with other types of molecules. For instance, it can be used to predict how proteins bind to nucleic acids, such as RNA.
Kozakov’s creation is not the first automated protein-docking server ever created, but it has consistently outperformed other automated servers in international competitions. And while ClusPro is designed primarily for professional scientific use, its simplicity – and global availability – opens it up to educators and students in college and even high school courses, according to SBU.
The digital nitty-gritty of the National Science Foundation-supported research performed by Kozakov’s team and their Boston University collaborators is detailed in a recent paper published by the online journal Nature Protocols.