Towards Fulfilling The Promise Of Genomic Medicine

David ShaywitzI was in eleventh grade when I first discovered The Eighth Day of Creation, Horace Freeland Judson’s wonderful, eloquent, deeply inspiring account of the history of DNA and the origins of molecular medicine.  I’ve been hooked ever since.

My choice of college was driven by the opportunity to study with many of the key players in the story – and sealed when a biochemistry department tour guide pointed to what he said was the exact centrifuge used in the famous Meselson-Stahl experiment.

It was thrilling to learn about molecular biology, and even more exciting to become a molecular biologist, learning how to extract and recover DNA, how to splice it, and how to sequence it.

I spent my first year in graduate school studying tumor suppressor genes, then found myself seduced by the elegant power of yeast genetics, and chose this area for my thesis research.

When I returned to medical school and entered the clinic, I was surprised by what I discovered.  While medicine is often regarded as applied science, I was reminded every day just how limited and fragile our understanding of health and disease really is.  The molecular basis of illness often escapes us, and even in the relatively rare instances where the biology is well understood, a cure can be difficult to come by – consider sickle cell disease (considered the first illness defined at a molecular level), for instance, or cystic fibrosis.

Along with co-authors Dennis Ausiello and Joseph Martin, I wrote in 2000,

“Physicians and physician-scientists have become increasingly concerned with ensuring that the tremendous advances they have seen in basic science find expression in clinical practice.  While an understanding of the genetic basis of disease allows us to consider the development of molecular therapies, we have learned not to underestimate either the magnitude of this undertaking or the extent of preparation required.  Indeed, this endeavor is much more difficult than most have anticipated.

As Goldstein and Brown recently noted, paraphrasing Magritte, ‘a gene sequence is not a drug,’ and although the development of rational therapy for a disease may require an understanding of its molecular basis, the path from mechanistic understanding to clinical treatment is often difficult to define and hard to predict.  Proteins often behave differently in test tubes than in cells, and cells behave differently in culture than as part of a vital organism.  Finally, a patient’s experience of disease reflects more than simply an underlying biological defect.  It is, to quote Eric Cassell, ‘a process inextricably bound up with the unfolding story of this particular patient.’  Thus, the critical question we are now struggling with as physicians and physician-scientists is how to avail ourselves of the advances in molecular biology without losing sight of our primary goal – the care and treatment of our patients.”

In the fourteen years since this was written, our tools have changed, our computational power has improved exponentially, the volume and velocity of data have grown at almost unimaginable rates, yet the fundamental challenge remains the same: how can we translate promising science into improved patient care?

More generally, how can we harness large data flows to improve the human condition? I believe this represents the central scientific challenge — and greatest scientific opportunity — faced by our generation.

The starting and ending point for this vision must be patients.  Data collected from a patient belong to the patient.  With appropriate permissions and rigorous safeguards, these data can be shared in de-identified fashion, in effort to accelerate the sort of rapid knowledge turns Andy Grove has insightfully discussed, and that Josh Sommer of the Chordoma Foundation so poignantly described at the first Sage Bionetworks Congress.

Progress will require brave exploration — memorably championed by Pixar’s Ed Catmull — combined with seamless collaboration, so that, as Cloudera’s Jeff Hammerbacher says, everyone can “party on the data.”

Genomic sequencing data represent a natural foundation, not only providing an essential “parts list” (as Eric Lander has nicely described it), a sense of what you are starting with, but increasingly providing dynamic visibility into important clinical pathophysiology, such as the evolving molecular characteristics of circulating tumor cells, or treatment-resistant bacteria or virions.

I’m especially excited by the opportunities at the intersections of large data streams, loci that, like overlapping academic disciplines, promise to be especially rich sources of novelty and insight.

DNAnexus represents a natural home for these aspiration, offering a compelling, secure, cloud-based data management platform, an enabling tool for any healthcare organization – academic medical center, healthcare system, biopharma company, payor – who recognizes that getting a handle on large healthcare data flows is rapidly becoming table stakes, and that figuring out how to manage and leverage genomic data is a wise place to start.

I’m excited – and feel inordinately privileged – to join the DNAnexus team, and to work with passionate colleagues from throughout the healthcare system, in both the private and the public sector, and explore together how we can move from genome to value, from data to impact, from information to cure.

David Shaywitz, MD, PhD is the Chief Medical Officer of DNAnexus, and the co-author, with Lisa Suennen, of Tech Tonics: Can Passionate Entrepreneurs Heal Healthcare With Technology? (Hyperink Press, 2013).  You can follow him on Twitter: @dshaywitz.


Cloud Computing Insight with Omar Serang

omar serang

Omar Serang, our new Chief Cloud Officer, came to us from Amazon Web Services, where he formed the Enterprise IT Cloud Transformation consulting practice. Before that, he served as Amazon’s EC2 operations engineering manager. We chatted with Omar to get an expert’s view on cloud computing — from how it started to new innovations on their way and where DNAnexus fits in.

