Noisy data facilitates Dartmouth investigation of breast cancer gene expression

Researchers from Dartmouth’s Norris Cotton Cancer Center, led by Casey S. Greene, PhD, reported in Pacific Symposium on Biocomputing on the use of denoising autoencoders (DAs) to effectively extract key biological principles from gene expression data and summarize them into constructed features with convenient properties.

“Cancers are very complex,” explained Greene. “Our goal is to measure which genes are being expressed, and to what extent they’re being expressed, and then automatically summarize what the cancer is doing and how we might control it.”

Normally, it is difficult to apply computational models across different studies because the gene expression data is “noisy,” meaning that there are many factors that differ in the way gene expression is measured. To begin their analysis, Greene’s team added more noise to the data and then trained a computer to remove the noise. To remove the noise, the computer had to learn about key underlying features of breast cancer. “This approach of removing noise makes the models we constructed more generally applicable,” Greene said.

Greene and the Dartmouth team studied DAs, which train computers directly on the data without requiring researchers to provide known biological principles to the computer, as a method to identify and extract complex patterns from genomic data. The model that the computer constructs can then be compared to previous discoveries to understand where data supports those discoveries and where the data raises new questions. The performance of DAs was evaluated by applying them to a large collection of breast cancer gene expression data. Results show that DAs were able to recognize changes in gene expression that corresponded to the cancers’ molecular and clinical information.

“These techniques and findings will enable others to use the DAs to evaluate gene expression data in a variety of disease sites,” reported Greene. “While noise in data is usually viewed as a problem, adding noise to data can actually be a good thing because it can help reveal the underlying signal. When we did this to analyze data from breast cancers, we found gene expression features that generalize across studies and represent important clinical factors.”

Next for Greene’s research team are more complex models that take multiple levels of regulation into account. Their goal is to develop methods that not only model data but that can automatically explain to researchers what the models have learned.

Inherited Gene Mutations


Some inherited gene mutations have been linked to breast cancer. These include mutations in the following genes:

  BRCA1
  BRCA2
  p53
  CHEK2
  ATM
  PALB2

Other genes are under study and may also play a role in breast cancer. 

Inherited mutations known to increase the risk of breast cancer are rare in the general population. These mutations account for only five to 10 percent of all breast cancers diagnosed in the U.S..

BRCA1 and BRCA2 gene mutations

BRCA1 and BRCA2 (BReast CAncer genes 1 and 2) are the most well-known genes linked to breast cancer risk. BRCA1/2 mutations can be passed to you from either parent and can affect the risk of cancers in both women and men.

A person who has a BRCA1/2 mutation is sometimes called a BRCA1/2 carrier.

Like other gene mutations, BRCA1/2 mutations are rare in the general population. In the U.S., between one in 400 and one in 800 people in the general population have a BRCA1/2 mutation [33]. However, prevalence varies by ethnic group. Among Ashkenazi Jewish men and women, about one in 40 have a BRCA1/2 mutation.

###

The BRCA gene test is a blood test that uses DNA analysis to identify harmful changes (mutations) in either one of the two breast cancer susceptibility genes - BRCA1 and BRCA2. Women who have inherited mutations in these genes face a much higher risk of developing breast cancer and ovarian cancer compared with the general population.

The BRCA gene test is offered only to people who are likely to have an inherited mutation, based on personal or family history, or who have specific types of breast cancer. The BRCA gene test isn’t routinely performed on women at average risk of breast and ovarian cancers.

Having a BRCA gene mutation is uncommon. Inherited BRCA gene mutations are responsible for about 5 percent of breast cancers and about 10 to 15 percent of ovarian cancers.

From a BRCA gene test, you learn whether you carry an inherited BRCA gene mutation and receive an estimate of your personal risk of breast cancer and ovarian cancer. Genetic counseling is an important part of the BRCA gene test process.

###

Noisy data facilitates Dartmouth investigation of breast cancer gene expression Dr. Greene and his team of investigators do their research at Dartmouth’s Norris Cotton Cancer Center in Hanover and Lebanon, New Hampshire. Their work is supported in part by NIH funding P20 GM103534 and the American Cancer Society Grant #IRG-82-003-27.

About Norris Cotton Cancer Center at Dartmouth-Hitchcock

Norris Cotton Cancer Center combines advanced cancer research at Dartmouth and the Geisel School of Medicine with patient-centered cancer care provided at Dartmouth-Hitchcock Medical Center, at Dartmouth-Hitchcock regional locations in Manchester, Nashua, and Keene, NH, and St. Johnsbury, VT, and at 12 partner hospitals throughout New Hampshire and Vermont. It is one of 41 centers nationwide to earn the National Cancer Institute’s “Comprehensive Cancer Center” designation. Learn more about Norris Cotton Cancer Center research, programs, and clinical trials online at cancer.dartmouth.edu

###

kirk Cassels

.(JavaScript must be enabled to view this email address)
603-653-6177

###

The Geisel School of Medicine at Dartmouth

Funder -
  National Institutes of Health, American Cancer Society

Provided by ArmMed Media