June 1, 2017 (Vol. 37, No. 11)
DeeAnn Visk Ph.D. Founder and Principal Writer DeeAnn Visk Consulting
Consciousness-Expanding Biodata Exchanges Can Turn into Bad Trips—Commercial Losses and Security Failures
When technology activists cry, “Information wants to be free,” they relate just half of a thought first expressed by the writer, Silicon Valley futurist, and old sixties icon Stewart Brand. The rest of the thought, which dates back to a 1984 hackers’ convention, reflects a more balanced outlook. Information, Brand clarified, also “wants to be expensive—because it’s so valuable.”
When “free” enters information technology debates, it may refer to unemployed journalists, starving musicians, or frustrated movie executives. Or it may refer to leaks from whistleblowers, unfortunate e-mail caches, or cases of identity theft. Sometimes “free” is about the freedom to avoid paying for anything. Sometimes it’s about the freedom to skirt restrictions on distribution, use, modification, re-distribution, re-use, and so on.
Both senses of “free”—gratis and libre—apply not only to the sort of information at the center of most internet-related controversies, but also to the special kind of information that we call bioinformation. When bioinformation is free, it can be shared by researchers and lead to new life-enhancing products and services. But it can also be exploited by bad actors who might be interested in engineering new pathogens and sowing terror.
Plainly, bioinformation concerns both bioeconomic and biosecurity interests. Both wrestle with the need to balance freedom and control. They appreciate the freedom to innovate, provided it isn’t abused by free riders, or perverted by those who would fashion weapons to cripple freedom. Yet there are also risks in clamping down on bioinformation, not the least of which are lost opportunities for new insights, new cures, and new wealth.
How should bioinformation risks and opportunities be managed? Several answers to that question are provided by the experts who are cited in this article. By and large, they advocate a balanced approach, such as the one reached by Brand two decades after his LSD-fueled stint as a Merry Prankster. They recognize that any misuse of bioinformation could inspire bad actors to repeat Dr. Evil’s take on the swinging sixties and sneer, “Freedom failed.” But they also know that biosecurity advocates could echo Austin Powers, who once said, “Right now, we’ve got freedom and responsibility. It’s a very groovy time.”
Privacy and Competitiveness
Enormous compilations of data can be used to draw correlations about human health from genomic data and other data sources such as health history, leading to the development of new pharmaceuticals and treatments. The safeguarding of these compilations, especially as they pertain to human health and human genomics, concerns Edward You, supervisory special agent in the FBI’s Weapons of Mass Destruction Directorate, Biological Countermeasures Unit.
“With companies outsourcing the generation and storage of this data, they are leaving the safeguards in place within the United States,” says Mr. You. “Health Insurance Portability and Accountability Act (HIPAA) privacy standards are not readily enforceable once the data leaves the borders of the United States. Hence, sensitive data is vulnerable to misappropriation by both companies and nation states.”
The U.S. remains in the forefront in the development of novel therapies, medications, devices, and approaches to human health. Much of this work is accomplished through the judicious use of vast quantities of data. While the laws and handling of this data with regards to privacy are well established and understood within the bounds of the U.S., it is easy to see how this information could be mishandled or stolen abroad, due to the lack of consideration for biosecurity issues.
“Dealing with the convergence of biology and big data and trying to assess ahead of time what the security implications will be, is a current challenge facing the United States,” continues Mr. You. “For example, precision medicine approaches are a great application of big data and biological data, the bringing together of data from the ‘omics’ (genomics, metabolomics, proteomics, etc.) along with other health-related data such as family history and medical records.
“If this data is coopted by bad actors, or nation states, and aggregated, it could provide a distinct advantage to the development of new pharmaceuticals, treatments, and approaches by companies outside of the United States. Is the United States really ready to give up its preeminent role in the world of innovation to other countries by not considering these factors?”
“When companies consider outsourcing the sequencing of human genomes, they need to consider the possibility that the data may also be copied by other parties, as well as returned to the company that ordered the datam,” warns Mr. You. “Companies outside of the United States may offer great financial deals in the short term, but in the long term, this may prove detrimental to the dominant position of the United States in biotechnology and healthcare. All of the United States’ trade agreements protect finished products; the issue here is the aggregation of raw data and how it can be analyzed and turned into innovation.”
Another important aspect of safeguarding the U.S. from biological threats (including radiological, chemical, and biological threats) involves outreach by the FBI. At 56 local FBI offices, there are special agents who serve as weapons of mass destruction coordinators. These coordinators work with first responders, public health officials, and law-enforcement personnel to answer biosecurity questions and resolve potentially dangerous situations. Additionally, these coordinators assist entities with guidance and threat assessment and work with corporate security.
