QA: Why IBM Thinks about Links

Listen to the people at IBM long enough, and you’ll start to believe that they truly understand your problems. Now, they say, they’ve got a solution to what’s fast becoming the biggest bottleneck in genomics: data management and integration.

Janet Perna, general manager for data management solutions at IBM, is one of the staunchest advocates of the company’s products. Formerly a math teacher and a competitive swimmer, she began her IBM career 27 years ago as a programmer and is now one of 50 people on Chairman Lou Gerstner’s advisory board.

Perna works with IBM’s new product, DiscoveryLink, which she calls the “game changer” for genomic data integration because it is designed to access data where it’s stored instead of moving and standardizing it in a central location. GT’s Meredith Salisbury put a few question to Perna.

What’s the most pressing data integration need from the genomics field right now? How will it best be addressed?

PERNA: If you think about the biggest challenge today in the whole area, there are really two major problems: large amounts of data and integrating the different types of data — and it’s often geographically dispersed. Public data and private data have to be brought together and integrated. In any industry today the amount of data is growing tremendously — doubling every 18 to 20 months. In this industry, it’s doubling every six months. And if you look at things like proteomics versus genomics, there are 35,000 genes today versus a million proteins; we’re going to see even a greater explosion of the volume of data.

People pretty much focus on massively parallel machines, so you also need databases that can exploit this massively parallel hardware technology to access, store, and manage this huge volume of data.

Why the push in life sciences? How much of IBM’s business comes from the genomics market?

PERNA: This is really about growing and addressing new opportunities. We see a huge convergence here between information technology and biology. If you look at what it takes to run biology experiments today, they don’t take place without information technology; we wouldn’t have the life science industry without computers.

This is an emerging market, and we’re investing here to capture an early stake in this market. I’m obviously not going to tell you how much business we have in this field. We see this as a fast-growth business — in terms of high-performance computing, services, and data.

A Frost & Sullivan report shows the hardware, software, and services sector will be worth $43 billion by 2004. That’s a pretty good chunk of change. The data management component alone is going to be $4.5 billion — it’s very important to us to be out front working with these firms in this industry feeding the market.

Where do you see this industry going?

PERNA: I think what’s going to revolutionize it is this whole notion of federation technology for data integration. What really is inhibiting pharma, the number-one problem for them, is the integration and collaboration. Other approaches have all centered around moving the data into a central place and it doesn’t work. There’s too much data, different formats, it costs too much money, or it may be public or may not be within your company. Federated data technology enables these companies to leave the data where it is and access it in an optimized way.