e-Therapeutics Leverages GridGain for Complex and Memory-Intensive Network Pharmacology
Modeling and Computations are 20 Times Faster with GridGain In-Memory Data Fabric
“The scientists have no idea the platform is running on GridGain. Things just work.”— Jonny Wray, Head of Discovery
Informatics for e-Therapeutics
“[While impressive,] it’s more than the fact that we’re more than 20 times faster than without the GridGain solution. It’s the fact that we can now do what we couldn’t do before.”— Jonny Wray, Head of Discovery Informatics for e-Therapeutics
“GridGain’s programming model makes it really straightforward and easy. They provided all the infrastructure needed to allow us to write our code for the grid, which was one of the reasons we chose GridGain.”— Jonny Wray, Head of Discovery Informatics for e-Therapeutics
Founded in 2001, e-Therapeutics is a drug discovery and development company based in the United Kingdom. A pioneer of network pharmacology, the company’s mission is to apply network pharmacology to find new treatments for human disease.
Advanced Network Pharmacology Platform
Based on advances in chemical biology and network science, “network pharmacology” is a distinctive new approach to drug discovery. It involves the application of network analysis to determine the set of proteins most critical in any disease, and chemical biology to identify molecules capable of targeting that set of proteins. e-Therapeutics has developed a proprietary network pharmacology platform for analyzing networks of proteins associated with particular diseases. After these analyses—which number in the millions—drug candidates with optimal impact are identified.
The Challenge: Millions of Computations Running in Parallel
While designing an extension to the platform, e-Therapeutics had to address the challenge of running millions of in-memory analyses repeatedly, screening various compounds against multiple computer models of diseases with different parameters to see which compounds would be most active. Testing these would have taken weeks on the company’s system.
“We were looking into how we could implement the system that did the analyses, so we looked at doing it ourselves, and using both open source and commercial solutions,” said Jonny Wray, Head of Discovery Informatics for e-Therapeutics. Then Wray remembered another product he’d used in the past — a grid-based solution from GridGain Systems. “I had used GridGain’s In-Memory Data Fabric in the past so I knew we should consider it. Without a data fabric solution, our network pharmacology platform would take weeks to test all the compounds against one disease model, so we evaluated GridGain.”
The Solution: GridGain In-Memory Data Fabric
The GridGain In-Memory Data Fabric is an advanced data management system, available as an open source distribution, that offers distributed in-memory caching, distributed computations, and streaming. It addresses performance and provides significant scalability for e-Therapeutics by allowing the network pharmacology platform application to compute on it. The GridGain In-Memory Data Fabric is now a key component of e-Therapeutics’ network pharmacology platform. Within this heterogeneous environment, GridGain provides a high performance, low latency In-Memory Data Fabric enabling fast computations for multiple different analytic approaches, not just the original platform extension that triggered the use of GridGain.
According to Wray, with GridGain in place, the network pharmacology modeling and computations are 20 times faster than before. “While that’s impressive, it’s more than the fact that we’re more than 20 times faster than without the GridGain solution. It’s the fact that we can now do what we couldn’t do before,” said Wray. “Just a couple of these tests, without the data fabric, would take weeks and you would need to run many of them in parallel, so it wasn’t something we necessarily would have considered. With GridGain, they take just days and as such, made the platform extensions feasible.”
Benefits: Seamless Integration, Transparency and Effective Results
Wray appreciated the ease of integration for the GridGain solution. “While we had to write some code to integrate the solution, GridGain’s programming model makes it really straightforward and easy. They provided all the infrastructure needed to allow us to write our code for the grid, which was one of the reasons we chose GridGain.”
Wray says that the GridGain solution is transparent to the e-Therapeutics scientists — in a good way — until something goes wrong, which it seldom does. “The scientists have no idea the platform is running on GridGain. Once it was deployed, they don’t know any differently — things just work,” said Wray. That is until something goes wrong, which actually leaves them with a situation better than the one they had before GridGain anyway. “We have had a couple of technical issues, where some of our 20 compute nodes crashed, so we’ve had to adjust and work with the smaller grid, but the analysis still works faster than before we had the GridGain solution,” said Wray. “That actually speaks to the reliability of GridGain. If our prior solution had failed, there was no redundancy and we’d have had to shut down altogether.” e-Therapeutics is happy with its selection of the GridGain In-Memory Data Fabric for its high-performance computing requirements. “We’re pleased with the results, which give us 20x faster data processing performance,” said Wray. “We get great support from GridGain when we need it, and we are looking into working with them on a bigger solution and possibly an on-demand cloud solution.”
Sponsored by GridGain