Founder and CEO, LinusBio
In theory, precision medicine is a relatively straightforward concept. Clinicians use advanced techniques like genomic sequencing to identify disease-related biomarkers and develop a personalized treatment plan tailored to the individual’s unique genetic and molecular profile. These biological elements can help them rapidly diagnose a patient’s disease and direct them to the most effective therapies. This approach differs from the traditional “one-size-fits-all” method, where each disease is treated with a preset recipe of interventions without regard to what makes each patient unique.
In the bigger picture, biomarker-based tests can also help identify the right patients to include in clinical trials, potentially accelerating drug discovery and development. Biomarkers have been particularly important in rare diseases and oncology. Next-generation sequencing has helped identify individual patients’ unique diseases, directed them to treatment, and isolated potential therapeutic targets for further study.
Unfortunately, for other conditions, precision medicine theory and disease reality have yet to match up. In many cases, effective biomarker tests do not exist. This is particularly true for neurological conditions, such as autism spectrum disorder, where observational tools like questionnaires are still being used in the evaluation process, relying on subjective assessments rather than objective, measurable factors. Without the necessary diagnostics, clinicians and researchers can easily access essential data and the presence of disorders, delaying treatment. This creates a slow, cumbersome, and inefficient therapeutic progress.
Against this backdrop lies another more serious question: Will genomic (or transcriptomic or proteomic) testing provide all the necessary information we need to truly implement precision medicine? The short answer is no. The long answer is that billions of dollars of investment in sequencing the human genome and then developing tools for mass screening our populations have fallen far short of the hope and promise that all diseases would be better understood once we had the genomic blueprint of life.
The stark reality is that genomics alone only explains a small part of overall health and the most common disorders. This becomes clear when we confront a simple fact —if we sequence a newborn baby’s genome, how well can we predict their life long health journey? In most cases, we can’t predict much at all.
Missing the environmental picture
At the population level, few diseases are predominantly genetic. A better view is that our life’s health trajectory is determined by an interplay between our genes and environment. Over the past 20 years, epigenomic and other studies have confirmed how environmental influences can even shape gene expression–this only further enforces that we cannot partition our genes from our environment the way we have been doing in most genomics scientific research to date.
Still, the current biomarker landscape remains deeply rooted in genomic sequencing. This is like going to a gallery and only looking at the frames, ignoring the masterpieces that lie within.
The time has come to truly embrace environmental data, not to supplant genomics but rather to augment it. Understanding environmental influences has tremendous potential to clarify—and synergize with—existing genomic information.
This is particularly true for neurological conditions. The brain remains a (mostly) black box, and so far, sequencing has provided little illumination. We see this in twin studies of many brain disorders, including those that arise in childhood, such as autism spectrum disorder (ASD) or at other life stages, e.g., schizophrenia in adults to ALS in older adults. If a person’s genomic inheritance were truly their destiny, there would be no variations between identical twins.
Yet, there are many examples of twins’ neurological health diverging. Environmental factors influence their development even prenatally. This shouldn’t be surprising–many identical twins are not born with the same birth weight. Why? Their genes are identical, and so is their household, same mother, same father, same diet, etc. The reason identical newborns can differ in their body weight is that their internal environment–think of it like their metabolism–is different, even prenatally. Genetics, therefore, has almost no ability to explain something as simple as why two identical newborns don’t weigh the same.
Integrating environmental data
While scientists have known for many years that the environment directly impacts our health and influences gene expression, it’s only relatively recently that we’ve had the technology to develop biomarkers that can deal with the complexity of environmental exposures and our body’s response to those exposures.
First, we are now awash in environmental data. Air pollution sensors, satellites, and other instruments precisely measure particulate matter and other potentially toxic substances in the air we breathe. Wearable devices have become ubiquitous and can provide useful, longitudinal information about individuals’ vital signs and health habits, as well as how much time a person spends near sources of pollution.
More advanced technologies have provided never-before-seen insights into the temporal dynamics of our physiology by mapping molecules at hundreds of time points. These technologies are lasers that expel molecules along the growth rings in tissues like the scalp, hair, and teeth (similar to the annual rings of a tree).
The resultant datasets are massive in the range of different types of molecules that can be measured and in the number of time points in a person’s life that are captured (hundreds to thousands). These technologies are ushering in a new era of temporal biomarkers that allow clinicians and scientists to move away from a static to a dynamic view of human physiology.
As powerful as these developments are, they need a technological overlay to make the data actionable—big data and artificial intelligence (AI). With their incredible predictive potential and ability to harness huge datasets, machine learning, and other AI approaches can combine multiple data streams to produce precise conclusions.
These capabilities could provide a new window into disease. By combining genomic and environmental biomarkers, researchers and clinicians can develop sophisticated disease signatures to rapidly diagnose and treat ASD, Alzheimer’s disease, ALS, and possibly many other conditions where a static, gene-centric view of human physiology hasn’t moved the needle much.
Manish Arora, MD, is founder and CEO of LinusBio.
