Curoverse plans to use funds from a recent $1.5 million seed round to develop Arvados, a free open-source computational storage platform used for biomedical research projects at Harvard Medical School that will form the basis of a series of commercial products that the company is creating for the bioinformatics market.
The first of these products, the company said, will be a platform-as-a-service offering that will provide clients with Arvados' computational management and data storage capabilities for a yet-to-be disclosed subscription fee that would cover external costs such as installation on cloud platforms like Amazon Web Services and OpenStack as well as any system integration, maintenance, and management costs, Adam Berrey, Curoverse's CEO, told BioInform.
Curoverse has set up a private six-month beta that will allow participants who apply via its website, to test the platform ahead of its commercial release in summer 2014. It plans to develop and release other commercial products based on Arvados but it is not disclosing details about these at this time.
Separately, Curoverse will maintain a free version of Arvados that will offer the same features as the commercial product but will not include the added operational support. The company also has an agreement to maintain and upgrade HMS' version of the software — which runs on a private two-cluster system made up of 500 cores and 300 terabytes of storage — adding new features as they are developed.
Arvados — named for a fictional planet from the Star Trek: The Next Generation television series — was originally developed in George Church's lab to support the Personal Genome Project (PGP) by a team of scientists and software engineers led by Alexander Wait Zaranek, PGP's informatics director and also Curoverse's scientific director.
Its components include a content-addressable distributed storage system that stores data in a flexible, low-cost fashion, and a computational management system for running large-scale computations in parallel, Berrey explained. Arvados infrastructure makes it easy, he said, to reproduce complex computational operations ensuring more consistent results, and its security system complies with current regulatory standards. It "free[s] informaticians to write whatever algorithms they want" — for things like sequence alignment and variant calling and analysis — and then provides them with supporting infrastructure to run those pipelines easily on petabyte-sized datasets, Berrey said.
Legally incorporated in 2010 under the name Clinical Futures, the company changed its name Curoverse — an amalgamation of the Latin verb 'to cure' (curare) and the suffix –verse — to better reflect its focus as an infrastructure provider for biomedical research, molecular diagnostics, and eventually precision medicine; and to avoid being misconstrued as a testing lab or a clinical data interpretation firm. Since its inception, Berrey said, Curoverse has focused on developing its technology and developing a business model that addresses customers' needs and distinguishes it from competing vendors.
The model it has adopted isn’t one many bioinformatics vendors use, although it’s well established within the larger information technology community. Companies like Red Hat, for example, have built successful businesses around developing and providing a number of free open-source software products alongside subscription-based user support, Berrey noted. Curoverse believes it is a good fit for the bioinformatics market, striking the right balance between demands for flexibility and transparency, and building a profitable commercial business.
By offering both free and subscriber-supported versions of the system, Curoverse provides an alternative to prepackaged traditional software for researchers that prefer to control their own pipelines and computational infrastructure, and caters to clients who want transparent systems but prefer to outsource the accompanying operational burden.
Plus, since it has investors’ buy-in, Curoverse has the funds to keep Arvados afloat, pursue multiple lines of development, and to provide a level of customer support that large-scale open-source software projects that are maintained by single academic institutions often can't manage, Berrey said. Its backers include Hatteras Venture Partners, Point Judith Ventures, Common Angels, MassVentures, and Boston Global Ventures.
Initially, Curoverse will target customers in the clinical research market, specifically those working in bioinformatics cores operating at academic medical centers and universities. It will then expand into the molecular diagnostic market targeting labs that are developing clinical tests. Eventually, it hopes to tap into the precision medicine market, Berrey said.
"The way we envision it is that hospitals will run an EMR and a biomedical big data platform [Arvados] side by side" with that latter storing all of the sequencing data associated with patients in the hospital’s system, he said. Furthermore, "on top of that biomedical big data platform, there will be a lot of applications" for things like pharmacogenomics that will "really provide decision support," he said.