Skip to main content

Incidental Findings from Genome Sequencing Nuances and Caveats

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


You have your genome or exome (the protein-encoding part) sequenced to help diagnose a puzzling set of symptoms, and something totally unrelated, and unexpected, turns up – a so-called “incidental finding.”

Surprises, of course, aren’t new in medicine. The term “incidental finding” comes from “incidentaloma,” coined in 1995 to describe an adrenal tumor found on a scan looking for something else. I had one -- a CT scan of my appendix revealed a polycystic liver. A friend had it much worse. She volunteered to be a control in an Alzheimer’s imaging trial, and her scan revealed two brain aneurysms!

Geneticists have long expected an avalanche of incidental findings from clinical (exome or genome) sequencing. Researchers from Baylor College of Medicine and NHGRI and elsewhere described several cases at the American Society of Human Genetics annual meeting last fall. My favorites:


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


- A boy had his genome sequenced as part of a project to better diagnose syndromes of developmental delay, intellectual disability, and seizures. Researchers found the aorta weakening of Marfan syndrome, gave the boy a repurposed drug in clinical trials, and he’s ok.

- A family with several members having their genomes sequenced to evaluate heart disease discovered that their “writer’s cramp” is myoclonus dystonia, a neuromuscular disease.

- A man had his genome sequenced in a study to investigate atherosclerosis and learned he had a deafness mutation. Although he claimed he had normal hearing, further testing showed he didn’t – he’d adapted so well for so long that he hadn’t known he was missing a sense.

To provide guidelines for clinicians having to disclose a medical surprise, the American College of Medical Genetics and Genomics released much-anticipated recommendations on March 21, to kick off the annual meeting.

The most important points come near the end of the 27-page document, especially the table of conditions to be tested for. And while the report is very clear, the accompanying news release uses some fuzzy definitions that often crop up when genetics is simplified. I know from genetic counseling and writing textbooks that misuse of certain terms can confuse. (“Genetic code” for “DNA sequence,” and “carrier” in for “pre-symptomatic,” when it more traditionally refers to someone who has one recessive allele, and no corresponding illness or trait.) News aggregators that boil down news releases may miss such nuances.

So here are 12 major points I’ve distilled from the report.

1. Labs doing clinical sequencing should test for well-studied mutations in 57 genes that cause or can theoretically cause disease. The mutations are fairly common, albeit among the rare – I’ve seen 3 in patients just this week. Most are “actionable,” some even life-saving. Certain heart conditions and malignant hyperthermia, for example, can cause a first (and last) symptom of sudden death when a person takes a certain drug. A good thing to know.

2. The list of 57 doesn’t really mean 57 illnesses, from a patient’s point of view. Tumor/cancer syndromes account for 25 of the 57, cardiovascular problems 23, and mutations in seven genes cause Marfan and related syndromes. Add the two lone conditions (a type of Ehlers-Danlos syndrome and malignant hyperthermia) and the list collapses to 5 to a patient.

3. Of people having clinical sequencing, 1-2% are expected to have one of the 57 mutations. Technically, these aren’t incidental findings, because they aren’t found by accident, like my cystic liver – labs are looking for them. But that’s ok. Just semantics.

4. We don’t know if the 57 variants that may cause disease in families who are having clinical sequencing to evaluate symptoms may do so in others. Due to effects of other genes and the environment, a mutation that causes a disease in one person may not in another.

5. Clinical sequencing is not the sort of testing that some direct-to-consumer (DTC) companies offer. At the present time, for example, 23andMe’s $99 test for 247 illnesses and traits include carrier tests that would already be part of a diagnostic work-up based on symptoms, family history, or newborn screening; and risks based on genetic marker (SNP) patterns that may or may not predict anything about the person sending in spit, due to population differences between the spitter and the study on which the test is based. DTC tests in their current incarnation are informational, not diagnostic.

6. Patients and their families won’t be able to “opt-out” of knowing about the big 57, unless they refuse clinical sequencing. “Duty to warn” trumps patient autonomy.

