A panel of industry experts say that advanced computing techniques can solve some of the toughest problems in biotechnology, such as combing through massive datasets and modeling drug interactions.
Shannon Bruse, co-founder and head of discovery for a biotech firm called Empirico, explained during a recent panel discussion that his company integrates computing at nearly every level of operation. The virtual event was hosted by UW-Madison’s Innovate Network and the Forward BIO Institute as part of Madison’s Forward Festival.
“What we do is combine genetic data with health information from millions of individuals and we can learn about the effects of altered gene function and we can design drugs that mimic those human mutations,” Bruse said yesterday. “The problem is, the scale of the data we’re dealing with.”
Because Empirico combines information from so many individuals and disease variants, straightforward analysis can lead to billions and even trillions of data points, making it nearly impossible for humans to understand the results at face value. But by applying machine learning models to organize the results, Bruse said his team can reduce those numbers to the thousands, making results much more accessible.
“In the end, it still takes a human with domain expertise to make the call,” he said. “But that’s the way that our computing helps, because the scale of our data is just so massive.”
Jalal Sulaiman, president and CEO of PROMISS Diagnostics in Wauwatosa, explained that artificial intelligence algorithms can help diagnosticians make connections between disease indicators that might be difficult to otherwise identify.
“Because of that advanced computing, machine learning techniques, you are able to analyze such a large set of data and extract information and develop insights,” he said.
Aside from parsing through enormous amounts of complex data, panelists also noted that advanced computing can model the way that different treatments will interact in a patient when given simultaneously. Joseph Grudzinski, co-founder of Voximetry and senior scientist in UW-Madison’s Department of Radiology, pointed to the example of radiopharmaceutical therapy and immunotherapy.
He said new computing techniques could shed light on how biological targets for these therapies might differ or overlap. Without using some form of computing techniques, getting to this level of understanding could take years or even decades, he noted.
“I think being able to use biocomputing to sort of model the biophysical drug interaction … could allow us to provide better understanding for how these combination therapies work, so that we have a better way of possibly combining them for safety and efficacy to allow for the best synergy possible,” Grudzinksi said.
See more Forward Festival events being held this week: https://forwardfest.org/events/2021
