SALT LAKE CITY — Veronica O'Connor lifted the glass divider just enough for her arms to reach through the bottom, into a vented lab workspace behind the glass.
A round glass boiling flask sat on a machine, spinning nickel in an organic solvent. The first-year graduate student lifted up a small piece of paper, with a small pile of an organic molecule piled on top.
The powdery-looking substance slid off the paper, into the flask. The solvent began to change color to a blueish-green.
"No one’s ever done this before, which is really exciting," O'Conner said of the chemical reaction she was creating.
O'Connor, who is studying alternative energy storage materials, is working on a project for the Department of Energy. The tests she's running could translate into advancements for household solar panels as well as wind turbines.
"Solar panels or wind power, anything that's a renewable energy," she said. "We are trying to find a way to store that energy when the wind isn't blowing, and when the sun is not shining."
Her experiment in the lab is all thanks to computer models, O'Connor explained. The work starts on the computer screen with predicting, computing, and finding models, before testing them out in the lab.
It's called "machine learning."
It's that aspect of the science world that the University of Utah is making a name in, after unveiling -- or should we say, releasing -- a database to be used around the world.
"Release the Kraken," said Matthew Sigman, a distinguished professor and the department chair for the U's Department of Chemistry. "It’s so big and unique in the field."
It's called Kraken, and no, it's not the giant cephalopod that rises from the seas.
Kraken is a monstrous system, a map if you will, of 300,000 calculated organic compounds that the U of U research group created in partnership with the University of Toronto.
Sigman said former U of U postdoctoral student Tobias Gensch started the whole project. Sigman oversaw it. It allows scientists and researchers to log on, hover over thousands of points on a chart, and see different compounds generated by the computer.
"Whereas the old workflow was much more by hand. We’d have to guess a little bit, guess and check," Sigman explained. "We will not have to guess and check at all anymore."
Kraken is beyond what any scientist's brain could come up with, Sigman indicated. More than any grad student could measure out and test in a lab.
"What we’re doing is predicting molecules now. A lot of them that are outside of what has been explored in the history of this field," he said.
While this may seem foreign to most people, the application of Kraken won't. Take, for example, the fact that the University of Utah collaborates with pharmaceutical companies on medications.
Kraken will allow them to create new drugs at rapid speed. Sigman said the system will help drugs get to clinical trials faster.
"We're mostly in the area of small molecule drugs. So, stuff that you're more familiar with like aspirin or Lipitor or something like that. We find ways to help companies do this much more accelerated," he said.
And it all happens at the click of a mouse.
"What Kraken does is save an enormous amount of time between where you start and where you end in trying to make a chemical process efficient and cost-effective," Sigman said. "And you'll bring the drug cost down, you will bring the timeline by which they will be evaluated in trials down."
Sigman described how the same kind of machine learning technology used for the COVID-19 vaccine. It's how the vaccine was able to get to trial so quickly, he indicated, leading to an accelerated release to the public.
Of course, lab work is still very much essential to the process.
And if you ask O'Connor, it's the part she loves.
"So that’s the exciting part for me," she said of her experiment. "I have no idea what it's going to look like."
That's where the fun happens — with help from AI technology, at the department's fingertips.