IBM, Cellomics Partner on ‘Turnkey’ HCS IT Infrastructure
IBM and Cellomics said at DDT last week that they have co-developed an IT infrastructure designed for storing and managing data from Cellomics’ High-Content Screening cell-based assay platform.
The companies said that the first of these “turnkey” IT systems has already been installed at EPOPlus, a biotech firm developing erythropoietin therapies.
The infrastructure, called the HCi Appliance, includes Cellomics’ High Content Informatics software with servers, storage, and management software from IBM in two configurations: Professional and Enterprise. The Enterprise version includes two IBM eServers with dual 2.8 GHz processors and 1.0 GB RAM. One server comes installed with the Cellomics Store database, while the other server runs the Cellomics HCS Application Server and IBM’s Tivoli storage manager. In addition, the setup includes 10 146.8 GB 10K disks for 1.2 TB of usable storage, along with a modular tape library for backup storage. The system also comes with Cellomics’ vHCS Discovery Toolbox, a suite of tools to manage, analyze, and visualize high-content screening data. The companies did not provide details on the Professional configuration.
Kip Peterson, worldwide segment executive for devices and diagnostics at IBM healthcare and life sciences, said that the Cellomics partnership is a “perfect fit” for his group’s charter within IBM, which is to “identify innovative new devices and build partnerships with them to create a joint solution.” Peterson said that “domain experts” within the IBM healthcare and life science group targeted Cellomics as “the leader in high-content screening technology” — a market that the company has identified as a promising step toward “closing the gap” that currently exists between genomic research and medical applications.
Mark Collins, senior product manager for Cellomics, told BioInform that the relationship with IBM will not impact a storage partnership with EMC that Cellomics began in December [BioInform 12-15-03]. The EMC solution, built on the company’s Centera system, is for “static storage,” Collins said, while the IBM system is designed for data that researchers will want on hand. Collins said that Cellomics had previously provided “an appliance that we built with Dell hardware,” but this offering proved “difficult to support.” IBM, Collins said, “has a strategic interest in the life science market,” and is therefore able — and willing — to address end-user IT demands, while “Dell was just about selling boxes.”
Collins noted that data management and storage is a crucial aspect of cell-based assay experiments. The amount of data generated per assay can vary widely, but Collins estimated that a “typical” experiment using 100 384-well plates can generate around a quarter of a terabyte of data in one shot.
Collins said that Cellomics is currently expanding its informatics offering to work with other cell-based screening platforms through a collaboration with GE Healthcare that it announced last month [BioInform 07-05-04]. In addition, he said, the company’s informatics team is currently addressing the “next big challenge,” which he described as integrating cell-based screening data with genomics, proteomics, and cheminformatics data in order to get “a picture of what’s going on in systems biology.”
Cellomics may have a head start in the HCS informatics market, but it is no longer the only player in the space. Image informatics provider Scimagix launched a data management product for cell-based assays called CellMine at DDT. Bob Marchmont, director of marketing at Scimagix, said that the company plans to add more analytical capabilities to the product, which is currently focused on data management.
Nonlinear Partners with Matrix Science as it Readies Launch of Two New Products
Nonlinear Dynamics announced a collaboration with Matrix Science at DDT last week as the first step in several product launches planned for late 2004 and early 2005.
Nonlinear said that it will integrate Matrix Science’s Mascot protein identification software with its upcoming “Pro” line of protein informatics tools.
The first of these, ProMST, is scheduled for launch in the fourth quarter of 2004. The system is a modular protein informatics workflow platform designed to manage experimental set-up, 2D gel image storage and analysis, data mining, and mass-spec-based protein identification. ProMST is built on a central database that integrates all the data generated during the workflow. Nonlinear said that the agreement with Matrix Science will allow users to link directly to a Mascot server from ProMST. John Spreadbury, group sales and marketing director of Nonlinear Dynamics, said the product will be the first protein informatics platform to integrate 2D-gel and mass spec analysis.
Another product in the works, Progenera PI, is a client/server platform that will run on a distributed computing architecture. Mark Bennett, general manager, Asia Pacific, for Nonlinear, said that the system will allow researchers to submit a query to a database on a central server, which then gets “farmed out” to available resources in the network, with the results coming back to the database and then the user’s desktop. Spreadbury said that the distributed computing model will offer “flexible licensing,” which should keep the total cost of the system well below that of the company’s Progenesis enterprise software system, which is typically over $100,000.
The company said it also plans to upgrade its Phoretix 1D software in the next 12-18 months.
Bio-Rad Partners with Strand, others, to Beef up ADME-Tox Informatics Suite
Bio-Rad last week announced a stable of new partners — Strand Genomics, Equbits, Lighthouse Data Solutions, and Sierra Analytics — in an effort to expand the ADME-tox prediction capabilities of its KnowItAll informatics system.
The company will collaborate with Strand Genomics to integrate Strand’s predictive human pharmacokinetic models in the KnowItAll platform. Kas Subramanian, CSO of Strand, told BioInform that the company began its ADME-tox program in mid-2002. He said the company offers five pharmacokinetic models for licensing — plasma protein binding, bioavailability, volume of distribution, elimination half-life, and rate of absorption — and also offers custom PK modeling services.
In the Equbits collaboration, Bio-Rad customers will be able to use Equbits’ Foresight SVM-based modeling application to build their own predictive models using proprietary experimental data — so called “local” models that can complement the “global” ADME-tox prediction models in the KnowItAll system, Bio-Rad said.
Bio-Rad will partner with Lighthouse, meanwhile, by hosting Lighthouse’s ADME Index database of experimental ADME data for 350 compounds, including FDA-approved drugs and non-approved compounds with drug-like properties. All experimental data for the database has been generated at Lighthouse’s laboratory.
Finally, Bio-Rad said it will modify KnowItAll so that a list of possible metabolites can be directly exported into Sierra Analytics’ Apex software for metabolite identification. Apex compares the likely mass spec fragmentation of the possible metabolites with LC/MS or LC/MS/MS files from a number of instrument vendors to predict the “metabolic fate” of the potential drug, Bio-Rad said.