Defining accuracy through two lenses: predictability and actionability

The first thing to appreciate about the word “accuracy” is that it can describe many different aspects of a genetic test. What accuracy really boils down to is two fundamental questions that matter most to you:

1. How predictive is the test? — Is what’s measured actually associated with what you’re trying to find out? (disease risk, genetic diagnosis, physical trait, etc.)
2. How actionable are the results of the test? — Once you find out about your results, are there things you can do about the information learned?

These two axes are on a continuum, and you can think of the four quadrants as ways to characterize a genetic test:

So, if you were hoping to gain health information from a genetic test, ideally the test would fall in the upper right quadrant — a test that has been shown to be 1) highly predictive of the characteristic of interest and 2) actionable by way of making medical or personal decisions based on the results.

Evaluating how predictive a test is

How can you tell if a test is predictive for what you’d like to find out?

Accuracy in obtaining genetic data

The first aspect is analytical validity. This refers to how accurately and reliably the DNA is being read. There are different ways to read your DNA, but put simply: if the test is designed to detect specific genetic variants — can the test accurately detect whether those genetic variants are present or absent? In other words, if the result says you have a specific genetic variant, do you actually have it?

If the test is designed to look at the combined effect of many genetic variants to assess for genetic risk of certain conditions, like what is done here at Orchid — are all those data points being correctly identified?

The two most common technologies used in genetic testing is genotyping and sequencing. SNP genotyping checks predetermined hotspots (less than 2% of the genome) while sequencing actually reads each letter base. Sequencing collects more information than genotyping, though tends to cost more.

So if sequence data is over 99% accurate across the entire genome, that means when you process all three billion bases of the human genome, the data accurately reflects what’s actually present in the genome over 99% of the time.

One way to ensure analytical validity is to see if testing is performed in a CLIA-certified laboratory. CLIA standards oversee how tests are performed and the quality of laboratory practices.

How well genetic data relates to disease

The second equally important criteria is clinical validity. This refers to how well a specific genetic variant or calculated genetic risk relates to a particular health condition.

Establishing the clinical validity of single genetic variants causing single gene disorders like cystic fibrosis is determined by a process known as “variant interpretation.” Pieces of evidence currently available such as clinical observations and experimental studies are considered to “classify” each genetic variant into buckets — ranging from benign (variation that does not impact disease) to pathogenic (variation that very likely impacts disease).

Meanwhile, measuring genetic risk scores involves assessing a combination of genetic variants across the genome. This is done by comparing the genetic makeup of hundreds of thousands of individuals diagnosed with a disease (“cases”) to individuals without the disease (“controls”). In this case, the predictive ability can be defined as how well a genetic score predicts the risk of developing a specific health condition.

Two metrics that have been commonly used by scientists to assess the strength of a genetic risk score include “SNP heritability” (what percentage of a disease outcome can be explained by the genetic risk score) and “AUC” (probability that a randomly selected individual with a disease will have a higher genetic risk than an individual without the disease).

So, how predictive is a genetic risk score?

Large-scale genetic studies have shown that people with the highest percentiles of genetic risk for a condition are 2 to 5 times more likely to develop it compared to average. Furthermore, quantifying genetic risk accurately identifies people who are more likely to develop a disease — critical information that can be easily missed by looking at family history, routine blood tests, or traditional genetic testing alone. These findings have been consistently demonstrated and validated through aggregated data from hundreds of thousands of individuals.

Of note, while genetics plays a role in susceptibility for many conditions, it is not the only factor, especially when it comes to the probabilistic nature of quantifying genetic risk.

At Orchid, 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.

Determine what actions you can take based on the results

The right genetic test for you depends on what you want to find out. Sometimes, the results of a test can help guide clinical or personal choices to improve health outcomes and help prevent disease.

If you’re planning on starting or growing your family, knowing you and your partner’s genetic risks helps you optimize your family’s health. You can choose to better prepare for the future through personalized preventative measures and increased awareness. Healthy lifestyle changes can offset genetic risks to chronic conditions and being able to measure these risks puts you one step ahead of the game. You can also mitigate genetic risk for your future child through follow-up IVF and embryo screening.

Takeaways

• When evaluating the accuracy of a genetic test, you should ask: how predictive is the test? How actionable are the results of the test?
• Predictiveness is determined by the accuracy of reading your DNA and how well what’s analyzed relates to what you want to find out 
• The right genetic test for you depends on what you want to find out — Are you and your partner planning on growing your family and want to minimize disease risks for your children? Does your family have a history of certain genetic disorders that your doctor wants to investigate? Do you want to explore your family’s ancestral roots?

Orchid offers advanced genetic testing for couples planning on building their family. We use advanced tools and smart, caring humans to help you give your future children the best shot of a healthy life. Conceive with greater confidence and peace of mind