Looking ahead: the future bioinformatics of genetic testing and precision medicine
One of the most effective ways to reach an audacious goal is to envision, in detail, exactly what you aspire to achieve. In that spirit, I’ve been thinking about what the genetic testing landscape could look like ten years from now so that we as a community can begin reverse-engineering solutions to overcome the obstacles facing us today.
Let’s imagine that it’s May 2027. Precision medicine is standard practice, and genetic testing is routine and essentially ubiquitous. Clinical labs run genetic tests of varying sizes and complexities, and genome-wide testing is far more affordable and broadly used for a range of indications. Genome sequencing is no longer a one-time test for each patient; physicians have patients’ genomes sequenced repeatedly over time to monitor for changes and answer new clinical questions. Clinical labs and patients themselves contribute a great deal to secure, cloud-based personal data repositories, both as active participants in their and their loved ones’ care, and as contributors to global R&D efforts that extract new medical insights from large-scale, privacy-compliant, research-accessible genome repositories. Finally, since genetic and genomic testing is routine, molecular data specialists are more directly involved in patient care.
Now, let’s take this future apart, piece by piece, and pinpoint the challenges we must overcome to make it a reality.
Molecular data must infuse healthcare IT
We will not fully embrace the era of genomic medicine without significant improvements to the computational infrastructure and policies that govern healthcare data. Today’s labs and hospitals lack shared data standards and technical integration capabilities robust enough to share, manage, manipulate, and make effective use of rich molecular datasets. Medical record systems are increasingly electronic in nature, and show great promise for managing patient data from a variety of sources, but today’s systems still lack broad support for storage, representation, and useful querying of molecular test data. We need electronic medical records that can manage genomic datasets, integrate them with other large- and small-scale clinical data, and enable integrative querying of the comprehensive patient record. Simple representation of clinical test reports as PDF documents is not enough; our clinical tools must support the richness and complexity of genomic and other molecular test results as well as their associated clinical metadata.
Data privacy will be another key component of our future IT landscape. We must address concerns about the privacy of patient data to take full advantage of large-scale analysis of aggregated patient datasets, and to allow the healthcare industry to fully realize the economic benefits of cloud computing. Techniques like differential privacy and other privacy-preserving computing and data transfer methods require broader adoption in order to address legal and regulatory concerns around sharing and transfer of molecular data. Addressing privacy concerns effectively will be critical to allow patients to share their data for research studies easily and safely.
Integration of genetic insights with new sources of health data
Insights from genetic testing will be integrated with increasingly available patient data from consumer devices and novel assay platforms. With the proliferation of health-related mobile apps and wearable devices, in a decade’s time we should expect that clinical interpretation of patient test results will occur within the context of multiple patient-generated streams of health data. Whether that’s from sleep sensors measuring a range of physiological parameters or smart toilets with built-in urine antigen screening assays, there will be plenty of data feeds that, if properly quality-controlled and integrated, can help make diagnoses more accurate and personalized. In addition, as molecular assay costs continue to drop, genome-based testing will expand to include microbiome, metabolome, proteome, and transcriptome data. While such multi-omics studies will likely be common in research, in ten years they may still be too expensive to be done routinely for the broad patient population. Instead, these hybrid assays will be validated for specific indications where single assay modalities show limited clinical utility, and from there will gain traction in mainstream healthcare.
IA will lead AI in clinical applications
That is, intelligence augmentation (IA) will come before artificial intelligence (AI). With this abundance of data, it will be more important than ever for clinical labs and physicians to have high-quality, sophisticated tools for accelerated, evidence-based interpretation of test results and other information used to drive clinical decisions. While such tools will undoubtedly leverage algorithm innovation from computer science—and artificial intelligence and machine learning in particular—we are unlikely to see AI drive a radical shift in medical practice for some time. In ten years, the bulk of healthcare and precision medicine will still be operating with trained physicians, oncologists, and primary care givers—but with improved effectiveness thanks to these algorithmic techniques. Well-designed software systems with powerful embedded computational engines will enable doctors to acquire, filter, integrate, and make sense of all information relevant to a patient and his or her clinical context. These decision-support systems will give clinical geneticists, pathologists, and physicians the right information at the right time to drive the right decision.
Technologies that aim to improve human decision making must be radically transparent, evidence-based, and customizable in their function. Systems that aim to improve decision making must provide full access to every data point, every reasoning step, and every calculation that supports a certain clinical conclusion—and provide clear mechanisms that allow expert human insight to guide the system performance. This design philosophy will make it possible to generate not only a simple one-page report for people who need a concise, high-quality clinical insight, but also a 100-page report with supporting evidence which illustrates the reasoning behind each summary result. Clinicians will review these details as needed and be confident in system correctness and performance. Well-designed systems will further enable customization by allowing users to adjust analytic workflows on a case-by-case basis, change filtering assumptions and decision logic, adjust computer-based diagnostic and treatment guidelines, and introduce new sources of evidence.
Molecular expertise across the care continuum
Greater involvement of clinical lab personnel across the patient care continuum will also be key. As molecular testing becomes routine, clinical lab professionals will play a greater role in the direct care and treatment of patients. This transition is already happening in oncology, where tumor boards have created partnerships among oncologists, pathologists, lab specialists, and other care providers to foster a holistic perspective on patient care. To apply this team-based approach in other indications that can benefit from deep molecular insights, we need better organizational integration between clinical lab specialists and medical practitioners. We also need IT systems with robust support for molecular testing data.
The fundamental thing to note about this vision of genetic testing just ten years from now is that there is no “insert miracle here” step. While the goals are ambitious, every aspect involved is realistic and achievable. Technological improvements are feasible; many are already underway. Policy changes are certainly achievable if we come together as a community with clinical and patient advocacy groups. I believe that if we join forces and push for change, we can usher in a new era of clinical testing and significantly improved healthcare for patients around the world.
Ramon M. Felciano, PhD, serves as QIAGEN Bioinformatics’ Chief Technology Officer and was a founder of Ingenuity Systems, a QIAGEN Company. He leads strategy for the QIAGEN Clinical Insight (QCI) solutions, a clinical decision support platform designed to help labs streamline and accelerate the interpretation of genetic test results.