Clinical OMICS

JAN-FEB 2017

Healthcare magazine for research scientists, labs, pathologists, hospitals, cancer centers, physicians and biopharma companies providing news articles, expert interviews and videos about molecular diagnostics in precision medicine

Issue link:

Contents of this Issue


Page 30 of 47 January/February 2017 Clinical OMICs 29 "The nice thing about cloud com- puting is the underlying companies like Amazon, Google, Microsoft, and others is that while they don't have unlimited capacity, it is pretty elastic," says Spielman. "Remember, [Amazon] was built for Christmas Day. If you are Roche or Memorial Sloan Kettering, you are really only going to be able to build so much capacity internally and the amount of that data, while not infinite, is quite significant." While companies were attracted to the expanded computing capacity and flexibility the cloud promised, many were reluctant to embrace it early on due to security concerns, Dr. Shaywitz points out. In an industry that uses patient data for clinical trials, and for broader research studies, working in a secure environment is paramount. He heard those concerns when DNAnexus was helping to design PrecisionFDA. "A well-designed platform will address these concerns about keeping data private," he says. "When we were implementing it, there were conspir- acy theories about the FDA wanting to look into people's data. But people see they can upload data and keep it com- pletely private. The last thing FDA wanted was to be responsible for other people's data." Moving to the cloud also alleviates organizations from having to invest in significant expertise in data handling and security, efforts that can take focus away from their core competencies. "Companies like Amazon, and like us, worry about compliance, worry about security," Dr. Shaywitz notes. "Azure and Amazon are living and breathing this every day." Many customers see this expertise as a major benefit. "Security of cloud- based platforms is a huge, huge issue," adds Shon. "But people now realize that [data] is a lot safer in the cloud than on their own local systems." Not so long ago, the ability to deliver pre- cision therapy fell in narrow band. "You could answer very specific questions, just not very well," says Bryan Spielman, evp of strategy and corporate development with Seven Bridges. "You could answer some of the questions for a particular patient, if you knew exactly what bio- marker you were looking for." The poster child for this situation was the companion diagnostic that detected the HER2 mutation for breast cancer and the therapeutic that targeted it, Herceptin. "You look where we were five years ago. You ran a test, and it was one test, on one gene, that led you to one drug," says Chris Cournoyer, CEO of molecular clinical decision support com- pany N-of-One. "Look where we are today," she adds. "We are running 50-gene panels, or 400- gene panels. We are getting multiple mutations and are trying to figure out the interactions among those mutations, which is leading us to combi- nation therapies." David Shaywitz, M.D., CMO of cloud- based genomic interpretation company DNAnexus puts it this way: "Now we have the opportunity to look at every single base at a remarkably affordable cost." That means instead of having limited information—a spotlight of understand- ing on only particular regions of the genome and mutations—scientists are developing a much more complete pic- ture of genetic variants and interactions. "Instead of looking where the light is, you are shining the light on the whole genome. This whole range of possibil- ities results in a much more sophisti- cated understanding of the relationship between genetics and disease, genetics and phenotype," Dr. Shaywitz adds. Today, companies like N-of-One help health systems marry genomic data and information on known cancer-causing genetic variants with the specific health information of individual patients con- tained in their electronic health record (EHR). This allows for a significantly more sophisticated, and nuanced, approach to providing clinical guidance. While combining genomic data and data from EHRs may be the holy grail of precision medicine, there remain signif- icant hurdles. These include integration of the data and the reliability of the data contained in EHRs, which can often be misleading or incomplete, says Spielman. With EHRS "we have a system that was designed for what it needed to be designed for at the time, it wasn't designed for where the puck was going," he says. "But people like to think technology is infinite. It's not. How often do peo- ple change their iPhone? So, hopefully, EHR/EMRs will be improved." But even having lim- ited information from a patient's EHR to pro- vide insight about their past health issues, current health, and medications is better than no data at all, Cournoyer contends, especially when combined with the annotated set of known cancer-causing genetic variants and which medications are approved to fight against these driver mutations. In the future, she believes molecular diag- nostics will bring even more data to the table in the fight against cancer. "In five years, we will regularly sequence the genome. But I think we will augment that with a variety of other tests," Cournoyer says. "We will be aug- menting that with RNA data, protein expression data, and copy-number anal- ysis. We will be bringing all of these var- ious testing [results] to provide a more complete patient profile." n Genomic Data Enabling Drive Toward Precision Medicine Chris Cournoyer, CEO of N-of-One

Articles in this issue

Links on this page

Archives of this issue

view archives of Clinical OMICS - JAN-FEB 2017