Mike Thomas
Breaking into Bioinformatics at ISU
Like the conscientious scientist he is, Mike Thomas expected to devote most of the fall semester setting up the new bioinformatics lab at Idaho State University.
An assistant professor of comparative genomics, Thomas joined ISU in 2003 to establish a center that would support modern biomedical research — and to teach the university’s first course in bioinformatics.
When Clusters Are a Full-time Job
Thomas thought he’d need months to create the center based on a cluster of Linux machines.
It seemed reasonable. After all, at the Medical Center of Wisconsin where Thomas conducted postdoctoral research in molecular evolution, dozens of programmers populated the bioinformatics lab, working around the clock to manage the clusters and write custom applications.
“In fact,” Thomas says, “we hired a programmer with a masters in bioinformatics in anticipation of having to spend a lot of time setting up and putting together a computing cluster at ISU. I thought it would take him working around the clock just to administer it.”
Computing in Quarter Time
“But when we explored alternatives,” Thomas counters, “and saw the Apple Workgroup Cluster for Bioinformatics, we immediately knew that we could put a cluster together very quickly and concentrate on biology rather than computer science.”
Today the lab’s programmer, Luobin Yang, spends only a quarter of his effort managing the cluster. He uses the rest of his time helping Thomas, faculty and students create or modify the programs they need for ecological and evolutionary research that includes genomics or proteomics components.
A computer scientist by trade, Luobin initially viewed the Xserve platform with some skepticism. “But now he loves it,” Thomas says, “because he can concentrate more on bioinformatics and less on system administration.”
Speedy Setup
Thomas says the cluster of Apple Xserve units running Mac OS X and the BioTeam’s iNquiry software was up and running within “a couple of days. That was really impressive. The Xserve units arrived in late September, and I thought it would take most of the fall semester to set up. But they were up and ready to be used almost immediately.”
The cluster took so little time to install, in fact, that Thomas was able to offer the first course in bioinformatics much sooner than he expected. “If we’d used any other system,” he says, “we would still have a long way to go. I wouldn’t have been in a position to teach a bioinformatics course for the spring semester.”
Platform Agnostic
Because researchers access the Apple Workgroup Cluster through a web portal, they can use any platform to run their research queries from anywhere they have Internet access. “Our researchers use Power Mac G5s, the latest PowerBooks, legacy Macs, PCs and Linux boxes,” Thomas points out. “All of them can access the Apple Workgroup Cluster perfectly here, on the road, at their colleague’s office, across the country.”
“The Xserve units arrived in late September, and I thought it would take most of the fall semester to set up. But they were up and ready to be used almost immediately.”
Before Thomas installed the cluster and iNquiry, ISU researchers were limited to standalone programs on their desktop computers, a kiosk in the Molecular Research Core Facility, the university’s sequencing center, or the BLAST server at the National Center for Biotechnology Information (NCBI). “Nothing was central, nothing had very much computational power, and nothing was flexible,” Thomas says.
Any Bioinformatics Tool They Need
Now ISU researchers can find almost any tool they need in BioTeam’s iNquiry package, which includes more than 200 bioinformatics tools, each with a web interface. “For instance,” says Thomas, “iNquiry comes with EMBOSS [European Molecular Biology Open Software Suite], a package of common molecular biology tools for manipulating and analyzing nucleic acid sequences and several useful molecular databases. We’ve added dozens of additional databases to suit the needs of our users.
“The cluster’s BLAST server is distributed so, instead of submitting sequences one at a time, users can submit thousands all at once. Other tools parse the results into “interesting bits” and return the results to the user.
“Users also can choose which databases they want to use and how they want to search,” Thomas says. “Our cluster gives them a lot of options they don’t have on NCBI unless they’re power users.”
1 2 Next Page >
