Technological advancements in computing power are enabling breakthroughs in research and medicine. Researchers use computational techniques to gain insights on how diseases develop and potential therapy methods. However, accessing enough computational power for supercomputers can be expensive. Furthermore, these resources are not easily accessible.
Distributed computing allows a computing project to be distributed across multiple machines. Some research teams now rely on an international network of volunteers who donate their spare CPU time to crunch scientific calculations. These volunteer distributed computing projects tap into the computing power of millions of personal computers globally. Through this, they discover new treatments and therapy methods for diseases, such as cancer, COVID-19, and Alzheimer’s disease.
A Crash Course Into Protein Structure
There are dozens of volunteer distributed computing projects. But most of the prominent ones focus on understanding protein structure. Proteins are biological molecules essential for most of our bodily processes. As our cells produce proteins, they fold and reassemble into extremely complex structures. These structures consist of smaller pieces called amino acids. Even small changes in protein structure can cause drastic changes in function.
Understanding the structure of certain proteins is critical to identifying possible treatments and therapy for diseases. For example, some research on Alzheimer’s disease suggests that misfolded amyloid proteins lead to brain damage seen in patients. Infectious agents, such as the virus that cause COVID-19 to rely on protein receptors to invade cells. Identifying the structure of the receptor protein can help researchers create treatments to block it.
Unfortunately, identifying the complete structure of proteins is difficult. Common experimental techniques involve crystallizing the protein and using X-rays to obtain information. This method requires a proper crystallization technique for the protein. However, this may not exist for newly discovered ones yet. Additionally, the use of reagents and specialized machines also increases the cost of this technique.
An alternative to experimental methods is to calculate the protein structure. It’s possible through analyzing how protein components interact with each other and with the environment. Proteins can contain millions of atoms and take on a myriad of conformations. Therefore, massive amounts of computational power are necessary.
Volunteer Distributed Computing Projects
Supercomputers can provide the calculation effort needed for analyzing proteins. But they are not always available to research teams. Volunteer distributed computing is an alternative solution. It splits the workload across multiple computers, making analysis more accessible for research teams.
A server splits a computational job into smaller work units that a regular computer can handle. This occurs in a typical distributed computing setup. Volunteers use a program to connect to the server and download the work units. All the processing happens locally, and the client program sends results from completed units back to the server.
The client program for these distributed computing projects handles the computations automatically. Users can continue with their daily tasks while the program crunches calculations in the background. Many client programs are designed to run at a lower priority than other computer programs. This way, they won’t noticeably slow down the computer. They also allow users to fine-tune the amount of computing power allocated to the computing task.
Rosetta@home is one of the classic volunteer distributed computing programs in biology. The Baker Laboratory runs the project at the University of Washington. It has been active in researching malaria, cancer, Alzheimer’s, and HIV since 2005.
Rosetta@home takes a creative approach in determining correct protein structures. They base this approach on a fundamental principle. All molecules adopt structures that require the least energy as they provide the most stability. A program can repeatedly determine its most stable form. It can do so by making small changes to a protein structure model and calculating its energy changes.
The processing software of Rosetta@home works with an initial structure model. It repeatedly makes tweaks until it finds the most stable configuration. Then, it sends the final structure back to servers for compilation. Using the initial model can influence the final calculated structure. With this, the server uses randomized starting models for each work unit to ensure better validity.
Signing up for Rosetta@home is straightforward. Users only need to install BOINC, a platform for distributed computing projects. Within the BOINC interface, users can select the Rosetta@home project to start calculations. Participants also earn points for each completed work unit. The client program is compatible with computers running Windows, Mac, and Linux operating systems. It can even run on mobile phones, tablets, and customized hardware.
The Rosetta@home project has accomplished many achievements. One of which is leading the development of a new cancer-drug candidate and a potential COVID-19 vaccine. Several drug candidates for treating COVID-19 have also emerged. These are thanks to the computational efforts of volunteers. During the COVID-19 pandemic, the Rosetta@home project reached 1.26 petaFLOPS of computing power in March 2020. This value is equivalent to around two quadrillion calculations per second.
Another project, Folding@home, was released back in 2000. It has seen a recent surge in popularity due to the COVID-19 pandemic. The project operates from Washington University in St. Louis and works on COVID-19, cancer, Alzheimer’s disease, and Huntington’s disease.
Folding@home tackles a different approach for computing protein structures. While Rosetta@home focuses on the final structure of the protein, Folding@home runs simulations. Its method identifies how protein structures evolve as proteins. This approach allows a more in-depth analysis of each protein. However, it reduces the number of proteins that can be calculated.
While Folding@home does not utilize the BOINC platform, registration is also easy. Volunteers have to download and run the software from the website. Users can work in teams and collectively earn points to rank in the public leaderboards.
Calculations from Folding@home work units have contributed to over 225 research papers. The entire network became the world’s first exaFLOP system. This means that the network was able to perform more than a billion calculations per second.
Computers are becoming more and more powerful. Thus, the world can expect even more results from volunteer distributed computing projects. Maybe one day, the cure for cancer, Alzheimer’s, or even mental illnesses will come from computers across the globe. These computing projects empower citizens to contribute toward solving the greatest health problems. What better way for humanity to come together than to save numerous lives?