Skip to content

When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.


How the Media Oversimplifies DNA Testing of Separated Families

Signing the Genetic Information Nondiscrimination Act (NHGRI)

When the Genetic Information Nondiscrimination Act of 2008 (GINA) was passed to unify “the existing patchwork of State and Federal laws,” the language was broad enough to apply to just about any use of information gleaned from  DNA. The law is meant to prevent discrimination in health insurance and employment based on results of a “genetic test,” defined as “an analysis of human DNA, RNA, chromosomes, proteins, or metabolites, that detects genotypes, mutations, or chromosomal changes.”

Use of “genotypes” alone covers any application of determining sequences of A, T, C, and G that I can think of, and if that’s not enough, the first sentence of the act mentions sequencing the human genome.

A decade ago also came the first direct-to-consumer (DTC) DNA tests, from a handful of companies. Since then the number and types of “spit” and cheek swab tests and companies offering them have mushroomed, probing traits, tendencies, risks of future illness, metabolic quirks, behavioral characteristics, carrier status for single-gene diseases, and of course ancestry. But with the spreading tentacles of DNA testing, a lack of precision in describing what, exactly, is being considered, can lead to misunderstanding. That’s apparent in the use of DNA testing to help to reunite children separated from their parents as part of the Trump administration’s “zero tolerance” policy at U.S. borders.

Don’t Echo Politicians Who Gloss Over the Science

Initial confusion about DNA testing unfurled June 21 when California Representative Jackie Speier called for DTC DNA testing company 23andMe to donate kits for swabbing children’s cheeks at the borders.

But that wasn’t quite right. 23andMe’s DNA tests, for such traits as “genetic weight,” “unibrow,” whether Brussels sprouts taste bitter, and Viking ancestry, aren’t the 4-base-long “short tandem repeat” (STR) forensic markers that the U.S. Citizenship and Immigration Services used to reunite immigrant children to “anchor” parents living in the U.S. who are citizens, a program that stopped taking new cases last year. Those labs are accredited by the American Association of Blood Banks, and as far as I know the government hasn’t released the identity of the testing lab or labs.

Rep. Speier likely hadn’t at first perused 23andMe’s website, which clearly cites use of a chip from llumina that detects single sites in the genome that differ among individuals (single nucleotide polymorphisms, aka SNPs) – not the STRs of forensics. Buzzfeed soon updated that the company would donate kits, but still not indicating what would be tested. Maybe just the collection.

But when a politician speaks, the media echoes, even when the science is glommed into a sound bite. Soon the “DNA testing” mantra began to reverberate, with only a few reports explaining the nuances of the different types of DNA tests – which has consequences for privacy concerns.

An Epidemic of Vagueness

“Just as commercial spit kits have brought together long-lost cousins or unknown siblings, the thinking goes, they might also be able to help in the current crisis,” according to the portion of a STAT News report that I could read for free. That’s accurate if they just mean the sampled cells, and not the DNA probes, because 23andMe’s standard offerings wouldn’t be of much help; lots of people hate Brussels sprouts and have unibrows. What would help are the 20 STR markers that form the FBI’s Combined DNA Index System (CODIS) of criminal profiles and are used in immigration cases to identify people who share half their genomes – kids, siblings, and full parents.

In 2017 the FBI upped the number of CODIS STRs to 20.

The power of STR testing lies in the math. Each of us has specific numbers of the STRs at 20 sites in the genome, times two because we have two of each chromosome. Overall, our STR profiles can vary more ways than there are people on Earth – hence the specificity. At the same time, STRs alone reveal nothing about the individual – no disease risks, no vegetable preferences – because repeats don’t encode proteins, which provide traits.

The Associated Press July 11 issued “Test DNA of migrant kids only as last resort, ethicists say.” The article evokes “DNA testing” ten times, but doesn’t mention what, exactly, is being tested. A genome? An exome? Specific genes? Weird chromosomes? STRs? Thousands of SNP markers? The report quotes two bioethicists, an HHS official, a U.S. district judge, and two lawyers, who pontificate over the horror of it all.

Although The Atlantic’s “Which Company Will Test the DNA of Separated Families?” cites the precedent in immigration and explains forensic STR testing, it also includes a quote making the rounds from Jennifer Falcon, of the immigrants-rights group RAICES, about the DNA testing:

“It’s a gross violation of human rights. These are minors with no legal guardian to be able to advise on their legal right, not to mention they’re so young how can they consent to their personal information being used in this way?”

Huh? 40 STRs is personal information? It would take a lot of analysis to collate the DNA regions surrounding the STRs to SNP maps to derive anything about a phenotype – trait or illness – from the cheek swabs at the borders.

Happily, USA Today, despite the headline “DNA tests used to reunite migrant families spark worries they’ll be used for much more,” quotes several prominent geneticists, and distinguishes and explains STR testing.

Navigating the Web of DNA Test Types

Other DNA testing scenarios can be confusing too, judging from the emails I get asking me to interpret results.

Paternity tests are easy, just the simple STR patterns – dad and child should share at least half of the 40 markers. This is perhaps why on soap operas it is routine for a nefarious character to sneak into a lab and alter results.


Women on my breast cancer Facebook group who know only of BRCA1 and BRCA2 are confused when results indicate mutation in CHEK2 or TP53 or any of dozens of other cancer genes – which indicate risk, not certainty. Some women, and even some health care providers, don’t realize that genetic testing is important before deciding on treatment. An inherited elevated risk suggests a double mastectomy makes more sense than a single mastectomy or lumpectomy.

