Matt Might is a husband, father, Director of the Hugh Kaul Precision Medicine Institute at University of Alabama at Birmingham, President of the NGLY1 Foundation, Professor at Harvard Medical School, and former USDSer.
After seeing a video dubbed ‘Hunting Down His Son’s Killer’ that was circulating through the USDS family, we wanted to share this inspiring story. After four years of no diagnosis, the Might family found out their child was the first-ever with NGLY1 deficiency. There were no documented patients and no treatments. With no time to waste, Matt ultimately went from being a professor of computer science to a professor of medicine.
When no one could figure out their son’s disease, Matt and his wife, Cristina, made it their mission to find the answer to save their son’s life.
It’s truly inspiring to see what Matt and his wife have accomplished in such a short amount of time in the worlds of medicine and science. They have helped to identify 70 patients with NGYL1 deficiency — none of whom would have known they had this deficiency otherwise.
Matt is now building a community of patients, researchers and doctors to figure out how to help his son and those with NGLY1 deficiency. His primary research area is precision medicine — the use of data (particularly genomic data) to personalize treatments and optimize healthcare outcomes.
While at USDS, Matt was a strategist detailed to the Office of Science and Technology Policy, working on precision medicine, the Million Veterans Program, cybersecurity, and biosecurity.
In the Superhuman video, you said “We’re going to do everything possible, and when we run out of the possible, we’ll do the impossible” — how did you know where to start, when doing the impossible?
What I mean when I say “We’re going to do the impossible” is “We’re going to do science.”
Science is the human endeavor that transforms the unknown into the known, and so it is also how we transform the impossible into the possible. At every step of my son’s journey, we needed a different kind of science, so the question was always, what is the kind of science we need to take the very next step?
First, we needed to know what he was suffering from, so my wife and I turned to a new kind of genome sequencing (exome sequencing) to get the answer. Once we had the answer, we realized we had to do glycobiology to understand what was going on inside his cells, and to find ways that we might correct that with therapies.
Once we understood what was going on, the science to find a treatment became pharmaceutical chemistry, medicinal chemistry, and bioinformatics. Ultimately, doing the impossible is a very step by step process: the catch is that you can’t see the entire path at the start; the next step only becomes clear once the last one has been taken.
Within two weeks of writing this blog post, you were able to identify the next patient with NGLY1 deficiency. What were you feeling when you first found out there was another human-being with the same deficiency as your son?
It was incredible to find out we were no longer alone. It had been wrenching to live without a diagnosis for four years. It was shocking to find out Bertrand was the first. It was inspiring to realize there were others — that there was hope we could build a community and go after treatments for this disease.
You went from being a professor of computer science to a professor of medicine. Computer science and natural science are not the same thing. How do we connect the dots here?
It’s certainly true that computer science and medicine are very different fields, but I think there are two answers here.
The first is a general answer about the nature of what you’re trained to do when you get a Ph.D. Some people think that you’re gaining a deep specialization in some field as you extend human knowledge.
I made an illustration a few years ago to explain what a Ph.D. really is.
So, it’s true that one does gain a specialty, but the primary skill one acquires in earning a Ph.D. is how to extend human knowledge, and this skill seems to transfer across fields.
The more specific answer is that the way we think in computer science carries a lot of advantages in medicine.
Biology is messy, but there are abstractions a computer scientist would find familiar. DNA is a string. Genes are programs written in a (universal) instruction set for creating proteins. Enzymes act like functions. Metabolic pathways behave like categories in category theory. Genes regulate each other as Boolean networks. Cellular signaling looks like networking theory. In fact, to some extent, cells are just quirky little Turing machines.
In effect, biology and medicine need computer science. But, what they need even more are computer scientists.
So, for my own transformation, I was lucky: I had unbounded motivation to help my son, coupled with the good fortune that computer science happened to be an excellent tool for understanding biology.
Perhaps most importantly, I had a supportive and understanding wife and family that worked side by side with me and let me pursue this as far as it could go.
You’re now an advocate for precision medicine, delivering the right drug to the right patient, at the right time. How does this incorporate your two worlds of computer science and biology?
Precision medicine uses data to optimize the health of a patient or of an entire population. So, precision medicine casts healthcare in terms of a constrained optimization problem. Precision medicine is data-driven computational medicine.
At my institute, we have created an “Algorithm for Precision Medicine” that helps us identify the next step to take for any given patient. Often times, that step involves research, and so it allows us to incorporate research into the personalized care of patients in a very consistent way.
To some degree, when we work with patients, it feels like we are “debugging” them.
At times, we very directly apply computer science: my faculty member Will Byrd and I have created a tool called mediKanren that has digested all of the abstracts in PubMed with natural languages processing into simple relations (e.g., X inhibits Y, A activates B). By layering artificial intelligence and logical inference on top of this, mediKanren can infer novel treatments for patients and diseases by connecting disparate parts of the medical literature.
In addition, there are a growing set of situations where molecular simulations can help identify precisely targeted therapies for a patient or a disease.
When we look back at the end of the 21st century, we’re going to find the best drug we discovered was data — and that the limiting reagent in saving human life is computation.
So, how did you end up at USDS? Why did you decide to sign up for a tour of duty?
Prior to the launch of Precision Medicine Initiative in 2015, I got a call to come to the White House. I ended up meeting with President Obama, and he asked if I would be willing to help with the soon-to-be-announced initiative. Of course, I said, “Yes!” I started going back and forth to the White House to help out, and after about six months, the precision medicine team in OSTP encouraged me to apply to USDS and join them on the inside — so I did!
What would you tell someone who is considering joining USDS?
Do it! You’re going to have awesome colleagues. Everyone I worked with at USDS brought talent, a can-do spirit and a belief that together we could make the country better. Getting to work alongside fellow USDSers on any task would make it a great job, but when the mission is to make an impact at a national scale, it becomes the best job you will ever have.
What do you hope is in the future for the worlds of medicine and science (computer science AND natural science) :)?
I hope that, increasingly, they’re the same world. Halford Mackinder noted that “Knowledge is one. Its division into subjects is a concession to human weakness.” I believe that computation is the intellectual glue that can help bring subjects like medicine and computer science back together.