Analytics in Action News

Blood Samples, Predictive Analytics Determine COVID-19 Outcomes

By taking blood samples, researchers can use predictive analytics to determine a COVID-19 patient’s survival rate.

predictive analytics COVID-19

Source: Getty Images

By Erin McNemar, MPA

- A new study of blood samples from nearly 200 COVID-19 patients reveals underlying metabolic changes that regulate how immune cells respond to the infection. These changes are associated with disease severity and, with predictive analytics, could determine patient survival.

COVID-19 patients have different immune responses that impact disease outcomes ranging from asymptomatic infection to death.

“We analyzed thousands of biological markers linked to metabolic pathways that underlie the immune system and found some clues as to what immune-metabolic changes may be pivotal in severe disease,” co-first author Jihoon Lee said in a press release.

“Our hope is that these observations of immune function will help others piece together the body’s response to COVID-19. The deeper understanding gained here may eventually lead to better therapies that can more precisely target the most problematic immune or metabolic changes.” 

The researchers collected 374 blood samples, two draws from each patient during the first week of their COVID-19 diagnosis. The team then analyzed the plasma and single immune cells from the samples. The analysis included 1,387 genes involved in metabolic pathways and 1,050 plasma metabolites.

In plasma samples, the teams discovered an increased COVID-19 severity associated with metabolite alterations, indicating increased immune-related activity. Additionally, researchers found that each major immune cell type had a distinct metabolic signature through single-cell sequencing.

According to co-first and co-corresponding author Yapeng Su, PhD, the researchers found that metabolic reprogramming is highly specific to individual immune cell classes and cell subtypes. Additionally, metabolic reprogramming of the immune system is associated with the plasma global metabolome and, with predictive analytics, can determine disease severity and even death.

 “Such deep and clinically relevant insights on sophisticated metabolic reprogramming within our heterogeneous immune systems are otherwise impossible to gain without advanced single-cell multi-omic analysis,” Su said.

The research will serve as an important tool in developing more effective treatments against COVID-19. It also represents a major technological hurdle, according to Jim Heath, PhD, president and professor at the Institute for Systems Biology (ISB) and co-corresponding author on the paper.

“Many of the data sets that are collected from these patients tend to measure very different aspects of the disease and are analyzed in isolation. Of course, one would like these different views to contribute to an overall picture of the patient. The approach described here allows for the sum of the different data sets to be much greater than the parts and provides for a much richer interpretation of the disease,” Heath said.

The research was conducted by ISB, Fred Hutchinson Cancer Research Center, Stanford University, Swedish Medical Center St. John’s Cancer Institute at Saint John’s Health Center, the University of Washington, the Howard Hughes Medical Institute.