A leading precision medicine academic medical institution needed a data management solution to store high volumes of whole genome sequence information and link it with clinical data for patient treatment and research purposes.
Slow ingest speeds and analytic performance were getting in the way of bioinformatician productivity. In addition, the chief enterprise architect was looking for a solution that would evolve as science and users demanded – both in terms of the type of analysis and the variety of complex data being integrated. Of course, given the volumes of personal health information, privacy, security and governance was a key requirement without the proliferation of identified and de-identified datasets.
- Ingest large genomic files at wire speed
- Look up information in a fraction of a second
- Build cohorts and understand treatment response options for patients with similar genotype/phenotype
- Analyze multi-structured data at scale and speed
- Maintain privacy, security and governance across many stakeholders and collaboration partners
PHEMI set up a secure, PHI-ready system in Microsoft Azure in under two weeks and the customer started to load data. PHEMI Central automatically de-identified data on ingest, and access policies were set to control who could access PHI, and who could only see de-identified data.
Bioinformaticians used a rapid prototyping environment and Global Alliance for Genomic and Health APIs to quickly look up genotype/phenotype cohorts. Moreover, users could plot Kaplan Meier survival curves, consider personalized treatment response options and provide input to the tumor review board. Based on user feedback, data models were easily modified and new data types were added on the fly.
PHEMI Central was able to ingest data 12 times faster than the existing relational database solution, bringing hours down to minutes. Information lookup times were reduced to a fraction of a second, and full genome-wide association studies could be performed in less than two hours.