By Malorye Branca
January 13, 2003
| Three years ago, Clayton Naeve experienced every research director's dream: His annual budget tripled overnight. Naeve is director of what was formerly called The Biotechnology Center, at St. Jude Children's Research Hospital. The windfall came from an anonymous donor who pledged $5 million a year to the Center — since renamed the Hartwell Center for Bioinformatics and Biotechnology — for the next five years.
"[The benefactor] has a particular view that computers will revolutionize biomedical research," Naeve says. "He thought that by investing in St. Jude's and, in particular, in the Center, he would impact everyone's research program and would raise our science to the next level ... It was an astute observation."
The donor's timing was good. Many of St. Jude's researchers were already jumping into advanced, data-intensive genomic technologies, including microarrays, genotyping, and proteomics, to help understand the molecular pathogenesis of serious childhood diseases. The number of users of the Center's genetic analysis tools jumped from barely a dozen to more than 800 today.
Before the donation, the Center offered high-throughput sequencing, macromolecular synthesis, and bioinformatics resources. Suddenly, Naeve had to determine the best way to ramp up the Center's level of service and facilities. Best of all, some important things could be added at once. "Our benefactor allowed us to compress our strategic plan," he says.
Naeve focused on three areas: building a solid IT infrastructure, bringing in key missing tools, and upgrading the spot from where most of the data was breaking loose: gene expression analysis. Within the past three years, the Hartwell Center has added:
- A cDNA microarray facility, which has generated more than 1,000 spotted arrays
- An Affymetrix microarray lab, which produced about 3,500 GeneChip data sets in the first two years of operation
- A proteomics lab capable of doing 2-D gels and featuring robotic spot picking, automated in-gel digestion, sample spotting workstations, five mass spectrometers (including an Applied Biosystems 4700 Proteomics Analyzer), and a Biacore system for surface plasmon resonance (studying molecular interactions)
- A new data center
The group has also been expanding its suite of informatics tools (both home-built and store-bought).
In addition to a high-caliber workforce, the Hartwell Center benefits from a priceless St. Jude's asset: a repository of high-quality, well- annotated tumor samples. According to Jim Downing, chairman of the pathology department at St. Jude's, the latest methods in tumor analysis "involve expensive, specialized machines and an increasing amount of computer horsepower. The Hartwell Center offers us economy of scale, because researchers don't have to learn each step of every process to get an experiment done."
Budget windfall: A donor's gift has enabled the Hartwell Center's Clayton Naeve to fortify IT.
Persuasive evidence for the effectiveness of the Center can be found in the results of a recent collaboration between Hartwell staffers and Downing's group on a study of acute lymphoblastic leukemia (ALL), the most common type of cancer in children. About 80 percent of children with this disease are cured, but there is a wide range in patient responses to existing therapies. The treatment plan is largely determined by information on the precise subtype of ALL in each patient. But diagnosing the subtype is a complicated process.
Downing's group used gene chips to see if ALL patients could be separated into subtypes based on gene expression patterns. The results were published in the March 2002 issue of Cancer Cell. "It's a wonderful paper," says Dietrich Stephan, associate director at the Research Center for Genetic Medicine, Children's National Medical Center, in Washington, D.C. "It gives you a very broad perspective of what is behind the variability we see in this disease."
The study evaluated thousands of genes in samples from 360 children with ALL. The authors associated discrete gene expression profiles with the six most important subtypes of leukemia. Gene expression analysis may thus point to better and easier ways to determine an ALL subtype.
"It's probably the largest gene expression study done to date," Downing says. Naturally, the larger the sample size, the more likely that the results are significant. And although this is just a taste of the power of applying gene expression analysis to cancer, it's a promising beginning.
"It was because we had the Hartwell Center that we could initiate and complete that study in one year," Downing says. It's not just the quality of the tools they have, either. "Today, instead of looking at three genes in about 100 patients, we're looking at 30,000 genes in hundreds of patients. That's 4 million to 6 million data points. There is no way to do that without sophisticated algorithms, and there are hundreds of ways to analyze it. You need expert guidance."
Riding the Data Wave
Given the pace of the field, St. Jude's is constantly evaluating the Hartwell Center's future needs, Downing says. In the chip arena, for example, the ability to do genomewide studies is having a profound impact. "It is giving us [not only] a tremendous amount of data but also new hypotheses," Downing says.
|Hartwell Center System Highlights
|Hardware, software and more...
All this experience with gene expression should pay off in other fields — notably, proteomics and genotyping, where St. Jude's has already made its mark (see "St. Jude's Test Makes It Better," Bio·IT World
, Sept. 2002, page 56).
But Naeve isn't resting on his laurels. "I would not say we have solved all the problems," he says. "Like everyone else, we are struggling with managing this data, and there are all kinds of issues." For example, as the tools evolve, the characteristics of the data change. "With gene expression, we collect 2,000 or 3,000 files, then we have to go back and reanalyze so we have consistent data."
The Center has several relational databases, each with a particular type of data. "Now we're thinking of a federated database," Naeve says. Naeve is also eyeing protein chips, and the Center is adopting the ICAT (isotope-coded affinity tag) approach to eliminate the need for 2-D gels.
Like many of his peers, Naeve is also puzzling over how to integrate so much data. "There are 15 ways to do this," he explains. As the amount of data grows, the idea of integration grows increasingly appealing. St. Jude's already has 0.5 terabytes of gene expression data, it sequences more than 40 million bases of DNA per year, and its new ABI TOF-TOF mass spectrometer can analyze 1,000 spectra per hour. So the pressure is on.
"The challenges in this field are immense," Naeve admits. With the help of "friends" such as the donor who stepped up for the Hartwell Center, St. Jude's should be able to keep riding that data wave, rather than being drowned in it.