Dr. Geraci’s efforts mainly concern precision medicine, using mathematical and computational methods to construct models of disease that go beyond classical top-down clinical definitions. After completing postdocs in oncology, biological psychiatry, and artificial intelligence he created NetraMark Corp where he has been developing novel technologies that aid in the understanding of our molecular and brain circuitry in addition to novel machine learning algorithms specialized to help understand complex patient populations. He is also a professor of Molecular Medicine at Queen’s University in Ontario.
Dr. Geraci has a strong interest in advancing the mathematical methods being employed in the study of our molecular circuitry (protein, microRNA, mRNA), the analysis of brain MRIs, and machine learning that can use variables that are beginning to emerge due to our interaction with technologies like fitness watches and smart buildings. A major interest of his is an ongoing project involving translating the vast amount of genetic and proteomic patient data, coupled with our current knowledge of our molecular circuitry, into a scoring scheme that can reveal potential new drug targets. His primary areas of research are in oncology, biological psychiatry, mathematics, physics, and machine learning. Additionally he is a quantum computation researcher and is looking to augment machine learning technologies with quantum based algorithms as quantum computation matures.
An excerpt from the NetraMark webpage reads: “We provide pharmaceutical companies with exciting and new analytical tools that reveal insights about safety, placebo, and response that lead to massive savings in time and money. Our system allows human clinical trial experts to inform the analysis through an interface that allows them to interact with high dimensional data like nothing else. We unpack the black box!”
The following has been paraphrased from an interview with Dr. Joseph Geraci on February 2, 2018.
(Click here for the full audio version)
What aspects of this new age of computing (artificial intelligence (AI), quantum computing) is going to have the biggest impact on our health?
Let’s start simply, what’s happened in AI lately is that some mathematical innovations merged with some powerful computational innovations that ended the so-called ‘AI winter’ and made machine learning feasible. We are now able to train our machines to recognize images and to generalize. What I mean by that is that in machine learning you want a machine to be able to recognize something, for example, will this patient respond to this chemotherapy? Or is this an image of a cat? We can now get the computer itself to do the recognition. This was all the product of advances made in mathematics coupled with advances in computing hardware.
This allowed a lot of people to jump into machine learning, and one of the areas where people saw use in this was in medicine. I was one of those people, 10 years ago I started modeling Huntington’s and Parkinson’s and some types of Cancers using these machine learning methods. It opened the doors for people in other fields, like mathematics or physics, to start focusing on these problems, which sparked a lot of innovation. There are drugs out there now that are more precise because of this influx of talent and advances in machine learning. But there is still a lot of work to be done to further improve precision.
The next phase of advancements will come from our ability to understand the shape of proteins that go into making small molecule drugs. This is something people have been doing for decades and which machines are getting better and better at, but quantum computers hold a very unique promise that will greatly improve our ability to make these drugs. The promise is this, classically, when you want to understand the shape of a protein or how it interacts with something, which is essentially what drugs do, you make a model of the protein you are using and try to map how it will disrupt certain pathways in our biochemical circuitry. But when you want to really get down to the level of small molecules and figure out how to engineer a protein to configure itself in a precise way, that becomes incredibly difficult. The reason is because by their nature, proteins are quantum. They are hundreds of amino acids that clump together, yet they emerge as this very specific shape that has a very specific job as a result of a variety of evolutionary forces. But ultimately they are quantum in their nature, they ‘live’ in a world governed by the laws of quantum physics. So now, thanks to recent advances in quantum computing, we are finally able to start examining these quantum interactions and create much more accurate simulations of molecules in a way that a classical computer could never do. This is where advances in medicine are going to explode.
How far away are from true precision medicine, where we can get drugs tailored for each of our specific biological makeup?
This is something I devote a lot of my attention to, I’ll start by telling you where we are and where we are going. Right now, doctors have been trained to look at things statistically. But now we have this combination of pure biologists, computer scientists and mathematicians working together to solve this. This has allowed us to say, this woman has this mutation which resulted in this type of breast cancer, and instead of giving her chemotherapy A we should give her chemotherapy B. That’s where we are for several diseases.
These advances came from being able to identify single mutations that create individual proteins, however there are over 20,000 genes that can code for thousands of different proteins that talk to each other in millions of different ways. Only by using very advanced computation and adding that data to all that we know about the effects of people’s diet, lifestyles, their environment, and their psychology, are we ever going to get to true precision medicine. That is starting to happen now, it will allow us to instead of having two options for a person’s cancer treatment, we will have a bunch of different options and also be able to better identify which sub-type of cancer they have. But we are just at the beginning of being able to identify these sub-types and being able to intervene in someone’s disease process, I believe we are about 25-30 years away from truly getting to precision medicine. But within the next 6-7 years we should have the computational power needed in our quantum computers to really start pushing this field forward.
What impact will quantum computing and machine learning have in helping us better understand the neurodegenerative process and map someone’s disease trajectory?
It will have a significant impact, I’ll give you an example from studies in depression. We get patients to sit inside fMRIs, which allow us to see what parts of the brain are firing together. We take those patterns and try to explain why some patients respond to a given therapy and some don’t. Using mathematical analysis we extracted the network and this enabled us to do a kind of Google search of their brain and out came this finding that specific networks in their brain weren’t functioning properly. We then ran it through some machine learning algorithms to try and predict responses to various therapies.
Now we are also looking at the different states that a person’s brain with a neurodegenerative disease fluctuates between to see some of the effects on the circuitry. This will allow us to identify patterns to better understand what is actually happening. What we need now is data, we need funding for these programs and we need patients to volunteer to be a part of this.