The Protein Structure Initiative — a 10-year effort sponsored by the National Institute of General Medical Sciences that kicked off in 2000 with the goal of determining the three-dimensional structure of thousands of proteins — has not yet identified a host for a project database that was expected to launch this fall.
Nevertheless, NIGMS still expects to launch the database, called the Structural Genomics Knowledgebase, by next fall.
A request for applications for the Knowledgebase was originally issued in December 2005, and the $2.5 million award was expected to be granted this September.
“We did receive applications, and the applications were reviewed. However, following the review of applications, there were no awards made,” Jerry Li, program director at the NIGMS Center for Bioinformatics and Computational Biology, told BioInform this week.
“Right now, we’re still trying to decide what would be the best approach,” Li added. “We’re considering re-announcing the RFA, but that hasn’t been decided yet.”
Li declined to provide further details on why the grant was not awarded, but John Norvell, PSI director, said that NIGMS is “still highly motivated to have a Knowledgebase. It’s very important, and we’re considering another RFA or other options right now.”
Norvell added, “The intent is to have a Knowledgebase established within the next year.”
More Than Just Structures
According to the NIGMS RFA, the Knowledgebase will serve as a “central information hub” for the 10 research centers funded under the PSI project as well as for the broader research community.
Both Li and Norvell stressed that the Knowledgebase is not intended to compete with the Protein Data Bank, the primary resource for experimentally derived protein structures managed by the Research Collaboratory for Structural Bioinformatics. PSI participants are required to deposit solved protein structures in the PDB within four weeks of their completion.
“The goals of the PDB and the PSI Knowledgebase are different,” Li explained, noting that the PDB is “very focused” on structures, while the Knowledgebase will be “more comprehensive” and include all the information associated with high-throughput structure-determination experiments.
The PSI database will include links to 3D structures in the PDB, “but in addition to that, the Knowledgebase is charged to make the other information and products produced by the structural genomics centers available to the community,” Li said. “That includes the expression clones, the protein production methods, the crystallization procedures, and NMR procedures that should be more useful.”
The Knowledgebase will also include links to the PSI Materials Repository, a warehouse for PSI-generated clones that launched this week.
The database is a key component for Phase 2 of the PSI effort, called PSI-2, which began in 2005 following a five-year pilot phase.
The goal of the pilot phase, Norvell said, “was just to show that this could work. And it did.” The goal over the next five years, he added, is to “produce a lot of structures and data related to producing the proteins, and in particular to make this information available and useful to the whole biological community so that researchers can more readily incorporate structures into their thinking and their planning, and the Knowledgebase is a key component of that.”
During the pilot phase, the RFA noted, much of the information related to experimental procedures was captured and maintained by individual research centers in their own formats. “Centralization and standardization is needed by end users for easier cross-mining,” the RFA stated. “A user-friendly integrated interface of information repository and web portal is imperative for PSI-2.”
“We did receive applications, and the applications were reviewed. However, following the review of applications, there were no awards made.”
During the pilot phase, the PDB hosted two databases to help centralize PSI information: TargetDB, a resource for coordinating target selection across the PSI centers; and the Protein Expression Purification and Crystallization Database, or PepcDB, for storing experimental protocols.
NIGMS intends to roll the current functions of these two databases into the new Knowledgebase.
The primary role for the Knowledgebase will be as a web portal and information repository for all resources related to the PSI. The resource is also expected to provide “a system of structural and functional annotation for the PSI-generated structures,” according to the RFA. While the Knowledgebase will not be responsible for performing annotations, it should “provide a user-friendly environment for deposition and query of the annotations generated by PSI projects.”
Another role for the Knowledgebase will be to help researchers computationally model protein structures. In August, NIGMS awarded two three-year, $2.1 million grants to create PSI molecular modeling centers at Fox Chase Cancer Center and the Burnham Institute for Medical Research. The goal of these centers is to improve homology modeling methods using information generated by the PSI experimental centers.
The Knowledgebase will host “a limited number of high-quality models of proteins” and will provide links to modeling tools and servers developed by other groups, according to the RFA.
Li said that NIGMS has not outlined any specific guidelines for how the Knowledgebase handles computational models. “That’s clearly an objective, but exactly how it should be implemented we haven’t specifically laid out,” he said. “It’s up to the applicants to propose how to best support that kind of project and activity in the community.