7. Kids count, contrary to precedent. Huntington disease set the pattern here – people under age 18 with an affected parent are generally not tested to see if they will develop HD 20 years later, because there’s no treatment. But for the clinical sequencing guidelines, informing parents and even children is okay, because so many of the conditions can be prevented, treated, or risk reduced with lifestyle choices, even if an illness such as cancer won’t start for many years.

8. A negative result doesn’t assure health. I’m thinking of Bruce Springsteen’s “57 Channels (And Nothin’ On).” Finding nothing on the 57 tests doesn’t mean you’re healthy or will stay that way. A person who has normal BRCA genes, for example, can still develop breast cancer from other genes going haywire. And sequencing won’t spot missing, extra, or moved DNA; unusual mutations; or those lurking in genome regions with light sequencing coverage. (See “The Battle of the Prenatal Tests”)

9. Genetic counseling. Clinicians should provide it or refer for it – before and after testing. To help, ACMG will soon release informed consent guidelines for clinical sequencing. But physicians-in-training still get woefully little training in genetics and genomics.

10. Genetic testing and sequencing are on a collision course, with mutation databases from patients edging toward the rapidly-growing databases from healthy people sending samples to DTC companies. (23andMe expects a million submissions by year’s end.) But this convergence will be to everyone’s benefit, I think. With all of these data, in a few years we’ll know what the variations in the human genome, point by point, actually mean.

11. Price is driving clinical sequencing. It’s already cheaper to do a whole exome than to sequence a “big chunky” gene and all of its known mutations.

If Amazon offered you a dozen books for the price of one, wouldn’t you take it?

12. Finally, at the end of the recommendations come some interesting terms that capture the disconnect I’ve long noticed between those of us who are more hesitant about sequencing and its enthusiasts: the “genetic empiricist” versus the “genetic libertarian”.

I’ve been an empiricist, against genetic testing when the info may be ambiguous because we don’t yet know enough. A report in this week’s Proceedings of the National Academy of Sciences from Yuval Itan and colleagues at Rockefeller University takes a major step in providing this needed context. They introduce the human gene connectome, a computational way to sort out gene-gene interactions that will be vital for making sense of genome information. Does a mutation in one gene protect against another? Imagine finding out you have a 3-fold increased risk of Alzheimer’s, before researchers discover a protective variant that you also have?

I still fear that too casual an approach to testing – like urging people to give spit kits for Christmas -- could create a population of Woody Allen-like hypochondriacs who overtax the health care system and take too many not-without-risk tests. Others will be falsely reassured.

We genetic empiricists want evidence that a testing outcome is beneficial, that it “does no harm.” That’s still not known for all genetic tests.

In contrast, “genetic libertarians” believe that everyone has the right to know everything. From responses to my blog posts, testimonials at the 23andMe website, and the streams of popular articles and books by people having their exomes/genomes “done,” the genetic libertarians seem to far outnumber the genetic empiricists. Or at least they’re more vocal.

The new guidelines on clinical DNA sequencing present a starting point for handling what our genomes are telling us, focusing first on revelations that we can do something with, under the care of trained medical professionals. So by the time that devices the size of a dorm fridge are sequencing patients’ DNA and spitting out risks, diagnoses, and suggested courses of action in the average internist’s office, we’ll be able to make the most of the information in our genomes.

Ricki Lewis received her PhD in genetics from Indiana University. Her ninth book, The Forever Fix: Gene Therapy and the Boy Who Saved It, narrative nonfiction, was just published by St. Martin's Press. Most of her other books are college life science textbooks, including "Human Genetics: Concepts and Applications," (10th edition, 2012) from McGraw-Hill Higher Education. Routledge Press published "Human Genetics: The Basics" in 2010. Ricki has published thousands of magazine articles, from Discover to Playgirl, but mostly in The Scientist. She is a genetic counselor at CareNet Medical Group in Schenectady, NY and teaches "Genethics" online for the Alden March Bioethics Institute of Albany Medical College. Ricki is a hospice volunteer and a frequent public speaker (Macmillan Speaker's Bureau). Ricki's blog Genetic Linkage is at www.rickilewis.com and she tweets at @rickilewis.

More by Ricki Lewis