• News
  • Genetic Networks

Paper sheds new light on genetic risk factors for breast cancer

by Eva Voinigescu Oct 13 / 17
Banner image: Mammary Gland Mast Cell Tumour via iStock

Although we know of about 100 genes that play a role in breast cancer, the majority of genetic factors in breast cancer risk remain unknown. Now new research has uncovered 25 biological pathways that are implicated in hereditary breast cancer, and could point the way to uncovering many more genetic interactions with implications for health.

“These genetic interactions basically define subtypes of the disease and as soon as you can subtype individuals that can be relevant for how you treat them,” said CIFAR Senior Fellow Chad Myers, a professor at The University of Minnesota and a co-author of the study.

Approximately 5 to 10 per cent of breast cancers are thought to be hereditary, caused by genes including BRCA1 and BRCA2. But known genes only explain about a third of the genetic contribution to breast cancer risk. The remaining genetic factors could be explained by the combined effects of multiple genes whose genetic interactions remain unstudied.

The research is outlined in new pair of papers describing BridGE, a computational method for identifying genetic interactions that impact disease risk. The papers are written by researchers at The University of Minnesota and The University of Toronto, including Gang Fang, Vipin Kumar and CIFAR Fellows Myers and Charles Boone. The researchers used BridGE to evaluate genetic data from six cohorts of breast cancer patients, and were able to identify 25 pathways where genetic interactions implicated in hereditary breast cancer risk were occurring.

Historically, statistical analysis of gene interactions tests pairs of gene variants or mutants. In humans the potential pairings amount to more than half a trillion, presenting a significant challenge for computational statistics.

“Formulating the problem like that where you’re testing individual pairs of mutations is statistically doomed from the beginning unless we have cohorts of patients that are hundreds of thousands of people,” said Myers.


In the between pathway model, variations in two pathways that share a function (like hormone production) can result in disease. If only one pathway has a variation, the other may make up the difference. In the within pathway model, one variant will decrease the pathway’s ability to perform its function but two variants could prevent it from doing its job all together, leading to disease. The hub pathway is one where a central gene or genes interact with many others potentially increasing or decreasing disease risk.

Myers, Boone and their colleagues were able to circumvent this problem by building on crucial learnings about yeast from work done by Boone, the co-director of CIFAR’s Genetic Networks program, and Senior Fellow Brenda Andrews. The Genetic Networks program explores how interactions among genes influence health and development and how we can translate genetic knowledge across species. 

“What we find in yeast is that genetic interactions appear in these big groups together,” said Myers. “The foundation for [BridGE] was to change how we search for genetic interaction from human population data to reflect the fact that we expect them to be structured.”  

As the researchers describe in the paper Discovering genetic interactions bridging pathways in genome-wide association studies, BridGE takes existing genome-wide association study data and searches for three types of patterns between and within biological pathways. Pathways are a series of actions within a cell that lead to a certain change such as turning a gene on or off. Because the number of possible pathway pairs in the human genome is significantly less than the number of gene pair combinations, researchers can extract statistically significant information about which ones play a role in diseases like breast cancer.

In a second paper published in PLOS Genetics last month, the researchers used BridGE to analyze data from six groups of breast cancer patients distinct in their disease type and spanning multiple ethnicities including European, Japanese, Latina, African American and Chinese. They identified 25 pathways that were major modifiers of hereditary breast cancer risk, nine of which included genes that had previously been identified as having an impact on breast cancer risk.

One of the pathways found to be involved in a number of modifying genetic interactions among European patients was the steroid hormone biosynthesis pathway. This pathway controls the production of steroid hormones such as estrogen, progesterone and cortisol.

“Steroid hormones are growth signals for many tumours, so from that perspective the fact that this pathway shows up isn’t surprising,” said Myers.

“This gives people new leads on what to look at for the genetic causes of breast cancer”

While the research found that certain genetic interactions were distinct to the cohorts who shared a particular type of breast cancer, they also found a core set of interactions that cut across multiple types, such as in the Glutathiome Conjugation pathway which works to process and eliminate toxins in the body. It is also involved in modulating other signaling pathways that control growth in cells. Most of the enzymes within this pathway had not previously been identified as genetic indicators of breast cancer risk.

“This gives people new leads on what to look at for the genetic causes of breast cancer,” said Myers. He thinks understanding the mechanisms of the Glutathiome Conjugation pathway is an important area for further investigation.

Meanwhile, the team has already also applied the BridGE method to data from patients with Parkinson’s disease, schizophrenia, prostate cancer hypertension and type 2 diabetes.

“That’s one of the benefits of the CIFAR network. It makes us bold about crossing these very traditional scientific boundaries. We took results in yeast that everyone knows about but that no one in the human genetics community was picking up and doing something useful with. This is a nice success story where you do that and all of a sudden you can solve a problem that is unsolvable just by leveraging insights from model species,” Myers said.