If you were handed $6 million to ramp up your biomedical research infrastructure, how would you spend it? A number of universities across the US have been given the opportunity to answer that question through the NIH’s $144 million Biomedical Research Infrastructure Network (BRIN) program, which doles out awards to state universities traditionally underserved by NIH grants, and the responses are about as varied as the program’s participants.
BioInform caught up with a number of state universities that were awarded $6 million apiece over three years in October 2001, in order to determine where bioinformatics falls into the mix of resources the BRIN program was created to support. The awards were granted through the NIH Institutional Development Award (IdeA) program, which serves 23 eligible states and Puerto Rico, in which investigators accounted for only eight percent of the total number of NIH research grant applications in 1998. The BRIN program was initiated to balance this inequity by giving these states a stronger biomedical research foundation.
Each BRIN network is required to support a bioinformatics core, along with an administrative core and at least one other core in an area left to the institution’s discretion. According to NIH documents describing the program, “the Bioinformatics Core is essential to provide investigators access to the technical expertise and data management and analysis tools required for competitive, multidisciplinary biomedical research.” The core must provide access to bioinformatics and other related tools for investigators in the network, promote informatics training and education, and develop methods for multi-center research and resource sharing as well as secure and confidential data sharing.
So how well are the BRIN institutions meeting these requirements so far? Of the 14 BRIN bioinformatics core directors BioInform spoke to, the majority said that their bioinformatics infrastructure was at least halfway toward completion, only one year into the three-year grant period (see table). Of course, the definition — and cost — of a complete bioinformatics infrastructure varied quite a bit. Depending on a number of factors, including existing computational resources, the budget set aside for the bioinformatics core ranged from a low of $25,000 at the University of South Carolina (0.4% of the total $6 million BRIN grant) to a high of $2.4 million at Louisiana State University (40% of the total grant).
For those directors who broke down their bioinformatics spending, hardware emerged as the clear budgetary priority, dwarfing bioinformatics software spending in all but a few cases. BRIN grants have funded the construction of at least three Linux clusters so far, and three more are planned in the near future. Four bioinformatics cores have installed SGI systems.
Not surprisingly, most bioinformatics cores are relying on public domain software resources, but this decision isn’t purely based on cost considerations. According to Marcella McClure, bioinformatics director at Montana State University, Bozeman, “there are no commercial packages that out-perform free software, so there is really no point in purchasing such products.”
Other programs, such as the Universities of Oklahoma and South Dakota, are devoting their resources to writing new bioinformatics applications. Said Tyrell Conway of the University of Oklahoma, “There are no adequate commercial solutions for microarray database needs. Thus, we are writing one.”
Of those universities that have opted for commercial packages, InforMax’ Vector NTI is the clear leader, with five programs already licensing or planning to license the sequence analysis software.
BRIN bioinformatics resources are generally open to all biomedical researchers within their states, but the current number of active users tends to hover between 10 and 15. Most core directors anticipate that this number will increase two-fold or more over the next five years, however, as infrastructures mature and training programs begin to bear fruit.
“There are a number of educational programs under development and as students are trained in bioinformatics and these programs and resources become better known, we expect an explosive increase in the use of these resources,” said Steve Jennings, bioinformatics director at the University of Arkansas, Little Rock. Most directors were willing to project growth out to five years, although ten years was a bit much to ask — “light years in bioinformatics,” according to Conway.
Effective data sharing across distributed research teams remains the largest obstacle to achieving a fully operational bioinformatics infrastructure, according to most BRIN bioinformatics directors: Five identified this as their top priority going forward, and many are currently reviewing technologies for collaborative research, including Access Grid technology, a community-based development effort that supports distributed meetings, remote visualization, and distance education over a computational grid (www.accessgrid.org).
For the most part, BRIN bioinformatics support is limited to a small core staff of one to four people, charged with meeting the computational needs of researchers working on everything from microarray analysis to phylogenetic studies to protein folding and molecular modeling. Thus, while the BRIN grants gave underfunded state universities a much-needed boost, there’s always room for further improvement. Noted Austin Hughes of the University of South Carolina, “If there is a BRIN renewal, I will fight for more money for bioinformatics.”