Share This

e-Therapeutics Case Study

e-Therapeutics Runs Hundreds of Thousands of Computational Analyses in Minutes to Find Treatments for Biocomplex Diseases

Founded in 2003, e-Therapeutics plc (LSE ETX) is a U.K.-based drug discovery and development group focused on finding treatments for biocomplex diseases, such as cancer and neurodegeneration. The company currently has two drug candidates undergoing clinical trials, and its multiple discovery programs include a focus on cancer immunotherapy and on treating resistance to modern “targeted” cancer therapies.

The Challenge – Parallelizing Hundreds of Thousands of Analyses

Using an approach called Network Pharmacology, e-Therapeutics identifies and analyzes a specific network of proteins associated with a particular disease. It then identifies multiple intervention points that, if impacted simultaneously, can disrupt this network of proteins. The team then seeks drug molecules with the best overall impact on the protein network.

This approach is based on computational analysis of disease cells, not work done in a lab. While performing a single analysis is relatively straightforward and does not take a lot of time, the e-Therapeutics approach involves hundreds of thousands of analyses across multiple parameter sets and assumptions, resulting in an extraordinarily compute-intensive environment. In 2012, to accelerate processing, the company recognized it needed to parallelize its algorithms across a grid.

The Solution – The GridGain In-Memory Data Fabric

While investigating a solution, Jonny Wray, PhD, Head of Discovery Informatics at e-Therapeutics, remembered the GridGain In-Memory Data Fabric, which he had used at a previous company.

“I looked at the other compute grid technologies, but there really wasn’t much that could compete with GridGain,” said Wray. “It was actually a pretty simple, straightforward decision. GridGain had more features and worked more effectively out of the box.”

The GridGain In-Memory Data Fabric, based on Apache® Ignite™, enables high-performance transactions that run 1,000x faster than disk-based approaches. It provides high speed transactions, real-time streaming and fast analytics in a single, comprehensive data access and processing layer. The solution powers existing and new applications in a distributed, massively parallel architecture on affordable, industry-standard hardware, which can be easily scaled by adding more nodes to the compute grid.

The Benefits

New Capabilities

The GridGain In-Memory Data Fabric now powers e-Therapeutics’ Network Pharmacology platform. The platform launched with 20 nodes on one 20-core commodity server and has now grown to 100 nodes on five servers. The platform delivers approximately an 80x speed increase over the non-parallelized environment.

“GridGain has allowed us to complete in just a few hours or even minutes analysis projects that used to take weeks,” said Wray. “Just as important, we’ve been able to launch initiatives that were simply computationally infeasible before. We’ve been able to build far larger and more complex models of cells, which has opened up fundamentally new approaches to tackling diseases. And our scientists can more quickly explore many more hypotheses.”

Increased Productivity

GridGain delivers increased productivity for e-Therapeutics’ disease biology specialists, who lack a computational informatics background. The company developed a simple, web-based interface that connects to an internally developed micro-service using the GridGain API, so the scientists can now run analyses without having to work from a command line or having to consult with a computational informatics specialist. “They simply submit their jobs through our web application and get their answers back far faster than they used to,” said Wray. “It’s that simple.”

In addition to simplicity and ease-of-use, GridGain allows multiple biologists to work on multiple projects at the same time. This effectively changes how they work and allows them to accomplish far more in less time. “Over the last year and a half, we have been able to run 10 concurrent discovery projects despite a rather small team of scientists,” added Wray. “We have moved a number of these discoveries into testing. We could not have achieved this pace without GridGain.”

Peace of Mind With Apache Ignite

Because the GridGain solution is based on Apache Ignite, Wray also sees long-term peace of mind. “The stability and longevity of the Apache Software Foundation means we can move forward knowing that the technology powering our platform will continue to be developed and well maintained for years to come,” said Wray. “That provides us with a lot of peace of mind.”

Download Case Study