AWS Public Sector Blog

Proteins wiggling and jiggling: The University of Nottingham’s Crossbow project paves a new path for biomolecular research using high-performance computing (HPC) and the cloud

An introduction to the team behind Crossbow

The University of Nottingham has a history dating back to 1881, and while the university is now global with campuses in China and Malaysia, its flagship campus remains in the UK.

Today, the university’s research efforts span nearly every discipline. One current project is Crossbow, a new, open source software project conceived and developed by Dr. Christian Suess, a research fellow at The University of Nottingham in conjunction with principal investigator Prof. Charlie Laughton, professor of computational pharmaceutical science. Crossbow is based out of the school of pharmacy at the University of Nottingham, where there is a particular interest in researching the design of new medicines using computer simulations of drugs and proteins.

Together, the two researchers set out to study how proteins move, interact, and behave in a variety of complex situations using newly developed biomolecular simulations run through the cloud. When describing Crossbow, Suess says, “we look at molecules and build them up into bigger systems, into proteins, little machines that control all functions of the body. We make them wiggle and jiggle, which is how they behave in the body. When we simulate their movement, we can look at them, and understand them. Once we understand what they are doing, we can explore them, and then we can design medicines that make them work better.”

Research then versus now: Crossbow in action on the cloud

Researching how proteins “wiggle and jiggle” is not new. But, the way in which Dr. Suess and Prof. Laughton have approached the situation is. Typically, a researcher has to go beyond his or her standard bounds of research in order to learn an entirely distinct set of technical knowledge just to run the necessary simulations – essentially, researchers must take time to learn computer science to complete projects. This represented a large barrier to entry for many. Dr. Suess explains that this friction is what spurred the idea for Crossbow. “If you’ve got a computational scientist who is good at producing a simulation, they’re not necessarily able to dedicate time to learning computer science. So, their research gets stuck on a laptop,” says Dr. Suess. To this point, Prof. Laughton adds, “a researcher might be a biologist or a chemist, not necessarily a ‘computer person.’ What we’re interested in doing is to develop tools so that non-specialists can use high performance computing (HPC) efficiently, and as easily as possible.” Dr. Suess and Prof. Laughton are effectively closing the technical knowledge gap by creating cloud-based tools that provide a pathway for an accelerated pace of time to science.

Furthermore, in addition to the baseline of computer knowledge necessary to run simulations, the types of simulations that need to be run in order to gather meaningful results require a massive amount of computing power – far beyond what a typical laptop is capable of providing. Because of this, researchers were required to navigate a complex, time-intensive process of securing time in an HPC lab, or running stripped-down, basic simulations that yield less complete results than what a simulation run on HPC can provide. The drawbacks of this model are immense. Prof. Laughton explains, “If you want to do the research well, you need a lot of computing power. You can do simple simulations without a lot of power, but they might not be accurate.” Inaccurate results, simplified questions due to inadequate computing resources, and lengthy wait times all result in lower quality research.

To complete their own research, Dr. Suess and Prof. Laughton’s interest was piqued by what the cloud could provide. “On-demand resources, and we can build it however we want? That was a very attractive prospect for us. We were interested in these capabilities, and we figured that others in our community would be interested in these tools, too,” says Dr. Suess. They are the first to build and offer these tools to other researchers, which recently went through a beta test, and are now publicly available.

Crossbow explained

Crossbow has two packages, Xbow and Xbowflow. Xbow provides tools for computational scientists who are not experienced with cloud computing. This allows them to launch personal compute clusters in the cloud, tailored to their individual research, and run simple jobs from their personal laptop. Xbowflow has tools geared toward more experienced biomolecular scientists; it allows them to create custom workflows that run complex, large-scale calculations across an Xbow cluster.

In order to build these tools, Dr. Suess and Prof. Laughton relied on Amazon Web Services (AWS). Among other services, the team uses Amazon Elastic Compute Cloud (Amazon EC2), Amazon Elastic File System (Amazon EFS), and Amazon Simple Storage Service (Amazon S3).

Beyond proteins: Crossbow’s implications for future research

The ripple effect of Crossbow is poised to change the landscape for many universities. Researchers are able to leap beyond technical knowledge gaps, and bypass long wait times for computing resources in a lab. “The fact that we can get up and running in 5 minutes with instant and on-demand resources is a massive benefit. We can scale up instantly. The library of resources is immense. We can try things, fail, and figure out why – all in an afternoon. We won’t have to spend 6 weeks trying to find out why something didn’t work” says Dr. Suess. With faster run times, ease of use, and cost reductions of about 70%, Dr. Suess and Prof. Laughton are very pleased with the immediate results of working in the cloud.

Currently, the tools were developed specifically to research new drugs and study protein behavior. But, looking forward, the tools are poised to change the way universities fundamentally view, fund, and implement research. Prof. Laughton says, “what’s interesting is the fact that the cloud provides a totally different model for how we approach problems. It allows us to completely rethink how to frame a problem, and therefore, how we may solve problems we were unable to solve before. It has helped us see that a lot of the ways people think about the work they want to do and the sorts of methods they want to use to solve the problems, are a bit constrained by a language and a set of tools that aren’t modern. Now, we have the cloud, which allows us to think differently, to start new.”

Beyond proteins wiggling and jiggling, Prof. Laughton believes the cloud will augment many research avenues, including but not limited to: “drug design, materials science, understanding the basics of new proteins used for biotechnology to remediate waste, or to produce new fuel types.” And, all of this research will happen at an accelerated pace and a fraction of the cost compared to legacy methods of research, according to Prof. Laughton: “you can run simulations in a matter of minutes, hours, whatever the workload is, it’s now faster and easier than it was before.”