Decoding Phenotype: The Holy Grail for Today's Medical Scientists

We've figured out DNA, but we need to start watching and measuring with greater precision how it interacts with the environment.

Starting with the elucidation of its structure by Watson and Crick (PDF) in 1953, medicine has been captivated by the power of DNA, by the ability to understand our health and treat disease based on subtle differences in the four-letter code that comprises our genetic information -- our genome.

The technology of DNA sequencing has progressed to the point that the readout of an entire human sequence -- a feat that just a decade or so ago was considered medical science's holy grail, worthy of a massive, Manhattan-Project style mission, and ultimately celebrated by a Presidential press conference -- has evolved into an almost routine activity for a number of technology companies, at a cost now closing in on $1,000 per genome, and dropping fast.

Perhaps not surprisingly, the spate of data emerging from sequencing efforts has not only revolutionized our understanding of disease, but has also highlighted fundamental limitations in our scientific understanding. As we try to wring meaning from the petabytes (that's over a billion million bytes) of data, it's become increasingly clear that sophisticated -- and actionable -- understanding of biology and disease requires not only a parts list, but also a nuanced readout of how the parts operate together in the context of a cell, an organ, a person. In a word: phenotype.

Phenotype is how something looks, acts, or behaves; in contrast to DNA sequence, which is fundamentally discrete and universal, phenotype tends to be much "messier," more challenging to reliably assess. Two investigators across the world can easily agree on the exact DNA sequence within a specific cell, say, but might come to very different conclusions about how the cell behaves in culture.

The greatest challenge may also be the most important: measuring complex human phenotypes, such as how a patient is experiencing a particular disease, or responding to a given treatment. Too often, and quite understandably, the approaches used by physicians and medical researchers have been relatively simple, episodic assessments -- measuring a patient's blood sodium, or blood pressure, for example, at the time of an annual physical. Such evaluations can provide important and useful information, but rarely capture the complexity of a patient's health and experience over time.

We envision improved measurement of phenotype as the underlying basis for the next generation of medical progress. Improved measurements of patients can guide -- immediately -- the treatment approach used by physicians, who often have very little visibility into what happens after a patient leaves the office. Better measurement can also guide medical product development, focusing attention on a patient's true unmet needs.

The FDA, to its credit, has recognized the need for improved measurement, and has been an early champion of the need for better "assessment science." Speaking at a conference on the subject last year, the director the FDA's Center for Drug Evaluation and Research, Dr. Janet Woodcock, noted (PDF, p.13) that "the identification, development, and qualification of new clinical trial outcome assessments has not been aggressively pursued by the scientific community," adding that "the consequent lack of assessment tools has been impeding, I think, the development of new drugs because we really, in many cases, don't know how to measure the impacts, both for good and ill, of the drugs we test in people."

The ability to measure with greater precision the real-world impact of a patient's illness would also enable improved assessment of the impact of both treatment approaches and of the providers themselves, giving us an opportunity to better assess the value of each, and to enable the iterative improvement of patient care and health delivery.

The improved measurement of complex patient phenotypes will also provide enormous benefit to basic researchers, enabling them to link this new information with existing, rich genetic data to form coherent datasets that can help identify key underpinnings of disease, and enable researchers to develop more targeted, and in many cases more personalized, interventions. Integrated datasets will also fuel increasingly sophisticated computer-driven "in silico" modeling approaches, capturing the benefits of empiric, "big data" technologies and approaches already used to great effect in other industries and disciplines.

The improved assessment of phenotype will be profoundly enabled by the explosion of mobile technologies, and the rapidly growing field of "mHealth." At the same time, we also recognize that technology and data collection alone are unlikely to provide the robust, integrated insight that science demands and patients deserve. While it's essential to understand the performance characteristics of the technology, and be sure it is reliable and robust, it's even more important that we're measuring the right things, focusing on the aspects of disease most troubling to patients, rather than simply the features we are able to quantify most easily.

This brings us to our final and perhaps most important point about phenotypic measurement: While it's essential for medical researchers and mHealth technology developers to view patients as partners in discovery, we must also recognize that most patients want to spend their time living their lives, not thinking about their health. Perhaps the most substantial design challenge for assessment science, and for mHealth in particular, will be delivering improved health without creating new burdens -- liberating patients from their illnesses, rather than dominating patients' lives with disease monitoring and management.

An approach to phenotype assessment that's sensitive enough to capture the real-world experience of patients, powerful enough to enable actionable change leading to improved health, and wise enough to do this without being intrusive -- this is the holy grail for today's medical scientists.

Image: dream designs/Shutterstock.