Skip to main content
Premium Trial:

Request an Annual Quote

High Performance Computing Cos Compete For Larger Piece of Bioinformatics Pie

Premium

IN 1996 Chris Dagdigian of Genetics Institute made a bold move: He opted for Compaq over the more popular SGI when it came time to buy his company’s first Unix system.

“At the time the heavy player was SGI,” said Dagdigian, 28, who is now the bioinformatics systems manager at Genetics Institute. “We definitely stepped away from the norm when we decided to standardize on Alpha for our heavy lifting stuff.”

Over the past few years, Compaq and Sun Microsystems have both elbowed their way to the forefront of the bioinformatics space, gobbling up some of SGI’s market share. Yet, in the fast-growing world of bioinformatics, the challenge of staying ahead will only become more difficult as an increasing number of companies start gunning for a piece of the growing bioinformatics pie.

IBM, which expects the life sciences industry to balloon from $3.5 billion today to $9 billion over the next two years, recently said it would invest $100 million over the next two and a half years to develop tools designed for the bioinformatics industry. And in its first foray into the genomics sector, the company last week announced that it would supply the software for Incyte Genomics’ new research tool.

While IBM has a long way to go to catch up with the likes of Sun, Compaq, and SGI, industry pundits agree it’s too soon to count a powerhouse like IBM out of the game.

“IBM will be a strong player,” said Rich Partridge, vice president of enterprise servers at computer consultant DH Brown & Associates. “They’re pervasive everywhere but Compaq is not to be ruled out. Compaq still has strength because of their high performance chips and that won’t go away.”

WHAT THE COMPETITION LOOKS LIKE

In fact, Compaq has invested considerable resources to shore up its position in the bioinformatics sector. The company has built a server that handles some of the most compute-intensive work in the industry and it has got an impressive list of customers to show for it. Celera Genomics, the Institute for Genomic Research, MIT’s Whitehead Institute, and the Sanger Centre all use Compaq.

Marshall Peterson, vice president of infrastructure technology at Celera, said “It’s a real problem getting the level of skilled people to design and run these systems. Even on the low end they tend to be very complex. Celera now has 80 terabytes of data—that’s more data than some countries have! Compaq provides us excellent engineering support. They understand bioinformatics.”

While Compaq has been passing some big tests and scoring some great deals, they are by no means the only high-performance computer player in the industry. Although no one tracks computer companies’ shares of the bioinformatics market, anecdotal evidence indicates that Sun, Compaq, and SGI each have significant footholds.

For seriously large technical computational tasks, Alpha-based systems are highly regarded for both their CPU performance and memory bandwidth. But it’s also important to consider which applications are needed before selecting the hardware. There is a wider selection of commercial applications available for Sun’s SPARC and HP’s PA-RISC platforms. Researchers using their own proprietary software have much more freedom to explore other options, such as IBM and SGI 64-bit RISC-based computers.

Although SGI has taken a beating over the past few years while it tries to restructure, the company is considered by bioinformaticists to have some of the best 3-dimensional visualization tools that are critical for such tasks as understanding the structure of proteins. SGI now also has plans to offer systems specifically designed for bioinformaticists.

The company recently debuted a modular computing technology called NUMAflex that it believes will be ideal for the demands of bioinformatics. Part of the Origin 3000 series, the “brick”-style system can be used to construct small to large systems from a common set of building blocks.

“We offer the modularity the computer industry has been talking about for decades,” said Juli Nash, SGI’s biology market manager. “You can buy the bricks that you need and build up. Before this, you had to buy all-in-one boxes.”

Meanwhile, Sun has some performance advantages for less specialized needs. Although its SPARC hardware performance tends to lag the competition, it has the advantage of offering the dominant platform for Oracle and Unix.

“Sun is not fast and it’s not cost-effective, but some standardize on Sun,” said one bioinformaticist who requested anonymity. “Sun is good if you want Oracle databases and it’s known as the common denominator of the Unix world.”

Indeed, Compaq’s Alpha servers account for only seven percent of the Unix market, compared with Sun’s 28 percent.

COMBINING HARDWARE

As a testimony to the various virtues of each of the leading hardware systems, many companies are finding the best approach to be a combination of technologies. Take Incyte Genomics, for example. The genomics database and technology giant uses hardware from nearly all the main vendors’ hardware tools.

Its computing muscle comes from SGI desktop machines and Origin 2000s. And it also has Sun equipment ranging from SPARC 5 workstations up to 32-processor E10,000 systems as well as a bunch of Compaq Alpha servers from DS 10s up to the ES series of processors. To boot, it has a huge Linux cluster.

“If you’re performing jobs that require a lot of interaction, Sun performs the best,” said Stu Jackson, bioinformatics production manager at Incyte.

But Jackson said SGI machines are the best in terms of providing visualization capabilities and high degrees of parallelism, which make them especially useful for proteomics research.

Meanwhile, Jackson noted that Compaq’s latest generation Alphas offer a lot of memory space fairly economically and the I/O channel is huge, so large amounts of data can be moved in and out rapidly.

As for Linux, Incyte has found that it is able to do smaller, stand-alone jobs very well and very cost-effectively. The shortcoming of Linux is that it only scales to four processors. For “larger science,” such as proteomics, bigger systems are needed.

All of this indicates that no one company can or will dominate the market. Certainly there will be much fighting in this space.

“IBM has a breadth of installations [at pharmaceutical companies] and that visibility will pay off, but they’re not going to put others out of business that have a stronghold in the market,” said DH Brown’s Partridge. He added that Compaq will likely retain its long-time role as the market leader.

Nevertheless, IBM is confident its scalable systems will give it an edge as researchers delve into compute-intensive proteomics work. As Caroline Kovac, vice president of IBM’s life science solutions software group put it: “The best-of-breed technology is there in our products. We need to get it to our partners and customers more effectively, and that’s what this $100 million investment is all about.”

—Leslie Doyle

Filed under

The Scan

Push Toward Approval

The Wall Street Journal reports the US Food and Drug Administration is under pressure to grant full approval to SARS-CoV-2 vaccines.

Deer Exposure

About 40 percent of deer in a handful of US states carry antibodies to SARS-CoV-2, according to Nature News.

Millions But Not Enough

NPR reports the US is set to send 110 million SARS-CoV-2 vaccine doses abroad, but that billions are needed.

PNAS Papers on CRISPR-Edited Cancer Models, Multiple Sclerosis Neuroinflammation, Parasitic Wasps

In PNAS this week: gene-editing approach for developing cancer models, role of extracellular proteins in multiple sclerosis, and more.