Precision medicine is the concept of using an individual’s genomic, environmental and lifestyle information for tailored treatments, and it has the potential to completely disrupt our approach to maintaining long and healthy lives.
To shed further light on where we are today, and how the future is going to look like I sat down to talk to Joel Dudley. As Joel appears on the screen in our online call he looks nothing like your stereotypical researcher, with a ball cap, grey hoodie and the size and build of an american football player. But make no mistake, Joel is the former leader of Mount Sinai’s center for precision medicine, he was leading the science division at the precision medicine company Tempus as they went through a massive scale-up. He is a serial entrepreneur, and can put a long list of high impact research published in some of the worlds most prestigious journals on his resume.
With an almost manic obsession over every detail from genetics and molecules, to systems biology into the world of healthcare, medical records and insurance data, Joel has worked to figure out how we can use that data to understand biology much more holistically and improve human health. “I am a bit of a health nut myself” Joel says, as he talks about his lifelong fascination of humans, and working on understanding the limits of human performance and human cognition.
For Joel precision medicine is an aspiration, because today we have very imprecise medicine. You see it every time you step into a hospital and your health is divided into symptoms and anatomical parts on the placards.
Today our physiological and anatomical understanding of disease, means that if something goes wrong in a given part of the body we call in a specialist in that particular part to fix it. But looking at cancer we start to realize that the molecular mechanisms are often more telling than the location of the disease. For instance, a lung cancer and a colon cancer sharing a given mutation pattern across pathways will respond to a drug targeting those pathways, rather than a generic lung cancer or colon cancer drug. Joel refers to one of the earliest studies he did showing that indeed the gene expression signatures of disease is stronger than that of a tissue specific gene signature [1], meaning there is a shared pathophysiology in disease that breaks our current “body parts” view of disease.
Unfortunately, the molecular mechanism of disease can also differ among individuals with the same disease, leading to a more muddy concept of neither a perfect “body parts” or a “disease specific” understanding. An example of this is atopic dermatitis [2] which have multiple different endotypes. Joel uses the analogy of a car with a smoking engine, we observe that the car is broken and that the engine is smoking, but the underlying reason could be leaking coolant, or 50 other reasons. In the case of a car we have the complete schematics and can methodically test each unit to see if it is working. We are just starting to obtain a similar schematic of human disease with multi-scale biological networks [3].
A major challenge in starting to build such schematics of disease is our lack of understanding of a healthy state. Today we observe disease when it is present, so we are likely looking at downstream effect of the actual causal drivers. In Alzheimers disease we have a focus on beta amyloid, but the actual causal events might happen somewhere in your 40’s or 50’s. “Think of your healthy state as a trajectory” Joel says, “a minute change in that trajectory for whatever reason will grow as the years pass, and you might end up in a very different (disease) state”. Mapping these state spaces and how we transition between them over time is really at the core of laying a foundation for any precision understanding of both health and disease. And laying such a foundation will require massive amounts of measurements across large populations for extended periods of time.
This leads us to discuss continuous glucose monitoring which is now available for consumers and could provide such foundational data. However, only on a single parameter which does not lend itself to a multi-scale understanding. A core challenge in this type of data collection is that people lead busy lives, and any collection would need to be almost passive and confer a clear value to the consumer. The catch 22 is that we have to collect the data first to tell you what the benefit and value is.
Data collection and value proposition is not the only challenge, as we start to grow the number of parameters we can measure the chance that multiple people have similar parameters decrease dramatically. Meaning we won’t be able to do inferences as nobody is sufficiently similar to you, so you will need to resort to N-of-1 trials where you are your own control. “The real empowerment in this space”, says Joel, “is enabling people to do interventions themselves and be able to draw conclusions. Do I actually sleep better when taking cold showers, than when I don’t etc.“ Another reason for N-of-1 trials is the dynamic nature of your entire biology. No matter what aspect of our body’s biology we are dealing with, it is constantly changing especially as we start to do interventions. While taking a particular supplement to enhance gut microbiome might have beneficial effects initially, continuing to do so might produce too much of the “good” microbes that they start to out compete other beneficial microbes. This balancing act within an optimized healthy state requires continues refinement and shifts in behaviour and intervention.
Such an empowered future is still a ways off. We have just started to see a movement from pure consumer diagnostic companies like 23andme, to companies that deliver products or actionable insights based on the diagnostic measurements. For Joel the most promising aspect of precision medicine is not within medicine and disease, but within health and wellness. “Going back to our health trajectory idea, the earlier you can act and change this trajectory the further you can push out onset of disease. This effectively moves us into a preventative precision medicine paradigm”, says Joel.
References
[1] Dudley, Joel T., Robert Tibshirani, Tarangini Deshpande, and Atul J. Butte. “Disease signatures are robust across tissues and experiments.” Molecular systems biology 5, no. 1 (2009): 307.
[2] Salimian, Jafar, Zahra Salehi, Ali Ahmadi, Alireza Emamvirdizadeh, Seyyed Masoud Davoudi, Mehrdad Karimi, Mohsen Korani, and Sadegh Azimzadeh Jamalkandi. “Atopic dermatitis: molecular, cellular, and clinical aspects.” Molecular Biology Reports 49, no. 4 (2022): 3333–3348.
[3] Kidd, Brian A., Lauren A. Peters, Eric E. Schadt, and Joel T. Dudley. “Unifying immunology with informatics and multiscale biology.” Nature immunology 15, no. 2 (2014): 118–127.