Technology and Security
The idea that technology and technological diffusion relate to national security at strategic and operational levels is central to the work of Margaret Kosal, Ph.D., associate professor, Sam Nunn School of International Affairs, Georgia Institute of Technology. Much of her research focuses on reducing the threat of weapons of mass destruction.
“Sometimes a seemingly innocuous project can take on more malevolent overtones,” Dr. Kosal explains. “For example, a biotech company in southeast Asia decided to engineer a more potent form of the botulinum toxin. From a commercial point of view this makes sense, as less of the product would be needed to have the same effect in cosmetic and medical treatments. Unfortunately, from a biosecurity standpoint, this means the potency of a potential biological weapon increased.”
Dr. Kosal argues that when biotech companies approach projects, they should do more than just keep the bottom line in mind. They should also think about the biosecurity repercussions of their work before deciding to move ahead.
Weak Links in the Supply Chain
Mark Greaves, Ph.D., technical director of analytics for the National Security Directorate at the Pacific Northwest National Laboratory, is preoccupied by a key question: “How can companies assist in promoting biosecurity?” He also entertains a couple of related questions: “How can the National Laboratories help biotech companies avoid becoming unwitting participants in the creation of dangerous pathogens or toxins? What techniques can we develop to help companies encourage responsible science?”
Part of what the National Security Directorate does is develop ways to mitigate security risks in biology supply chains without burdening normal commerce. “We can build from the methods we developed to help secure trade in dual-use nuclear technologies,” Dr. Greaves explains. “The parallel isn’t exact. For example, a major tool for limiting nuclear proliferation is to tightly control fissile material, but similarly controlling biological materials is not possible. So, we focus our attention on the biotech supply chain—both the data flows and the intermediate physical products.
“We are developing lightweight techniques that can help companies in the biotech supply chain flag worrisome orders, such as an unusual set of items or problematic gene sequence, or orders from a non-bona fide person or facility. Helping companies develop a corporate culture where employees are actively encouraged to raise concerns with an order is another important component of promoting responsible biotechnology.”
Such concerns have been taken up by the International Gene Synthesis Consortium (IGSC), which was formed to prevent the misuse of gene-synthesis technology. The IGSC currently includes 80% of commercial gene synthesis capacity worldwide.
In partnership with governments and other entities, the IGSC promotes the beneficial application of gene-synthesis technology and strives to safeguard biosecurity. One way the organization does this involves creating protocols for screening individual synthetic-gene orders as well as the customers who place them.
“One caveat to providing cybersecurity it the problem of evaluating the efficacy of the measures that are in place,” comments Robert Sloan, Ph.D., professor of computer science and head of the computer science department at the University of Illinois at Chicago. “In order to evaluate computer security, one has to practically be an expert oneself to determine if cybersecurity measures are efficacious and cost-effective.”
There are varied solutions out there for computer security, explains Dr. Sloan, who adds that these solutions also have varied prices, some of which exceed the cost of fixing the sort of data breaches that the solutions are meant to prevent.
“One way to conceptualize computer security is to think of home security,” suggests Dr. Sloan. “If a house contains many valuables (computer data), then security needs to be higher than a home without much information of value. Also, the home of well-known persons or entities needs more security than the average American.
Sharing Data, Sharing Genes
“Most genetic signals related to health can be found only by integrating information across many genomes,” observes Nathan Price, Ph.D., associate director and professor at the Institute for Systems Biology. “Thus, without data sharing, no individual’s genome data would be of much value, including to the person him or herself. If all the genomes in the world were sequenced and none of them were shared, then there would be little value in all that sequencing because we wouldn’t much understand what any of it meant.
“While data security is clearly very important, this need is counterbalanced by the need for data to be shared to be interpretable. This is the basis for the challenges that must be met for safe data-sharing with qualified people and trusted organizations so that the benefits to the individual and to society can be maximized.
“Most people who want to analyze human genomes are highly motivated by wanting to help people and advance the cause of human health and wellbeing. However, issues around potential misuses of these data are clearly important. The threat of an ill-intentioned group using genomic data, for example, to design a virus that afflicts one ethnic group over another would be a terrible outcome.
“However, such events are unlikely to occur. Often people who use genetics to determine their ethnic heritage are surprised to learn how mixed they are through the diversity of their ancestors. It is unlikely that a bioterrorist could use genomic data to design a bioweapon that affects just one ethnic group or, conversely, does not affect a particular ethnic group.”
“While this is clearly improving dramatically over time, humans as a group can tend to see themselves as different or separate from people of other ethnicities,” concludes Dr. Price. “In truth, we are much more mixed than some might think, based on genomic analysis. The beautiful truth emphasized by studying genomics is the oneness of the human family.”
