CHICAGO – The month-old partnership between bioinformatics mainstay BC Platforms and analytics giant SAS Institute is as much about filling in gaps in each company's product line as it is about data integration.
Zurich-based BC Platforms can pull data directly off next-generation sequencing instruments, something SAS can only do with a lot of custom data management, according to Scott McClain, principal industry consultant for agricultural technology and life sciences at SAS. "They've been doing that longer and have productized those processes, and what SAS is really interested in is getting our customers to the analytics," he said.
BC Platforms generally does not perform deep analytics in the way Cary, North Carolina-based SAS built its business on. "That's our sweet spot, so that literally did create this gravity" that brought the companies together, McClain explained. He presented on the partnership at last week's Bio-IT World Conference & Expo in Boston and subsequently spoke to GenomeWeb.
McClain called the partnership "synergistic" in that both SAS and BC Platforms had a couple of common customers — one in healthcare and one in life sciences — who asked the companies to integrate data for them.
As the firms noted when they announced their partnership April 12, BC Platforms and SAS are looking to provide life sciences and healthcare organizations with curated genomics and other patient data and help customers break down data silos and ultimately improve prediction of health outcomes. The companies said they will "discover new ways to derive insights" from genomics and other patient data, such as imaging and blood chemistry.
SAS is applying to life sciences some of the expertise it has built up in other industries, such as financial services, often using the same statistical analysis software, called JMP. McClain noted that a bank engaged in fraud detection is using pattern analysis on large datasets, much like a genomics researcher might be searching for a single variant in a whole-genome sequence.
The technology is also useful for large-scale polygenic risk scoring, according to McClain. Users can plug Plink, a widely used open-source tool for calculating polygenic risk scores, into the SAS platform through an application programming interface (API), take the data output from SAS to run through Plink, and then SAS can produce visualizations of the results, he said.
Other third-party tools and APIs can enable SAS to provide clinical decision support, though integration with electronic health records might take some custom programming, McClain said.
Another advantage of the partnership is the user experience, according to McClain. People on the front lines of research or clinical practice don't want to have to switch between multiple windows or log into numerous applications when looking for biomarkers, designing experiments, or treating patients.
McClain said that users can build models within SAS and then use what the firm calls the "competition model" to score each one for accuracy to help them pick the best algorithm for a given use case.
BC Platforms customers tend to have teams of bioinformaticists and biostatisticians on staff. "You don't want all those Ph.D. bioinformaticists sitting still," McClain said. BC Platforms can help them obtain datasets so these experts can focus on what SAS offers to help them apply analytics models to answer research questions or build cohorts.
"The data is piling up, and you've got a pipeline of the data into your analytics. BC Platforms helps pipeline that data in," McClain said, then SAS provides the analytics. "People get kind of worried about the scaling problem. We want them to know that the scaling's not a problem," he added.
SAS customers on the healthcare side tend to be integrated delivery networks, while life sciences customers generally are pharmaceutical companies.
McClain said that SAS does not always know how these organizations use their data, but he noticed that the word "multiomics" was popular at Bio-IT World, including in his own presentation. That suggests that researchers throughout healthcare and the life sciences are looking for ways to break down data silos, another longtime buzzword at informatics conferences; indeed, SAS and BC Platforms used that phrase in their partnership announcement.
BC Platforms has years of experience aggregating data, specializing in federating data for analysis. The multinational firm has built BC|Rquest, a global data network for analyzing and viewing aggregated genomic and clinical data from multiple biobanks.
Having both the aggregation and analytics teams at a given organization on the same schedule will accelerate research and discovery processes, McClain explained. It might also give both companies an entrée to a market for community hospitals that don't have bioinformatics teams.
McClain noted that companies that show up at conferences like Bio-IT World have changed their thinking in recent years. It used to be that their intellectual property was how they produced data.
"Now, producing the data is cheap and easy," he said. "Now, the startups that are going to speak at these conferences, what they're not going to tell you is what their algorithm was for sifting the variant out of the tumor cell. That is their IP."
McClain called genomics "foundational fuel" for digital health and precision medicine.