Q: What were the key concerns people had about cloud computing when it first began?

A: You really can’t look at any part of cloud computing without looking at Amazon as being the vanguard. When Amazon first came out with their cloud computing offering, it was perceived as a huge retail company that had excess capacity on its platform mainly because they have to engineer for very peaky volumes such as Cyber Monday. They got an idea to sell this capacity instead of having it just sit there. Early days were quite challenging for cloud: stability was certainly more of an issue five years ago, and security and compliance were probably the biggest barriers for people putting financial data or other sensitive data in the cloud.

Q: How did Amazon Web Services overcome those initial concerns and lead to a more accepted version of cloud computing?aws

A: Amazon made some very significant inroads in a number of ways. First was service quality; they are constantly learning to minimize failures. The excellence they’ve brought to bear has ended up demonstrating itself through increased platform stability and performance.

Also, they got certified for a huge raft of compliance regimes — some related to the federal government, and a broad range of compliance certification around SOC 1, SOC 2, and ISO 27001. These were certifications both for operational excellence and for security best practices, and they went a long way toward opening people’s eyes to AWS as an option. They capped that off by doing marquis projects with companies like FinQloud with NASDAQ, which is all about financial information security. That really served to knock down barriers to cloud adoption.

Finally, Amazon did a lot of work to expose the total cost of ownership so people could compare their on-premise versus cloud costs for the same thing. A study came out in Germany that showed a company could run an SAP cluster for 70 percent less cost than their on-premise dedicated infrastructure.

Q: Did any particular event help clear up security concerns around data in the cloud?

A: In late 2011, the CIA technology chief stated at an AWS conference that the cloud is more secure than traditional approaches. When that happened, IT managers started taking a serious look at how they could create reliable and secure infrastructures in the cloud that would actually surpass the reliability and security of their on-premise infrastructures.

Q: How is the cloud evolving right now?

A: The macro model that’s emerged includes infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). Infrastructure as a service is pretty well accepted. There are still some holdouts — people who think it’s better to do a dedicated setup — but I really feel like that is going to become a dying breed in three to five years. The idea that people can do it by themselves better than it can be done in the cloud is starting to recede.

What I’m seeing emerge now is this concept of managed services that sit on top of the infrastructure as a service layer. This is where DNAnexus comes in: providing a PaaS that runs on top of infrastructure as a service, and offers a really high degree of value in software and expertise. We’re finding that the real attraction is this concept of a managed service which brings to bear compute, software, and analytics along with the expertise that we have in genetic science, bioinformatics, and cloud computing. It’s not just a bunch of computers running in the cloud. It’s a group of people and a platform that they’ve developed based on their genomic analysis expertise that allows customers to take advantage of the cloud in a way that’s frictionless, secure and compliant, and enabling for their collaborative efforts.

Q: What are some upcoming innovations that will affect this space?

A: I think that global reach will increase for cloud providers. Storage — specifically block storage — is also going to see some major advances. I expect to see some advances in direct-attached storage instead of network-attached storage. Another will be long-haul networking; that will evolve in some very interesting ways.

Q: What drew you to DNAnexus?

A: The real key here is that there are two disruptive technologies coming together: next-generation sequencing and cloud computing. They need each other because of the bursty nature of genomic analysis, but you also need a wrapper of HIPAA and CLIA compliance and PHI protection all around it. It’s a really compelling story and it’s what makes DNAnexus really exciting: it’s at the confluence of these two game-changing technologies.

Relationships for Innovation

This week we announced new agreements with two premier healthcare institutions: Geisinger Health Systems (GHS) and the University of California, San Francisco (UCSF). We also announced, with Complete Genomics, our participation in its Genomics Discovery Partners program.

Each of these relationships opens exciting new opportunities. NGS technology generates terabytes of data requiring enormous storage capacities and supercomputing processing power to extract meaningful information. Academic research centers, university hospitals, and commercial organizations risk being overwhelmed by this rapidly growing amount of data. As researchers and clinicians seek to integrate these datasets into their work, industry leaders are increasingly investing to meet this data management analysis challenge.

We are excited to enable these industry leaders to innovate solutions with us on our cloud-based platform. GHS is integrating its genomic data assets with clinical applications, including parent-child trio studies for disease characterization and prevention. UCSF is uploading, managing, and analyzing sequencing data for large-scale genome sequencing research applications. Complete Genomics is offering our data management and visualization services to customers of their human genome sequencing service.

These relationships exemplify ways we help organizations to capitalize on opportunities created by the growing ubiquity of low-cost genomics data. Together we are applying these experiences to create new services and capabilities that support their customers, researchers, and clinicians.

Check back often for updates on these and other collaborations in the works.