A massive study of nearly 4,000 variants in a gene associated with cancer could help to pinpoint people at risk for breast or ovarian tumours.

The information is sorely needed: millions of people have had their BRCA1 gene sequenced. Some variations in the DNA sequence of BRCA1 are linked to breast and ovarian cancer; others are thought to be safe. But the effects of most variants are unknown , leaving patients and physicians alike at a loss to interpret the results.

The study, published on 12 September in Nature ¹ , examined the effects of thousands of such variants on the survival of cells grown in the laboratory. The findings could help physicians to interpret the mutations’ significance. For example, a variant that hampers a cell’s ability to repair DNA in the lab might also be linked to cancer in the clinic.

“Every patient is different. Every physician is different,” says Jay Shendure, a geneticist at the Brotman Baty Institute for Precision Medicine in Seattle, Washington, and a co-author of the study. “But if such a variant were present in a family member of mine, would I use this information? Absolutely. Otherwise, there’s no information.”

Uncertain future

The American College of Medical Genetics and Genomics recognizes about 60 genes for which screening might suggest a medical plan to preventor reduce the effects of a disease. Yet often when people find out that their genes include unusual DNA sequences , they are at a loss to interpret that finding.

“These variants are nightmarish,” says Alvaro Monteiro, a geneticist at the Moffitt Cancer Center in Tampa, Florida. “The result becomes: ‘Well, you have something, but we just don’t know exactly what it is.’”

Assays for genetic causes of hearing loss are a prime example: about half of the people who undergo such testing find out that they carry variants whose significance is unknown, says Heidi Rehm, a geneticist at the Broad Institute of MIT and Harvard in Cambridge, Massachusetts. “It’s a definite challenge in all genetic testing areas,” she says.

For BRCA1 , the stakes are particularly high: women whose BRCA1 sequence puts them at risk for cancer will sometimes have surgery to remove their breasts and ovaries. Rehm says that there are more than 2,500 known variants of uncertain significance in BRCA1 .

Casting a wide net

Shendure and fellow Brotman Baty Institute geneticist Lea Starita decided to tackle this problem using cells that die without a functioning BRCA1 protein, which is important for DNA repair. They used CRISPR–Cas9 gene editing to create mutations throughout the BRCA1 gene, and then looked to see which of the resulting cells survived.

The screen covered nearly every possible single-letter variation in regions of BRCA1 known to be important for the function of its associated protein. In cases where clinical data were available, lab results correlated with clinical findings more than 96% of the time.

The approach could be expanded to some other disease-related genes, particularly cancer-associated genes that are important for DNA repair, says Monteiro.

Ideally, these data would be combined with other genetic information about a given variant, but more data might not be available for rare sequences. In such cases, it will be up to the patient and their physician to decide how to interpret the results, says Starita. Some might choose to conduct further screening for cancer, in an effort to catch tumours early, she notes.

And eventually, researchers might incorporate lab results, such as those derived from the CRISPR screen, directly into the models they use to classify variants , Monteiro says. But he also notes that the field tends to be conservative, and will likely require more evidence before it makes the change.

“This is a very thorny subject,” he says. “When we try once to classify a variant, we really like that decision to never change again.”