This study included larger groups of people from African, Hispanic and Asian ancestries than many other previous studies.

A recent study involving scientists from the U.S. Department of Energy’s (DOE) Argonne National Laboratory has uncovered insights into the prostate cancer risks of people from a variety of genetic ancestries.

The project, which was led by the University of Southern California, included large increases in representation among men of African, Hispanic and Asian ancestries, that were contributed in part by an ongoing collaboration between the U.S. Department of Veterans Affairs (VA) and DOE as part of the Million Veteran Program Computational Health Analytics for Medical Precision to Improve Outcomes Now (MVP-CHAMPION) project.

Ravi Madduri, an Argonne researcher who is a senior author on the project, spoke during an interview about the paper about the limitations of previous studies.

“There have been a lot of studies to understand the relationship between the risk for prostate cancer and the genetic makeup of a particular individual,” Madduri said. ​“Most of the studies have been done in the context of people who are of European ancestry, and very little is known in terms of the contribution of genetic ancestry of other peoples to their risk for prostate cancer.”

While simply focusing on different ancestries made this project stand out, the truly unique factor of this research was the data provided by the MVP.

The MVP is one of the largest and most diverse collections of genetic information in the world, with one million veterans joining since the program’s launch in 2011.

The study of this genetic data was performed using the latest techniques in high performance computing (HPC) and artificial intelligence (AI), and this unprecedented biobank enabled researchers to reach further than any that came before it.

Studying populations

Prostate cancer is the most frequently diagnosed non-skin cancer globally, and as such, it is a problem that affects the entire human population.

To understand the likelihood of an individual or population developing prostate cancer, researchers create what is known as a Genetic Risk Score (GRS).

This is an estimate of the cumulative contribution of genetic factors to a specific health outcome.

“You look at the genetic sequence and compare the variants that you find in the genetic sequence to the variants that are already found that have an impact on the particular disease,” Madduri said. ​“And then you calculate the risk for that particular person.”

Specifically, the research group focused on single nucleotide polymorphisms (SNPs.)

These are variations at a single position in a DNA sequence that act as biological markers that help scientists locate genes that are associated with disease.

For this study of 156,319 prostate cancer cases and 788,443 control cases, the GRS is calibrated by adding up all the risks contributed by individual SNPs.

Here, the researchers found 187 new genetic markers that provide more accuracy in calculating the risk posed by a particular ancestry.

This increased the total number of known risk variants to 451.

Madduri stated that his background as a computing scientist has him working in various domains applying HPC to accelerate scientific discovery.

Work like this performed with the giant MVP database is ripe for HPC optimisation, as the scientists need to look at each genetic sequence and compare any discovered variants with previously discovered variants to determine if they will have an impact on a particular disease.

“The current algorithms that are typically used for this process can run about a thousand individuals at a time, and they are kind of slow,” Madduri said. ​

“They are not created or designed to run on hundreds and thousands of participants in parallel together. So, Alex Rodriguez and I have created a high-performance pipeline that calculates the risk variance for all of the participants in the cohort in a very short amount of time.” 

Alex Rodriguez is a computer scientist at the University of Chicago collaborating on this project.

Different ancestries, different risks

As Madduri pointed out, this work was centred around a desire to understand the genetic makeup that increases risks for diseases like prostate cancer.

Of course, genetic architecture is heavily based on a person’s ancestry, and previous studies did not include people who do not share a European ancestry.

This is especially true of people with African ancestry.

In this study, the case sample size was increased by 43% for people of European ancestry over previous studies, 26% for those with Asian ancestry, 45% for Hispanic groups and a whopping 87% for those with African ancestry.

This inclusive approach enabled unique discoveries. For instance, in terms of the GRS for the current 451 risk variants, men of African ancestry showed a greater risk of aggressive versus non-aggressive disease.

This increase of representation is obviously important for groups that are often marginalised and forgotten, but it also helps people outside of these groups.

“This study not only provides more insight into the genetic architecture of people with African ancestry, but it also provides a large population that helps validate genetic findings from other populations,” Madduri said.

Science involving such a large database like MVP can sometimes feel impersonal, as the individual can often be lost in such a large group.

But for researchers like Madduri, the contributions of every person involved is of the utmost importance — not only for what they contribute to science with their data, but also in terms of what they’ve provided for their country.

“Prostate cancer is the second-biggest killer of our nation’s veterans,” Madduri said. ​“For the people who have given everything for our country, it is our duty to help them when they need it.”

Journal: Nature

Source: DOE/Argonne National Laboratory