Clustered storage enhances prostate cancer and proteomics research

The researchers at Cedars-Sinai Medical Center’s Louis Warschaw Center for Applied Molecular Medicine in Los Angeles are highly regarded in the field of personalising existing cancer therapies and developing new drug therapies.

The Center uses mass spectrometers to analyse blood samples from tens of thousands of patients. From these samples they search for patterns of proteins that differentiate patients who respond to therapy, versus those who do not — a problem tantamount to looking for a proteomic needle in a haystack.

High-performance computing clusters are used to sort through, analyse and correlate their proteomic and clinical data. However, a proteomic analysis of a single drop of blood can generate more than 60 Gigabytes of data and with thousands of samples stored, the Center needed a storage solution that would scale up to hundreds of thousands of sample sets to allow for better analyses.

In mid-2006, the Center began evaluating a number of new storage solutions to complement its high-performance computing cluster. After an extensive evaluation period, Cedars-Sinai selected a new clustered storage system from Isilon Systems.

The Isilon IQ provides a clustered storage environment where each node is independent and intelligent. No one node stores all the file data and each cluster has copies or parity bits to allow it to reassemble any file in the event of the loss of any individual disk, node or multiple nodes. Nodes connect together like individual bricks in a wall, scaling from tens to thousands of terabytes.

“The clustered storage architecture allows us to meet our most demanding performance requirements, while scaling in a modular and highly cost-effective manner”, says Parag Mallick, Director of Proteomics for Cedars-Sinai Center for Applied Molecular Medicine.

The system also offered the unparalleled ability to add data points to its studies while simultaneously keeping all information available for further analysis as patterns are detected. Because the Isilon solution uses standard Gigabit Ethernet connectivity to interact with the computing cluster and its own InfiniBand Cluster Interconnect for high-speed, inter-node connection, the system proved remarkably easy to implement with no significant software changes other than pointing systems at the new storage pool.

“As we advance our research by expanding the number of processors in our high-performance computing clusters, Isilon allows us to flexibly add performance, capacity or both in ways that are simply not possible with other storage technologies,” Mallick adds.

While the Cedars-Sinai Center for Applied Molecular Medicine’s primary focus is to target therapies for prostate cancer, the clinical, research and technology innovations it develops could have far-reaching effects.

Their research and genome-scale methodologies, coupled with the emergence of comprehensive digital patient records, have the potential to benefit other medical research. With its data reliably stored on the storage cluster, Cedars-Sinai researchers are able to continuously expand their studies so that they can find new ways to use their existing data with high-performance computational analysis.

These data repositories, which are greatly expanding due to the emerging trend of standardised medical records, promise to vastly increase the information that can be leveraged for life-saving research initiatives.

The adaptive nature of the Isilon IQ allows Cedars-Sinai’s researchers to easily install, maintain and scale their storage and to accommodate adding more than one terabyte of new data per day. The ability to add capacity and performance linearly or independently in a 'pay-as-you-grow' expansion model frees the Center's technical staff to focus on more important matters such as expanding their collection, refining their drug therapies and working towards new treatments.