Who you are and what you eat: using AI for personalized nutrition plans

Nutrition has held a life-long fascination for Carmen Tocheniuk, even in childhood. As a young athlete growing up in small-town Alberta, Carmen wanted to know more about how her body used food and how it affected her performance.

"I was an athlete, and my sister was an athlete. We had the same parents and we ate the same food. But, I could recognize that my sister would eat the same foods as me, and she would lose weight, and I would gain weight," she says.

Same foods, different results

That perplexing question — why could two people eat the same thing and see such differences — stayed with Carmen. Nutrition became a big part of her life as she continued athletics, first competing in figure skating and hockey before discovering a passion for bodybuilding. Carefully planning her nutrition became a part of her training, which eventually led to her going pro and becoming a three-time Canadian champion.

In 2005, Carmen started Karmalife, a small company that offers nutrition and fitness guidance. Many organizations provide the same, but Karmalife has a unique angle. Carmen and her business partner, Michelle Mol, were able to secure a used gene sequencer — a device that can analyze the structure of a segment of DNA and spit out information on a person's genetics. It was an opportunity to investigate that question that Carmen had been carrying for decades. Why do different people react differently to the same food?

It is well-established that genetics can influence how we handle food. Research shows that our DNA can influence what we like to eat, how much and how our body metabolizes food. However, those studies examine general trends. Carmen was interested in whether or not individual genetic information could be a part of crafting personalized meal and exercise plans for her clients, something nearly unprecedented at the time.

"We really focus on what will work for the person, not just based on their DNA, but we take in their existing health concerns," she says. "How do they sleep? Do they work out? How much water do they drink: All of these things are important to understand each and every individual."

A wealth of data

That meant ridiculous amounts of data. At first, Carmen was doing things by hand, filling file cabinets with folders of data on her clients, everything from sleep schedules to genetic results. Then she used that data to build thousands of meal plans and monitored the results. That way, she could isolate common factors, including genetics, that influenced a person's nutritional needs.

"We started with three genes. Then five genes, eight genes. Now we do multi-hundreds of genes and have a way better portfolio and package to offer people," says Michelle.

As Karmalife's offerings and client base grew, so did the amount of work that went into collecting and using the data. Too much to do alone if Carmen wanted to achieve her goal of reaching as many people as possible. Then, through some friends, she heard about the Regional Innovation Ecosystems project, which helped guide health and wellness companies looking to adopt artificial intelligence.

Immediately, she saw the potential that machine learning could have with the work that Karmalife was doing, streamlining the process of crunching the data and creating individual fitness plans.

"With the amount of genetic data involved, we just knew that artificial intelligence had to take us to the next level. It wasn't humanly possible for our group to do it without AI at the volumes we are talking about," Michelle says.

It's been a learning process, and Karmalife is still early in its AI adoption. Currently, Carmen and Michelle are focusing on training their system on the data and results they've collected from their clients. But the hope is more than just doing their current work more efficiently. The pair is hopeful that artificial intelligence can also help identify new patterns in the genetic sequences they collect. And that might lead to more discoveries on the relationship between health and nutrition.

AI promises a wider reach

The future impact could be massive, they say. Healthcare resources are finite, and if proper nutrition can help people prevent health problems later in life, it could take pressure off a system already strained for resources.

Carmen adds that AI seems to be the only way for Karmalife to achieve her ultimate goal: bringing individualized health information to people on a wide scale. She wants people to understand their unique nutritional needs based on their genetics, habits, and other factors. So hopefully, they can find the answer to the same questions she's been asking since she was a child.

"Our dream is to truly assist doctors, assist nutritionists, assist people directly, all based on their own DNA. We weren't a true solution before; we were helping one person at a time.

I guess my endgame is to help as many people in the world as I can. And without AI, it wouldn't be possible."

This project was part of the Western Economic Diversification Canada Regional Innovation Ecosystems (RIE) program. The initiative brought together nine organizations from non-profit, business and academia to establish viable uses for artificial intelligence and machine learning in health and data analytics.