Skip to Main Content

Zooming in on the Human Genome

June 25, 2022 - 5 minute read

Parker Hicks ’22, a biology major, took a side interest in coding as an undergraduate, taught himself a programming language, then applied for a National Science Foundation research opportunity at the University of Colorado over the summer. There, he created a model that helps genome researchers create high-resolution data from low-resolution data. The work was then published in the industry-leading journal, Bioinformatics, an unusual accomplishment for an undergraduate student.

“It’s very, very uncommon at any university for an undergraduate to be a first author, meaning the one who primarily did the work and the writing,” says Concordia professor Sarah Karam, who was Hicks’ academic advisor on other projects. “To also be in a top journal is less common still. This is rare at Concordia and pretty rare at other places, too.”

Hicks employed methods used in enhancing photo resolutions to enhance data about chromosomal structures, with the hope of finding regions of biological significance.

“The information [we gain] could be used to identify a new gene for something, where that might be located within the genome, and what sorts of interactions could be occurring between genes that affect the expression of one,” he says.

Acquiring high-resolution data is computationally and chronologically expensive, meaning it takes a lot of processing power and time. But it is much more useful to “zoom in really close and see the intricate details of that picture,” he says.

Hicks’ model takes low-resolution data and predicts what a high-resolution image would look like based on what the model has “learned” about how high-resolution data should look.

It was deep waters for someone who had just started programming. Hicks had spent much of his time at Concordia among sea urchins, sharks and crabs as manager of the on-campus marine lab. Then he decided marine biology wasn’t his forte.

Parker Hicks

“It’s pretty stressful to have an entire lab’s worth of animal health on your hands,” he says. “You’re responsible for feeding them, keeping the water chemistry stable, checking to make sure everything’s still alive, cleaning everything, maintaining the equipment.”

He loved the animals, especially the horn shark and swell shark. “They would come up to the surface when you were around, and you could scratch their nose a little bit and hold them if you wanted to,” he says. “They’re only coming up to you because you have food usually, but it’s still cute.”

Drawn to a career in computer science and mathematics, he liked the idea of an interdisciplinary approach to those subjects and biology, so he taught himself how to write in Python, a widely-used programming language, while taking classes in biology and chemistry at Concordia.

“My goals were to get good enough at it to use in my research in biology,” he says. When an opportunity came up to do summer research at the University of Colorado, he applied to the program but didn’t think he would be accepted.

“These programs are very competitive, and I didn’t necessarily think I would get in because I taught myself how to program and didn’t have much experience in the field of bioinformatics,” he says. “But this was a chance to do big research at a large institution.”

To his surprise, he was invited to participate, and he moved to the Colorado Springs campus to work with Assistant Professor of Computer Science Oluwatosin Oluwadare. He began working 12 hours a day, six days a week, on a problem he had never encountered before: how to make low-resolution “images” of the genome sharper by training the computer model on high-resolution data, essentially teaching it to take bad images and turn them into good ones.

“I had no idea what any of this stuff was at the beginning of the summer,” he says.

He started by studying scientific literature about existing models used to enhance the resolutions, and the biology behind those methods. Then he developed his own model and collected a training data set and a test data set. Most of his time was spent troubleshooting programming issues and fixing errors in the code. Eventually, he had a well-trained model.

Karam explains that Hicks’ model takes in low-resolution “chunks” of data about the genome, analyzes the information, and repeatedly applies a filter to pull out areas of high concentration versus low concentration. The data undergoes a cycle of repeated sharpening, re-adding data, then sharpening again with the goal of making the data as high-resolution and accurate as possible.

“The biggest challenge in the beginning was coding because I was relatively new, and this is pretty complex stuff,” Hicks says. “It took me the entire three months to become fully proficient in coding these things.”

He and his summer advisor decided to submit the model for publication, another “entirely new and pretty difficult experience,” Hicks says.

It was rejected by a computer science conference, but the Bioinformatics journal sent back a list of comments with an open door to re-submit. Hicks and his advisor performed more testing and gathered and processed new data sets to answer the reviewer’s questions.

“We had to analyze our results four different ways to make sure the model accurately predicts what’s actually going on in the biological system,” Hicks says.

Their research was then accepted and published. That helped him get into his preferred graduate school, he says.

They teach us to think about science and really question why we’re doing what we’re doing, and why any of this matters.

“They knew I had work under review,” he says. “I’m excited to continue to do cool things like learning new methods of doing these things, publishing, and continuing to develop my skills as a scientist.”

Concordia, Hicks says, prepared him well to think in a comprehensive and creative way.

“Our biology department here [at Concordia] is outstanding compared to other institutions I’ve experienced,” he says. “They teach us to think about science and really question why we’re doing what we’re doing, and why any of this matters at all.”

Karam says Hicks has an uncommon ability to see connections between different fields, take an aspect from one arena and apply it in another. “I’ve not seen another student have quite that trajectory of being able to chart a path that wasn’t laid out,” Karam says. “Seeing the connections between computers and biology and saying, ’Ah, there’s something there,’ and making it happen.”

Hicks will attend the PhD program of the University of Colorado Anschutz Medical Campus for Human Medical Genetics and Genomics.


Back to top