Genetic screening before trying to conceive may show that both partners carry recessive mutations, but that’s only worrisome if they’re in the same gene (see “Mendel meets Tinder: GenePeeks screen maps babies’ genetic risks — before conception”). Ditto for prenatal screening. People who don’t remember how recessive inheritance works may become upset over learning they’re carriers of something. We all are.

DNA ancestry testing can be distressing in other ways, shattering a sense of identity or revealing biological relationships that might have better remained mysteries. Just a few months ago, when my online “genethics” course was discussing genetic genealogy, a student related how her elderly mother had become distraught when ancestry testing didn’t identify the small Native American group from which she’d been told she descended – but the testing company probably just didn’t have that info in their database. It’s easy to interpret lack of evidence as a true negative finding.

I find the crop of relatively recent “lifestyle DNA” DTC testing companies maddening. Some offered tests merely detect patterns of SNPs, each of which contributes a vanishingly small degree to a trait. I wouldn’t make a life-altering decision on finding a SNP association, not even to eat more Brussels sprouts. Reading a palm or tea leaves would be about as useful.

Some “lifestyle” companies test DNA tangentially linked to something like metabolism or muscle function and package the results as a pitch for costly and supposedly personalized vitamin supplements. This is an old story. See “GAO Concludes DTC Genetic Tests Mislead Consumers,” the classic take-down of “nutrigenetics” companies by investigators from the Government Accountability Office. Twelve years later, things haven’t changed. Companies exploit lack of consumer knowledge about genetics to sell pricy products.

12 Types of Genetic Tests, in Context

Here’s a rundown of the types of DNA tests, from largest to smallest.

  • Comparing genome sequences among living species illuminates shared ancestry and evolution.
  • Clinical whole genome sequencing can diagnose or initially describe a very rare disease that has eluded all other methods.
  • Exome sequencing of sick children and their parents – trios – can distinguish whether the child has a new and dominant mutation in a protein-encoding gene or has inherited two gene variants that cause the condition, one from each parent.
  • Karyotype charts display missing, extra, inverted, or moved chromosomes.
  • Chromosomal microarray tests reveal extra and missing chromosome regions.
  • Gene panels test for mutations in functionally related genes, such as for neurological and cardiac diseases.
  • Mutations in single genes may cause a disease or trait or raise the risk of developing a condition, like breast cancer.
  • Mutations in DNA sequences that do not encode protein may control the expression of protein-encoding genes or be repeats.
  • Genome-wide association studies amass single nucleotide polymorphism (SNP) patterns that may be linked to specific traits. Here’s a recent SNP study that made headlines about genetic variants associated with educational attainment. The 1271 catalogued SNPs explain about 10% of the variation in educational attainment – that’s not the same as explaining 10% of the trait itself, a common misconception in media coverage.
  • Y chromosome sequences and mitochondrial DNA sequences trace paternal and maternal ancestry, respectively, but only represent a tiny percentage of a person’s ancestors. Ancestry testing tells more when it includes autosomal (non-sex-chromosome) DNA sequences.
  • Repeats such as STRs are used to distinguish individuals, match a person to crime scene biological evidence, identify remains, and reunite families.
  • Gene expression profiling indicates how actively sets of genes are being transcribed into RNA, the first step in synthesizing proteins. Often news releases pitch stories about patterns of mutations that are really patterns of gene expression – not at all the same thing. That’s what happened with reporting earlier this year comparing gene expression patterns between an astronaut in space and his earthbound identical twin – of course they differed, everything we do changes gene expression!

Until media coverage of the consequences of DNA testing includes more quotes from geneticists, the folks who interpret DNA testing in the news – the activists, lawyers, bioethicists, journalists, and politicians – need to consider, and communicate, exactly what a specific DNA test actually tests.




  1. I suppose you don’t advocate using a household plumber to design the plumbing for a nuclear power plant cooling tower, then? ?. Great analysis, as usual, Ricki

  2. The mother who was upset because it showed she had no Native American autosomall dna, probably had been raised on a NA blood myth all her life, like the majority of White Americans!! Why would they have to test the NA group this mother supposedly descends from!! The Native American people descend from no moree than 250 ancestors who were the founding population of the Americas that all NA groups descend from, and they are more related to each other than any other group in the world. So if she had any NA dna it would definitely show on a DNA test!!

  3. NA DNA would only show on a DNA test if all NA DNA sequences are being probed. That’s not likely, because ancestry testing does not compare complete genomes. I do not think that all possible autosomal DNA sequences that have variations unique to NAs are tested. We might not even know all of them because some native populations have not allowed researchers to examine their DNA. The parts of the genome that ancestry companies test for include the male Y lineage, the mt female lineage, and selected sites among the autosomes. That leaves out so much that it is indeed plausible that someone could test non-NA and actually have some NA sequences. In addition, the independent assortment of chromosomes at meiosis, plus recombination (crossing over where chromosomes exchange parts), could, by chance, mix up gene combos in a way that minimizes or even removes the NA contribution, perhaps to the point that none of it matches a specific company’s probes. We are 99.something % alike. We should focus on that. No need for !!!. I’m just explaining the science. Yes, of course the woman could have been mislead. But the tests might not have been complete enough to give a definitive answer. Thanks for responding.

Leave a Reply

Your email address will not be published. Required fields are marked *

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Back to top