Reaching consensus on the interpretation of a raw dataset is another challenge inherent with Big Data. “If you present the same raw dataset to five different bioinformaticians, you will get five different answers,” contends Folker Meyer, Ph.D., computational biologist and deputy division director of biology at Argonne National Laboratory. “Additionally, the primary data in DNA/RNA sequencing is frequently lost. Only the processed data (which is frequently massively biased) is saved, making retrospective analysis with improved or different methods nearly impossible.
“Each person comes at data analysis with a slightly different approach. Also, the person doing the analysis is generally not the principle investigator, but a graduate student or postdoctoral fellow, which is to say, someone who is more comfortable running the software analysis of the raw data. Hence, the person in charge relies on the judgements of more junior members of the group.
“Another issue is getting actionable data from the analysis for human health. To truly evaluate what the impact is on patient treatment, an individual would need a deep understanding of the analysis and medical practice; those individuals are rare.”
“The last problem is that the cost of data analysis is not factored in,” adds Dr. Meyer. “Data acquisition is much cheaper than it was at the time of Darwin’s voyage. For example, generating DNA sequencing data is cheap. Still, costs tend to accumulate during data analysis, which requires massive computing power and a diverse set of computing, math, statistical, and analytical skills. Few biologist have the specialized training for this work. The cost of this is often overlooked in budgets.”
Data Science Investments
Big Data approaches are being used to analyze complex biological systems. For example, Ivo Dinov, Ph.D., associate professor in the department of health behavior and biological sciences at the University of Michigan, is working to improve the management of Big Data.
“We have developed protocols for data wrangling, harmonization, aggregation, processing, model-based and model-free inference, and visualization,” informs Dr. Dinov. “Once all the details are worked out, the implications for pharmaceutical companies may be enormous—from expediting drug discovery to improving prediction of treatment outcomes.
“Financial rewards are likely to be huge in the long run. In the immediate future, we need a massive investment in basic, transdisciplinary, and data-driven sciences.”
Dr. Dinov insists that long-term investments will yield substantial dividends, but he recognizes that tightfisted players will fail to seize the opportunities ahead. “The shortsighted may meet near-term fiscal expectations but miss the coming economic tidal wave,” he cautions.
“Benefits to patients will be substantial,” Dr. Dinov continues. “No medical treatments are guarantee complete success. Nonetheless, effective, predictive Big Data analytics will reduce risks, tighten confidence limits, improve forecasting, and expedite medical decision-making.”
With all these benefits promised by combining Big Data, biosecurity, and biotechnology, scientists, business people, and government officials need to be aware of the risks and opportunities of allowing domestically generated bioinformation to be amassed abroad. Protocols for the sharing of bioinformation must be carefully balanced to facilitate discovery while safeguarding privacy, protecting commercial interests, and maintaining the common defense.
Precision Medicine Exchange Consortium
Advances in genomic profiling are producing a wealth of information about the biological mechanisms of cancer. Translating this information into better treatments depends on a multistakeholder collaboration that promotes broad access to large, diverse datasets that can help us make connections between genomics and clinical outcomes.
Two years ago, Foundation Medicine (FMI) launched the Precision Medicine Exchange Consortium (PMEC), which is designed to facilitate information-sharing by breaking down barriers between institutions. PMEC unites academic centers, regional hospital systems, and community oncology networks under a single, secure data umbrella to establish a combined knowledgebase that can simultaneously inform patient treatment today while optimizing future drug discovery and development.
“Advancing cancer care is greater than any single institution,” said Gaurav Singal, M.D., vice president, data strategy and product development at FMI. “All of us in the global oncology community recognize the importance of collaboration to remove silos and enable discoveries that ensure patients benefit from the full potential of genomic data.”
Last year, FMI contributed 18,000 de-identified adult genomic profiles to the National Cancer Institute’s Genomic Data Commons Portal, more than doubling its size at the time. FMI also shared profiles of more than 1,200 pediatric tumors to stimulate research in childhood cancer. Additionally, a groundbreaking partnership between FMI and Flatiron Health facilitated a real-world dataset that links genomic and clinical outcomes data for over 20,000 patients to identify actionable biomarkers in lung and other cancers.
Newer precision technologies, like liquid biopsy, can especially benefit and advance by sharing insights. FMI is partnering with government, academia, pharmaceutical, and diagnostic experts via the Blood Profiling Atlas, which aggregates clinical data for researchers to speed development of blood profiling technologies.
As walls continue to fall, multi-institutional data-sharing initiatives hold the promise to make meaningful strides in improving diagnostic technologies, drug development, and cancer care.