OLATHE, Ks. — An Olathe North High School freshman's research on Alzheimer’s disease and breast cancer is being recognized for its important new insights into the diseases.
“I’m especially passionate about breast cancer research and Alzheimer's research," said Mahi Kohli, 15. "Just doing that and coding and learning about biology. It’s my dream.”
Kohli earned first place awards in multiple science competitions in the Midwest for her research
She earned a first place award in computational science at the Greater Kansas City Science and Engineering Fair.
Kohli will present her findings next month in Los Angeles at the Regenereon International Science and Engineering Fair.
“This is a real problem," Kohli said. "You can meet anyone, and they know two or three people who have this disease."
KSHB 41 met Kohli at the high school.
“I love math and science and I want to do an intersection of those, and for me, that’s computational biology,” she said.
She uses mathematical code to find out about biological diseases, specifically Alzheimer’s disease and breast cancer.
The inspiration comes from women in her life, including one of her mentors.
“She was diagnosed with breast cancer when I was in eighth grade,," Kohli said. "She was able to get it treated, they caught it early, but her process made me want to look into how that disease is diagnosed."
She found a way to increase the accuracy of one type of breast cancer test.
Kohli compared optimal machine learning models of cancer cells.
“The reason why some people may get a late diagnosis, currently the “gold standard” of diagnosis is a biopsy, but they are expensive and invasive,” she said. “But there is another method called a fine needle aspiration biopsy (FNAB), that’s a minimally invasive procedure, but it only has a 77% accuracy rate. So that was super low.”
Kohli went to work.
“The goal of that study was to see if we could make FNAB more accurate,” she said. “We found pictures of cells. We developed a machine learning model that could look at a cell and say, 'OK, based on the cells feature you either have cancer or you don’t.’”
The accuracy rate on the test soared in her research.
“We found out one of the models had over a 99.26% accuracy rate, which was really high," she said. "That means we can improve the FNAB accuracies up to the standard of biopsies, which means that people can get cheaper or more affordable, more accessible diagnoses."
Mahi’s freshman project received the Grand Award (1st Place) in the senior division of the science competition and she was awarded a scholarship.
Olathe North High School Principal Jason Herman knows Kohli's work is impressive.
“What Mahi’s done in the last couple of weeks is amazing,” Herman said. “We’re not surprised by it, but we are super thrilled that she gets the experiences she’s getting. I can absolutely not speak any language to what Mahi is capable of doing. I know it has to do with proteins, but that’s as far as I can go.”
She also continues her work in Alzheimer's disease.
“I met this woman, Miss Rose Jackson," Kohli said. "We’d become great friends, but I noticed she’d forget what we’d did or what we talked about."
Her next goal with her code was to develop a model to quickly and cost-effectively identify Alzheimer’s inhibitors.
“In our brains, there is this protein called Acetylcholine," she said. "What that does is it helps communication with other neurons, it helps with cognition and communication. It’s really important. “But there is also this protein called Acetylcholinesterase. We can think of it like a little Pac-Man and it starts to eat up that Acetylcholine. So, what’s problematic in Alzheimer’s patients is that the level of Acetylcholine is already low because of tangles around neurons. But what’s wrong is that Acetylcholinesterase, while it may be helpful in normal brains, it starts to eat up that already low level of Acetylcholine. That rapidly hurts people's cognition and memory who are affected by Alzheimer’s.”
Kohli found a data set where there was already some information on potential inhibitors. She analyzed it and found a high accuracy rate with a prediction whether a molecule could be an inhibitor of that Acetylcholinesterase.
“We then tested 100,000 molecules that we didn’t know if they could inhibit Acetylcholinesterase or not. We tested the decision tree model. "We found 221 of those molecules had a really high potential to be inhibitors, and then of those four of them had extremely high potential to be inhibitors. So, this project can be kind of used to advance therapeutics in the Alzheimer’s diseases.”
Her mentors, and her school and community are cheering her on.
“She’s put herself in a position of prepping herself, preparing herself for things, “Herman said. “You don’t see that happening in juniors or seniors and those are still the top of the top and she’s only a freshman. "If this were a basketball game, our student section would be cheering saying, 'She’s a freshman!' So we're super proud of her and can’t wait to see what she does.”
