MySQL Helps EraGen Biosciences Deliver Powerful Analytic Tools for Genome Research

EraGen Biosciences is a Madison, Wisconsin-based life science company that designs, develops and commercializes products and technologies that accelerate the drug discovery process, advance molecular diagnostics and enable functional proteomics -- the branch of genetics that studies the full set of proteins encoded by a genome. Once researchers map a genome, proteomics is the next step in moving toward new drugs that may some day mitigate or cure many of the diseases now common to humankind.

To speed researchers' efforts, EraGen developed its premier product, MasterCatalog™, an integrated phylogenomics software application that employs evolutionary analysis to predict protein function. This powerful data mining and analysis tool is built on the MySQL® database, the world's leading open -source database with more than 4 million installations worldwide.

EraGen considers MySQL a mission-critical application, not only as the cornerstone of its MasterCatalog product, but also as a powerful research and development tool for continuing product improvement and development.

According to Kyle Munn, bioinformatics production manager, when the company began its R&D efforts in 1998, Oracle was the database most commonly used within the genome research community. However, it didn't deliver the speed and reliability that EraGen required.

"When we began building the MasterCatalog database, we required the input/output of several hundred gigabytes of data," says Munn. "MySQL was the only database engine that allowed us to achieve the level of performance necessary to complete a build cycle within our specified time frame."

The specified time frame was 90 days, based on a quarterly release schedule. Difficulty in maintaining the schedule initially arose because several steps in the production pipeline required simultaneous analysis of large amounts of data that had to be fully synchronized with the database. To minimize this bottleneck, the MasterCatalog team needed to import and export quickly tens of gigabytes of data. The rule of thumb used was “no more than one day delay at any import/export step”.

The database servers used are dual Intel 2.0GHz servers with 2GB of memory. Each server also contains a 3ware 7xxx series controller to manage the RAID used by MySQL. According to Munn, running MySQL on a system with a 3ware IDE RAID is the most cost-effective solution when a high-performance database is required.

"We can quickly load and retrieve large amounts of data, and replicate selected portions of the database to other servers via non-CPU intensive file operations," says Munn. "MySQL also facilitates the use of optimized transfer protocols to quickly move databases, in full or in part, around the network to meet replication or archival requirements."

Munn states that, in principle, these goals could have been achieved with other databases. However, it would have required "a vast amount of fine tuning, and a very expensive DBA to do the tuning." Other solutions would also have required expensive hardware to meet EraGen's performance requirements. MySQL provided the ease of use and performance that EraGen needed to get the MasterCatalog to market in a timely manner.

Eragen markets the MasterCatalog as a complete solution, specifying everything from the hardware to the version of software packages, including MySQL. This optimizes the MySQL database to work with the company's client/server based graphical user interface. This, in turn, simplifies updating customers' installations by eliminating involved and complex SQL-based operations. Instead, most of the files issued are read-only, allowing the update of some tables via file operations.

"MySQL is highly reliable, so we rarely require support," says Munn. "However, the few times we've called to report a bug, or to request information, the response has been immediate and helpful."

EraGen recently upgraded to MySQL v. 4.0 to support future R&D and product efforts.

"We continue to use MySQL because it is second to none in performance, reliability and support," says Munn. "To get the same level of benefit from a comparable database would cost almost several hundred times more than the cost of all of our MySQL licenses combined, because we would likely need to hire a full-time database administrator. With the level of support provided by MySQL and the MySQL users groups, we are hard pressed to justify such an expense."