Stem cell research, high-end Linux solution

Published: July 3, 2004
137

Princeton University implements a Linux on POWER system from IBM for critical stem cell research. How research applications can benefit from Linux.

If Biotechnology is the next frontier, then stem cell research can truly blaze trails in conquering the devastation of disease. A research database at Princeton University is an exciting part of that story.

Princeton is using IBM’s technologies to run its stem cell research database—a unique repository to support knowledge-seeking efforts of scientists that may eventually result in a better future for those who fall victim to Alzheimer’s disease, heart disease, spinal cord injury, diabetes, and strokes. In using these technologies, Princeton’s scientists access, share, and collaborate on data with researchers worldwide.

In this interview, Princeton’s Dr. Douglas Welsh, Lecturer and Senior Professional Technical Staff Member, Department of Molecular Biology, discusses how Power Architecture and 64-bit Linux support the department’s research efforts.

What is your stem cell research database all about?

Dr. Welsh: The database is a collection of what we call “expression data.” This is a collection of data dealing with, for example, which genes are expressed higher in stem cells than in normal cells, or which genes are expressed lower in stem cells than in normal cells.

Using such information, we can determine which genes are important to stem cell development. Knowledge of this is important to learning more about a number of disease states, spinal cord injuries, and the Regeneration of various functions in the body.

At Princeton, our emphasis has been on collecting and also analyzing the data, correlating it with expressed data in other systems and other databases in the world. Expression analysis looks at thousands of genes for each sample to find the level of expression. We’re looking at 22,000 records at a time. There are a couple of hundred megabytes of data from each sample.

Our web site is linked to our database running Linux. So the database is both a repository for the data and also for information on what’s been done to analyze it. At Princeton, that is our emphasis: to not only collect all this expression data but to analyze the data.

So this is not the only stem cell research database in the world. What makes it unique?

Dr. Welsh: It is unique in our emphasis on bioinformatics-analyzing the gene expression, included in the database, rather than merely behaving as curators.

We understand you were running other types of systems before switching to an IBM platform. What made you look for a new computing approach?
Dr. Welsh: We were using Dell Intel processor-based servers running Linux, but we found that we needed more capacity.

Our technology lacked the power required to share and collaborate on those results globally. Based on other projects that we had in the department as well, we found that IBM’s Power Architecture and servers could best support our needs. We were basically looking for a platform that could bring reliability, stability, and support.

What were some of the must-haves on your checklist that made you go for the pSeries running Linux?

Dr. Welsh: The department chose a p630 system powered by a POWER4+ processor running SUSE LINUX. We were looking for a platform that could bring reliability and stability. The fact that it was running a standard version of Linux was important to us. And we were already familiar with SUSE LINUX.

We did not want a “specialized” distribution of Linux; we needed a generic type of Linux where we would have no issues over having to learn a new way of administering the system, problems with utilities, knowing where programs would be, the directory tree, those sorts of things.

Also, our choice was made just based on our experience in the department. I have been here for 20 years and we have been working with IBM all of those years. The department has an IBM Beowulf cluster and RS/6000 servers. We have had very good hardware and software support from IBM. So the other factor was support.

Is this pSeries with Linux platform up and running?

Dr. Welsh: Oh, yes. We installed the system last fall. Since then, I don’t know if it has had any downtime at all except maybe to install security patches. It’s been extremely reliable.

Do you have any thoughts about turning to the upcoming POWER5?

Dr. Welsh: We presently find all that we need in POWER4. The POWER system is good in that if we wanted to we could add processors, disk capacity, memory For now and the foreseeable future, we are very happy with POWER4+.

Linux, sponsored by IBM, Oracle and Sun Middle East
Sunday, July 04 – 2004 at 13:46 UAE local time (GMT+4)