Hitachi Software is one of Japan’s largest software companies, with $1.5 billion in revenues and over 5,000 employees. MiraiBio, based in San Francisco, is the life-science group of Hitachi Software’s US subsidiary and specializes in software, reagents and instruments for biological and biomedical research.
Looking to enhance their offering, the team at MiraiBio wanted to incorporate new scientific articles relevant to a researcher’s particular sequence. This new compute-intensive feature would require another dedicated server which cost time and money. Instead, the team decided to deploy the feature on Amazon EC2, allowing a lower cost to implement, faster time-to-market and easier scalability.
“Our initial goal was to build a good working prototype and that took us only about 2 weeks to build from scratch (including the programming) on EC2. Setting up a new server on EC2 was easy and very compatible with agile software development,” claims Daniel Reda a bioinformatics specialist working on the project. “To prepare the server ourselves would have taken us at least 2 weeks in itself. I estimate that Amazon EC2 saved us about 200 man hours and thousands of dollars in hardware costs.”
How it works: When a biologist adds a new DNA sequence to their DNASIS SmartNote notebook, the new feature sends the sequence to the EC2 server, which runs the BLAST (Basic Local Alignment Search Tool) algorithm to compare that sequence to every other known DNA sequence. After identifying the sequence, the EC2 server runs some proprietary algorithms to find the most relevant articles that have been recently published on that sequence. This information is displayed to the biologist giving them quick access to referenceable articles and allowing them to continue their research.
“The EC2 server provides a powerful and scalable solution. Using EC2, we can run this search for hundreds or even thousands of sequences while not interrupting the main application for our customers.” “The low cost and rapid setup of EC2 made it easier to get past initial doubts about whether our new application feature idea would work. By testing the idea so quickly and inexpensively, it was easier to get approval to proceed with a full rollout,” recalls Reda.
Amazon Case Study: MiraiBio
Date: 28-08-2013