On the science
What is a genetic risk score?
Most diseases have a genetic component. Genetic risk scores quantify the degree of risk conferred through an individual’s genes.
The human genome contains several million genetic markers that commonly vary across individuals and influence the risk of common diseases. Individually, each of these associated single nucleotide polymorphisms (SNPs) has a small impact on disease risk. Their combined effect, calculated as a genetic risk score (GRS), is frequently the best genetic estimate of disease risk, providing more predictive value than any single gene marker alone [1, 2, 3]. Genetic risk scores often outperform traditional clinical risk factors like smoking, BMI, and family history [3, 4]. This is a new development, over the last 5 years genome sequencing has become more affordable, allowing study sizes to grow to the order of hundreds of thousands of individuals. These larger study sizes provided large amounts of data and afforded the development of more accurate genetic risk scores [5, 6].
Learn more about genetic risk scores through a visual explainer.
How is this different from the genetic tests I usually order?
Orchid tests for genetic susceptibility to the most common conditions based on many genes instead of rare genetic disorders caused by single genes.
Historically, clinical genetics has focused on identifying rare genetic variants that diagnose a rare genetic disease. Monogenic diseases such as cystic fibrosis or spinal muscular atrophy can be diagnosed by interrogating genetic variation within single genes. These conditions may have autosomal dominant, autosomal recessive, or X-linked inheritance patterns. These are the tests that are usually ordered by physicians for individuals at risk of an inherited mutation, or as part of preconception or prenatal counseling.
However, the vast majority of diseases are not controlled by the presence or absence of a single gene. Most common disease susceptibility is governed by small changes in dozens to millions of locations across the genome [3].
A genetic risk score aggregates the effects of tens to thousands of genetic variants that mediate their effects through many genes: this is what is meant by a “polygenic” disease architecture [7].
Learn more with a deeper dive into our science.
Why are these results reported as a “risk” number?
An elevated genetic risk score identifies individuals with a genetic susceptibility to a disease of interest, however it is not a diagnosis of a disease.
While anyone can develop any of these conditions, genetics can tell us whether an individual has a combination of certain genetic markers that increase or decrease susceptibility to these conditions. Genetic risks are similar to other risk factors — a smoking history makes an individual more likely to develop lung cancer, for example.
Knowing this information can help individuals proactively manage their health and if desired, seek mitigation strategies to lower genetic risk for their future children, thus decreasing risk of disease development [8].
How is a genetic risk score developed? How is it validated?
Statistical geneticists develop genetic risk scores using cohorts of hundreds of thousands of individuals who have sequence data paired with physician verified diagnoses.
Statistical tests are performed to find variants that are found in cases (individuals with the disease of interest) at higher rates than in controls (individuals without the disease of interest).
This study design is called a genome-wide association study (GWAS) [9]. There is no single way to develop a polygenic risk score model, but the process generally involves:
Reducing the impact of potential confounding variables. In this case, non-genetic factors like age, socioeconomic status, or geography
Identifying a set of genetic variants with a strong association to the disease outcome
Quantifying the extent to which each genetic variant either increases or decreases disease risk
Once a genetic risk score has been developed, it must be tested on a separate population that has also been sequenced and has physician verified diagnostic records. This is known as a “validation dataset” and helps answer the question: “is this genetic risk score as predictive on this population as it was on the original cohort used to develop the model? [10]”
How predictive is a genetic risk score? What’s the evidence?
Using data from hundreds of thousands of individuals, we see that individuals with an elevated genetic risk for a particular condition are 2 to 5 times more likely to develop that given condition than the average population.
Genetic risk scores aggregate the effects of dozens to millions of variants across the genome to estimate overall risk more accurately than monogenic (single gene) status alone [1, 2, 3]. Genetic risk scores outperform conventional clinical risk factors [3, 4].
The evidence supporting genetic risk scores is based on curated databases of hundreds of thousands of individual’s genetic sequencing data paired with physician verified diagnoses. Large-scale studies have reproducibly found that genetic risk score percentile stratifies and predicts the fraction of individuals that develop a condition over time.
To illustrate, the figure below maps the percentile of a genetic risk score for coronary artery disease to the prevalence of coronary artery disease. The light blue dots represent actual patient cases, while the dark blue dots represent the predicted prevalence using the genetic risk score [3].
Khera A.V., et al. (2018); Supplementary Figure 2
We continually review the latest research in genomics — identifying results that meet our stringent quality standards. We only incorporate results reproduced by several research groups and based on population sizes of more than 10,000 individuals.
Learn more with a deeper dive into our science.
.svg){kind=small}
