Skip to main content
Premium Trial:

Request an Annual Quote

At ASHG, UWash, Color Genomics Launch Free Database of Breast Cancer Gene Candidate Variants

VANCOUVER, British Columbia (GenomeWeb) – A team of researchers from the University of Washington and Color Genomics this week introduced a publicly available database of variants in 27 known and candidate breast cancer genes from cancer-free women over 70 years old.

The database, nicknamed FLOSSIES for the "Fabulous Ladies Over Seventy" whose DNA was sequenced as part of the project, is intended to provide the research community with a resource of control cases to aid in the evaluation of variants of unknown significance in breast cancer — a resource that is sorely needed, according to its developers.

UW breast cancer researchers Mary-Claire King and Tom Walsh discussed the new database in a presentation and scientific poster at the American Society of Human Genetics conference held here this week. 

Dozens of genes and their variations have been suggested as having a role in breast cancer and ovarian cancer, but the strength of evidence for their role varies greatly. The role of some mutations, such as those in BRCA1 and BRCA2, are well known in breast cancer, while the role of variations in genes like RAD51C and RAD51D are less clear in breast cancer, but have a known tie to ovarian cancer susceptibility.

One way to distinguish pathogenic from benign variants in these genes is to characterize variant frequencies in appropriate control subjects — for instance, older women who have not developed breast cancer — then compare that with frequencies found in breast cancer patients.

And, as the researchers explained in their poster, although public databases exist that include exome sequences from thousands of individuals, very few of those sequences are from older women known to be free of breast or ovarian cancer.

"There isn't an appropriate control series out there in the world yet," King told GenomeWeb in an interview. "Some very large series of exome data have been evaluated in many different groups … but the difficulty there is that although they have exome-level information by age and by gender, they don't stratify on it. ... So, you have no idea if [a variation is] occurring in an elderly woman who is well, and that's terrific and an indication that this is an easily survivable variant without developing cancer."

In order to address this, the UW and Color researchers tapped into DNA samples obtained from women who between 1993 and 2005 joined the Women's Health Initiative, a major, long-term research program coordinated by Fred Hutchinson Cancer Center and designed to investigate cardiovascular disease, cancer, and osteoporosis.

The WHI "followed women really well, assiduously over years," King said. "Out of 160,000 women who ... gave DNA and said it could be used anonymously for women's health studies … the WHI staff selected for us 10,000 women at random after stratifying out women who were at least 70 years old as of last year and had never had cancer."

The initiative also stratified women with respect to race: 7,000 women who self-identified as European American and 3,000 who self-identified as African American.

Researchers from King's lab at UW and from Color Genomics then divvied up efforts to sequence 27 known and candidate breast cancer genes from those samples. The UW team used Illumina HiSeq 2500 platforms running BROCA, a gene panel originally developed by UW's Walsh; while Color Genomics used Illumina NextSeq platforms running Color's own commercial gene panel, which is mostly based upon BROCA.

"We used the same approaches … and almost the same but not exactly the same panel," King said. "We each took 5,000 [samples] and jointly sequenced a few dozen to make sure we were getting the same thing. Then we traded all of our data … and independently evaluated all the variants and decided on what all the calls were. These are just technical decisions about calls to be sure that we have good coverage. We're not making judgments about the pathogenicity of any genes; we're simply indicating what they are."

King, who is an unpaid advisor to Color Genomics, noted during an ASHG presentation that FLOSSIES is currently "undergoing dynamic revision," but later told GenomeWeb that it now contains "the great majority, 99-plus percent, of the variation that any investigator would encounter in exploring risk associated with one gene or another with respect to women's cancers."

Having just wrapped up the sequencing for the project in the last few weeks, and with the database going live this week, the researchers are already considering how they might expand and improve it. For one thing, King noted that the FLOSSIES project hopes to eventually go back and sequence DNA from women who also self-identify with other racial groups such as Hispanic and Asian.

In addition, they'd like to eventually increase the size of the database but can't do so without additional funding. As it is, King noted, UW conducted much of the work piggybacking on other grants; while Color provided its sequencing work pro bono.

"At one level one would like to have a sequence for every FLOSSIE on earth," King said. "We'd like to have it three times the size that it is so that we have a good sense of what really rare variants look like. At one level it will never be enough … but for the size that it is now, certainly for variants that have a frequency of a tenth of one percent, one can get a good sense. It's a start."