How one Parkinson’s patient drove a groundbreaking research effort

Passion // Project

It’s more than a job. These compulsively creative IBMers make innovation a way of life, creating amazing projects just for fun.

Share this page:

How one Parkinson’s patient drove a groundbreaking research effort

Jonathan Rezek

Business Development Executive,
National Innovation Team, Canada


His passion project

Improve quality of life for Parkinson’s patients

Discover what you can
do at IBM

Jobs at IBM

In Jonathan Rezek’s day job, speaking to large groups of people is a pretty ho-hum event. So when Rezek had an anxiety-driven “meltdown” (as he puts it) on stage five years ago in mid-presentation, he knew something unusual was going on.

“That was one of the early clues that I had Parkinson’s,” says Rezek, now a business development executive for the IBM National Innovation Team in Canada. “People don’t realize it, but massive anxiety is one symptom of this disease.”

By last year Rezek, whose job includes helping startups incorporate Watson into their business environments, knew the technology was being used in some areas of medical research. He wondered if Watson’s nearly boundless capacity for ingesting and analyzing data could be applied to help Parkinson’s patients like himself.

A teachable moment for scientists

Rezek broached the topic with his neurologist, Dr. Connie Marras. She brought the idea to her colleagues in research at the Morton and Gloria Shulman Movement Disorders Centre at the University Health Network.

“They were intrigued by it, but they didn’t really know much about it,” he says. “These are clinicians and scientists and their domain is wet labs. They do organic studies, not data digging. So I said, ‘Well, look, if you’re interested, I can find someone to talk with you about Watson and how it works with scientific research.’”

That subsequent call led to a face-to-face meeting in Toronto in early 2016, which Rezek eagerly facilitated. “I was basically the emcee. They had a lot of questions, and Dr. Alix Lacoste, who spearheaded the project from the IBM side and deserves a boatload of credit, worked tirelessly. She was so patient.”

Lacoste and her peers walked the researchers through several demonstrations of Watson for Drug Discovery’s capabilities, and together they came up with a strategy to put Watson to work. Her team also presented its ideas to the Ontario Brain Institute, which agreed to fund the initial study.

Dr. Naomi Visanji, a research scientist at the Morton and Gloria Shulman Movement Disorders Centre, was excited by the prospect of adding AI to the lab’s toolkit. “There are around 10 million people in the world with Parkinson’s, and that population is growing.”

Jonathan Rezek and the team of researchers working with Watson for Drug Discovery talk about the process of using AI to find a better treatment for Parkinson’s disease.

Jonathan Rezek and the team of researchers working with Watson for Drug Discovery talk about the process of using AI to find a better treatment for Parkinson’s disease.

Watson for Drug Discovery

Watson for Drug Discovery enables breakthrough insights by analyzing millions of articles and other corpus data in minutes

Three core components of Watson for Drug Discovery


Rich visualizations Rapid learning


Understands natural language

Evaluates and generates hypotheses

Learns and evolves over time

Deep NLP & text mining
Machine learning
Domain adaptation

Knowledge and data sources



Structured databases

Big data — structured and unstructured
Scalable architecture

The treatment is almost as bad as the disease

Parkinson’s disease is a chronic, neurodegenerative movement disorder caused by the progressive deterioration of neurons in the brain. Those cells are responsible for producing dopamine, a chemical that enables your brain to control movement throughout your body. Without enough dopamine, patients exhibit symptoms ranging from tremors, muscle rigidity and weakness to severe pain and a gradual wholesale loss of voluntary movement.

There is no known cure and the only drug approved to treat the symptoms, L-DOPA, has been around since the 1960s. An artificial dopamine, L-DOPA can ease the worst effects of the illness but it’s no dream solution.

“All the current treatments we have are symptomatic therapies. We’re literally replacing a chemical in the brain that's missing in people with Parkinson’s, says Visanji. “That’s effective for a period of time, but after a year or two, these therapies can start to cause debilitating side effects — unwanted movements, or dyskinesia. They start to writhe when they move like Michael J. Fox does nowadays. Those uncontrolled movements he has, that’s not the Parkinson’s, that’s a side effect of the medication.

“Up to 90 percent of patients experience those side effects after 10 years or so of treatment, and many find them as debilitating as the Parkinson’s itself. We’re in a catch-22, and it’s really quite tragic that this far along we don’t have anything better to offer.”

One of the problems with developing new treatments is the sheer cliff a researcher must climb to get a new drug on the market. “From conception to the time a new drug gets into a clinical setting takes on average 10 years and costs over $1 billion,” says Visanji. “Only about 1 in 10,000 drugs that gets invented actually becomes a treatment given to patients. The odds are heavily stacked against success.”

Repurposing allows for shortcuts

One strategy that’s found good success is taking a drug developed for one condition and repurposing it to treat another.

We’re in completely uncharted territory, which is where we need to be

Dr. Naomi Visanji


Top 5% of ranked candidate drugs comprised

7 candidates with proven antidyskinetic efficacy

25 candidates with mechanism of action (MOA) related to drugs with proven antidyskinetic efficacy

6 candidates with novel plausible antidyskinetic MOAs

139 drugs rejected due to MOA predicted antiparkinsonian actions, toxicity, lack of BBB permeability

“The example I sheepishly use is Viagra,“ says Rezek. “Viagra was originally designed as a heart medicine, and they were ready to kill it because guys couldn't lie on their stomachs. Then someone said, ‘Hey, this actually solves another problem.’ So the drug didn’t work out as a heart medicine, but it had already proven safe for people to take, and now it’s one of the biggest drug successes Pfizer’s ever produced.”

