Assigning the prevalence and cancer-causing potential of mutations in
the best understood genes is beset with unknown unknowns, since only
about 30% of the
BRCA1 variants in reported to the ClinVar
database and less than 5% of all possible single amino acid
substitutions have ever been seen in the largest collection of over
120,000 whole genome sequences
\cite{Karczewski_2019} assembled in the
gnomAD
database (2,129
BRCA1 variants of the BRCA Exchange variants are also seen in gnomAD). Gene
editing allows VUS to be assigned a pathogenicity score by testing
all possible
substitution mutations for cell survival or cell function before they
are discovered in individual people
\cite{Findlay_2018}.
BRCA1 and four other tumor
suppressor genes are essential in haploid cell lines. In those, most of
3893
BRCA1 mutations were just like known benign variants that
allowed cells. In contrast, 21 known cancer-causing mutations that had
no function were lethal, allowing 400 non-functional missense mutations
and 300 impairing gene expression to be labeled pathogenic for the first
time. Assays testing
homology-directed
repair of double strand DNA breaks
\cite{Starita_2018} can be used to determine the
activity of 6% or so mutations that showed partial loss of function in
the viability assay.
The smaller contributions of many genes throughout the genome are
thought to explain population
cancer risk, together with environmental factors such as hormone
exposure, smoking and aging \cite{Cronin_2018}. At the population level, about 18% of the
familial relative risk of breast cancer can currently be explained by
mostly common variants identified by
genome wide association
studies \cite{Michailidou_2017}. Using these variants for polygenic prediction, the 1% of
women with the highest predicted genetic risk would have a 3.5-fold
(42% lifetime risk) compared to the population average (12% lifetime
risk). Such tests are undergoing clinical investigation for their
utility in prioritizing early detection and prevention strategies,
particularly for those with sparse or unrecorded family history of
cancer.