Bioinformatics startup PanGenX is betting that its semantic approach to data integration will ultimately help drug developers and diagnostic firms develop more personalized treatments.
The firm, based in Auburndale, Mass., intends to help pharma and diagnostic customers integrate proprietary and public data in "meaningful ways" and then run analyses that reveal information on individuals' responses to treatment, Jeremy Sohn, PanGenX's chief operating officer, explained to BioInform this week
The company markets the PanGenX Knowledge Base, which uses a linked data approach to aggregate pharmacogenetic data, results from peer-reviewed literature, health outcomes, and claims data. It relies on a set of proprietary ontologies that specify scientific, clinical, and business concepts and relationships to structure that data for querying and analysis.
Among the data included in the knowledgebase is a version of the National Center for Biotechnology Information's Single Nucleotide Polymorphism Database, or dbSNP, that improves on the publicly available resource, according to PanGenX.
The company's version, dubbed LD-SNP, offers a cleaner, normalized version of the public resource, which makes it possible to find additional variants that aren't currently associated with some genes in dbSNP, Sohn explained.
For example, he said, dbSNP's record of the DPYD gene — which encodes for the dihydropyrimidine dehydrogenase enzyme that is involved in a metabolic pathway — reports about 6,000 variants for the gene, while PanGenX's approach identified nearly 12,000 variants.
This data, combined with the company's semantic-based approach to analysis, makes it possible to map and compare polymorphisms between different individuals as well as calculate distributions of variants from a gene, drug, or disease perspective, PanGenX said.
The product also includes so-called PURL (Persistent Uniform Resource Locator) Nexus, or Plexus, technology, which lets the system combine data from the knowledgebase with remote customer data.
Currently, PanGenX offers two versions of its knowledgebase under a software-as-a-service business model: PanGenx-KB, priced at $250,000 per year for a site license, is geared toward pharma and diagnostic companies; while PanGenX-KB Professional, priced at $35,000 per year, is meant for academic and commercial research groups and labs.
Both versions include access to LD-SNP, curated peer-reviewed literature, and analytical tools. The higher-priced version also offers access to third-party analytics engines, such as those offered by GNS Healthcare and Mathematica, through an application programming interface which is still being developed and should be available later this year, Sohn said.
Other planned updates include visualization tools that will allow end users to customize the sets of queries and reports that they would like to generate, he said.
PanGenX is targeting its wares at pharmaceutical companies who can use its tools to identify likely drug candidates in their research and development pipelines, select patient populations for clinical trials, and collect and analyze pharmacovigilance data. The company believes it could also find use in pharmacogenomics research projects or companion diagnostic development.
The company currently doesn’t have any customers but it is in the process of negotiating agreements with a number of big pharma, gene sequencing companies, and academic medical centers, Sohn said.
The firm also believes that its offering will be picked up by hospitals and health systems that could use its content to improve their clinical care guidelines to make them more tailored to specific populations, enable decision-support tools, and avoid adverse drug events.
Another market that PanGenX hopes to tap into would be insurers and pharmacy benefit managers who could use its resources to analyze comparative effectiveness, identify differentiated outcomes by subpopulations, and improve formulary selection.
PanGenX believes it doesn't have any direct competitors in the marketplace, though Sohn acknowledged that some companies offer products that are complementary to parts of its own, adding that the firm sees these groups as partners rather than rivals.
For example, companies like GNS Healthcare and Mathematica offer more in-depth analytics engines, whereas PanGenX has a much simpler interface and intends to link to these detailed systems, Sohn said.
Also, companies like IBM and Oracle offer big data warehouses that are meant for storing and integrating data in a single location, while PanGenX's approach relies on semantic technologies to link datasets regardless of where they are housed, he said.
Finally, firms like Thomson Reuters and Elsevier gather peer-reviewed literature, which PanGenX also includes in its knowledgebase, he said.
These latter companies are more likely to encroach on the market share PanGenX hopes to capture.
In 2010, Thomson Reuters bought privately held pathway informatics firm GeneGo to provide its clients with a better understanding of disease mechanisms and potential therapies (BI 12/3/2010) and (BI 05/06/2011).
More recently, Thomson Reuters released a set of APIs that let researchers combine information from its commercial drug development databases with their own internal data as well as information from public sources (BI 03/2/2012).
Meanwhile, Elsevier purchased pathway analysis firm Aridane Genomics to complement its chemistry, pre-clinical, and clinical workflow solutions (BI 12/9/2011).
Have topics you'd like to see covered in BioInform? Contact the editor at uthomas [at] genomeweb [.] com.