I’d like nothing better than for IBM to build a Watson that helps solve neurodegenerative disease

Jonathan Rezek

Thousands of drugs are approved for human use by both the FDA and HealthCanada. These drugs have been through extensive testing for safety, so if one of them shows potential for treating a different condition, the process of testing for efficacy against a new disease is a much shorter, less costly affair.

Visanji’s team worked with IBM to narrow the scope of the study to 3,500 drugs that met some basic criteria for impacting the neurological processes involved in Parkinson’s.

Watson goes to school

“The first step was training Watson to understand the problem — that we give Parkinson’s patients this drug that helps them, but eventually it causes this awful side effect, dyskinesia, and we want to stop that from happening,” she explains. “We took a list of the drugs and chemicals we’ve been researching for the past 40 years that help with dyskinesia in some way in animal studies. We’re telling Watson these drugs are related in some way and we’re interested in what they do. Then we fed in the set of candidate drugs — the 3,500 drugs we want to look at — and in a matter of minutes Watson was able to filter all the information available and point us in new and exciting directions.”

“What Watson is able to do that we can’t do is read the scientific literature at an alarmingly fast rate and draw parallels between related information,” says Visanji. “So we’ve got millions of abstracts and documents within the medical literature that would take us forever to read. Watson can read them all and digest them very quickly. It can understand the language, it can absorb medical literature, it can absorb patterns, it can absorb genetic information. And it’s putting all of this stuff together in a way that our human brain is not able to do.”

Dr. Naomi Visanji
and Jonathan Rezek

With its predictive analytics capabilities, Watson can do much more than look at the drugs in terms of their chemical structure. “That on its own wouldn’t be very sophisticated,” says Visanji. “It looks at the words and the phrases in the literature that are associated with the drugs. Watson built a very powerful and deep comparison between our candidate drugs and the training set.

“What we want is for Watson to draw new links between things we haven’t ever thought of, and it’s able to do that because it can read the entire medical literature in 10 minutes flat. I haven’t met a person who could do that.”

The eureka moment

Left to right:
Drs. Alix Lacoste
and Naomi Visanji

Watson put the drugs in order of the best to worst match to solve the problem, and Visanji’s team set to work sifting through those data. “We looked at the top 5%, and among those were seven drugs that we already know have some potential to treat dyskinesia, so the fact that Watson rated them highly gave us confidence that Watson understood the question.

“We were looking for the complete eureka moment, that thing that no one has ever thought of before. And after sifting through all the drugs, we’ve come up with six candidates that completely fit that bill, and we’re very excited about them. Their mechanism of action has never been thought of for dyskinesia, but Watson thinks they’ll be anti-dyskinetic, and we trust Watson because of the other drugs it picked that we know are very plausible.”

The first of Watson’s six picks is undergoing lab testing. “It’s a drug that’s already FDA approved, so it could quite easily be repurposed for Parkinson’s disease. And it’s showing some very exciting preliminary evidence in our animal models that it’s working.

In January 2018, Canadian Institutes of Health Research (CIHR) awarded the team a grant to cover the costs of testing the remaining top candidates in a well-established rodent model of dyskinesia, and performing additional pharmacological studies that will ease the translation of any findings to future studies in primates.

“I’ve been in Parkinson’s research for 15 years,” says Visanji, “and this is the first time where I can see a clear path to getting something new into the clinic.”

The long game

Ideally, of course, Visanji and her colleagues would like to find a cure for Parkinson’s. And they are also working with Watson to try and get at the root cause of the disease.

Researchers know that clumps of the protein alpha-synuclein are present in the brains of Parkinson’s patients. A second study asked Watson to predict drugs that might prevent those protein clumps.

Watson generated a ranked list of drugs that might be effective, and Visanji says those will be tested in the lab, but “the caveat is how long is it going to take to get into a clinical trial, and how long are those trials going to take? One of the reasons there isn’t a cure for neurodegenerative diseases like Alzheimer’s and Parkinson’s is those experiments are so much more complex and they take a very long time. Proving that a drug has an impact might take years and years, and it’s challenging to find someone to fund that.”

“That’s really the long game,” she adds. “I see what we have done to date with the dyskinesia project as something that could realistically impact and improve patient quality of life in a matter of three or four years. It’s not the solution, but it’s something that can help until we do have a cure.”

The patient as muse

Visanji adds that Rezek himself has been hugely motivational for her team. “You can’t undervalue the patient experience, and the courage it took for Jonathan to come forward with this idea, and everything he’s done in the last 18 months to drive this project.”

“He is a force of nature,” agrees Lacoste. “Without Jonathan there would be no project. He’s determined, and I’m grateful to him for persisting with this. I hope it will make a difference in his life.”

“This is my passion project,” says Rezek. “There are a lot of neurodegenerative diseases that are similar — Parkinson’s, Alzheimer’s, ALS. I’d like nothing better than for IBM to build a Watson that helps solve neurodegenerative disease.”

Rezek devotes as much energy as he can spare to fundraising efforts to support the necessary lab validation studies and clinical trials. “What I don't think is often recognized is that even with Watson identifying potential drugs that can be repurposed, the basic research has to continue. You have to do the testing, in a petri dish, in animal studies, in clinical trials. It’s expensive.

“It’s easy to get demoralized, but I’ve met some incredible people who all want to help. For me, it’s a tug of war between optimism and fear. Optimism is fine. Guarded optimism is better.”

Share this page: