Sun Microsystems said last week that it is supplying Mount Sinai Hospital’s Samuel Lunenfeld Research Institute (SLRI) in Toronto with more than $5 million in hardware, software, and storage to support development of the Biomolecular Interaction Network Database (BIND) — a contract that was originally slated to go to IBM.
In the first part of a “multi-stage partnership” with the Blueprint Initiative, which manages the BIND project at SLRI, Sun’s computational biology business development manager Stefan Unger said that Sun has already delivered six Sun Fire V880 servers, four Sun Fire 280R servers, two Sun Fire V480 servers, and 24 terabytes of StorEdge disk arrays to support the growing database. In addition, Sun is supplying software products from its Open Network Environment portfolio to design, implement, and manage BIND’s web services, including JAXB (Java architecture for XML binding) to manage submissions.
Christopher Hogue, the scientific leader of the Blueprint Initiative, noted that the project considered solutions from several hardware vendors, and “elected to use Sun because they were the most responsive in getting our operation going.” As an example, Hogue noted that the lack of a user-friendly Java interface for BIND has been “a sticking point for people who wanted to enter data,” but he was reluctant to move 90,000 lines of C/C++ code to Java. Using JAXB, however, it took only four weeks to create the Java interface, putting the release of the new data submission system “significantly ahead of schedule.”
Hogue has good reason to want pieces of the project running ahead of schedule now — it was supposed to have its hardware infrastructure in place over a year ago. BIND’s initial development occurred at SLRI, where Hogue began designing the database while an academic researcher. After taking the position of CIO of MDS Proteomics, Hogue found a temporary home for the public version of the resource at the Center for Molecular Medicine and Therapeutics at the University of British Columbia. MDS and IBM, along with a number of other partners, established a public/private non-profit entity in May 2001 called Blueprint Worldwide to more rapidly develop the database and provide a dedicated infrastructure for it. The two companies agreed to jointly contribute $9 million in cash and equipment.
BIND was built with IBM’s DB2, but the eServer systems and Shark disk storage planned as part of the original Blueprint Worldwide agreement were never delivered. Hogue, now back at SLRI, indicated that the involved parties could not agree on certain terms of the equipment deal with IBM, which led to the year-long delay and eventual termination of the hardware portion of the contract.
IBM spokeswoman Theo Chisholm explained via e-mail that because SLRI offered to host the database early last year, the Blueprint Worldwide consortium decided that “it made sense to take advantage of an existing infrastructure, facilities, administrative staff, rather than building from scratch, leasing, hiring, etc. With the move, Blueprint Worldwide lost its purpose.”
Blueprint Worldwide has since been dissolved and reborn as the Blueprint Initiative at SLRI, which has secured a total of Can$28 million (US$18.4 million) in funding from the Canadian government, Genome Canada, and the Ontario R&D Challenge Fund. Hogue said he is making up for lost time, and hopes to have around 40 curators working on the project within two years. Through a combined process of manual literature curation, text mining, and third-party submissions, Hogue said that BIND should soon have up to 80,000 entries. Hogue’s goal is for BIND to become the de facto repository for protein interaction data within the next two years.
IBM, meanwhile, “has a strong relationship with SLRI/Mount Sinai and we continue to be actively involved with BIND,” said Chisholm, who added that the company never had an exclusive relationship with Blueprint. “The more stakeholders in this effort, the better,” she said.