Determination would be made using person’s genetic make-up, medical history
by Caitlin Randle, MRT.com/Midland Reporter-Telegram – April 16, 2020
A Midland data scientist and his two partners have created an algorithm that uses a person’s genetic markers and medical history to predict someone’s likelihood of becoming infected with the coronavirus and suffering complications from it.
Midlander A.J. Rosenthal and his partners, Dan Brue of Oklahoma and Warren Gieck of Alberta, Canada, filed patents this week related to the algorithm.
Rosenthal said it could use a person’s genetic make-up in combination with various factors, such as their medical history and types of exposure they’ve had (i.e. a miner exposed to coal dust), to determine someone’s risk factor and assign them a correlating score.
“We’re describing potentially where a person would fall, give them a score, and that score allows them to either start going back to the workplace because they’re not going to succumb to the disease, or they won’t even be susceptible to it,” he said.
The algorithm would use the medical histories of those who have been hospitalized with COVID-19 to determine what markers could put a person at risk, Rosenthal said. He described inputting the data from past patients as “training the algorithm.”
The goal of this project is for the information to be widely accessible, Rosenthal said. He said the algorithm could potentially be on a website where a person could enter their medical information after signing a HIPPA privacy release.