Questions to Ask About DNA
Questions for students, patients and curious readers talking with a biology teacher, a genetic counselor or a researcher. They cover how DNA works, what consumer and clinical tests can and cannot show, how to read a risk figure, and what happens to your sample and your data afterwards.
The questions
Open any question for the note
If you had to explain what DNA actually is in a minute, how would you put it?
Why ask it
A good opener because it shows you how the person thinks before you get technical. Watch whether they reach for the blueprint metaphor, which most geneticists dislike, or something closer to a recipe or a parts list that gets read differently in different cells.
How much of my DNA is the same as a stranger's, and how much varies?
Why ask it
The answer reframes everything that follows: humans are overwhelmingly identical at the sequence level, so the interesting work happens in a small fraction of positions. Anyone who cannot put rough numbers on that is working from headlines.
What does a gene actually do, and how does a cell decide which genes to use?
Why ask it
Regulation is the part school biology usually skips. A strong answer gets to the fact that every cell carries the same sequence and differs in what it switches on, which is why a liver cell and a neuron are not the same thing.
Does most of our DNA code for proteins, or is it doing something else?
Why ask it
Only a small share of the genome is protein coding, and the label junk DNA has aged badly. Listen for whether they distinguish regions with regulatory function from stretches with no known function, because collapsing those two is a common error.
Where do new mutations come from, and do I carry any my parents did not have?
Why ask it
Everyone carries a handful of changes absent in both parents, which surprises people who assume inheritance is pure copying. The answer should cover both copying errors and outside causes such as radiation and certain chemicals.
When someone says a trait is heritable, what does that number mean?
Why ask it
Heritability describes variation within a particular population in a particular environment, not how much of your own height or intelligence came from genes. If the person you are asking uses it to mean the latter, treat the rest of their explanation carefully.
Which traits come down to a single gene, and which involve hundreds?
Why ask it
Single gene conditions are the ones that behave predictably and that testing handles well. Most things people are curious about, height, temperament, common disease risk, are spread thinly across many locations, which is exactly why prediction is weak.
What can a consumer ancestry test really tell me, and what is it inferring?
Why ask it
These results are comparisons against reference panels of living people, not readings of your past. The size and composition of the panel is the answer that matters, and it explains why some regions come back with confident detail and others are lumped together.
Why can two siblings get different ancestry breakdowns from the same company?
Why ask it
Siblings inherit different halves of each parent's DNA, so real differences exist, but part of the gap is estimation noise. This question tends to expose whether the person understands the reported percentages as estimates with error bars or as measurements.
How is a health report from a consumer test different from a clinical genetic test?
Why ask it
Consumer arrays check selected positions, while clinical tests often sequence entire genes and are interpreted against a patient's history. The practical consequence is that a reassuring consumer report can miss the variant that matters in your family.
If a result comes back as a variant of uncertain significance, what happens next?
Why ask it
These are common and they are not bad news or good news, they are unfinished business. Ask who is responsible for revisiting the classification later, because that is where uncertain results either get resolved or quietly forgotten.
How do you tell whether a variant causes a disease or just sits near one?
Why ask it
This is the daily problem in genetics, and the answer usually involves family studies, population frequency and whether the change plausibly breaks the protein. A confident causal claim from a single association study is a warning sign.
What can genetics say about how I will respond to a particular medication?
Why ask it
For a small set of drugs, including some blood thinners and psychiatric medicines, genotype changes dosing in real practice. For most drugs it does not yet, so the useful part of the answer is which specific medicines are affected.
When is it worth seeing a genetic counselor rather than reading the report myself?
Why ask it
Counselors mainly help with family history interpretation and with decisions that follow a result, not with explaining the science. If the person cannot say what a counselor would add in your particular situation, you probably do not need one yet.
If I test positive for something inherited, what does that mean for my siblings and children?
Why ask it
A result about you is partly a result about your relatives, which is the feature of genetic information that distinguishes it from other medical data. Ask who is expected to tell them, since clinics differ and most leave that to the patient.
What happens to my sample and my data after the company has processed it?
Why ask it
Retention, research consent and resale after a bankruptcy or acquisition are the parts people find out about later. Ask specifically whether deleting your account destroys the physical sample, because those are usually separate requests.
Can insurers or employers use genetic results where I live?
Why ask it
Protections vary a great deal by country and by type of insurance, and life, disability and long term care coverage are often treated differently from health coverage. The answer should be jurisdiction specific or it is not an answer.
How do police genetic genealogy searches work, and does a relative's test expose me?
Why ask it
Investigators identify people through partial matches to cousins who uploaded data, meaning your exposure depends on decisions made by relatives you may never have met. Ask which databases allow it and whether uploading is opt in.
What claims about DNA do you see repeated that are simply wrong?
Why ask it
Experts usually have a ready list, and it tells you where the popular coverage of their field goes off the rails. Common candidates are a gene for a behaviour, DNA as destiny, and treating small ancestry percentages as identity.
What do we still not know about how the genome works?
Why ask it
This closes on honesty rather than facts. Someone comfortable in the field will name open problems freely, and the specific gaps they choose tell you what is actually being argued about right now rather than what has settled.
How to use these questions
Practical guidance for the conversation itself
Getting past the textbook answer
Separate the science from your own results
Two different conversations hide inside this topic. How inheritance works is a teaching question anyone with a biology background can answer. What your own test result means is a clinical question that needs your family history in front of the person. Say which one you are asking, or you will get the wrong kind of answer.
Ask for the base rate alongside any risk
A doubled risk is meaningless without the starting figure, since doubling a one in ten thousand chance is not the same as doubling a one in ten chance. Whenever you hear a relative risk, ask what the absolute numbers are for someone like you.
Ask what the test does not look at
Most disappointment with genetic testing comes from assuming full coverage. Ask which genes were examined, whether the method detects large deletions and duplications, and what a negative result therefore rules out. This applies to consumer and clinical tests alike.
Write down the exact variant notation
If you are discussing a specific result, get the gene name and the variant written out rather than a summary phrase. Classifications change over time, and that notation is what lets you or a clinician look up the current interpretation years later.
Before you send off a saliva kit
- Decide what you would do differently depending on the result. If nothing changes, you are buying curiosity, which is a fine reason, but it sets expectations correctly.
- Consider who else the result implicates. Ancestry tests routinely reveal misattributed parentage and unknown half siblings, and there is no way to unlearn that once shown.
- Read what happens to your data on account deletion, and whether research consent is separate and revocable.
- Check whether the service lets law enforcement search matches, and whether that setting is on by default.
- If you have a family history of a specific condition, ask a clinician before buying anything, since a targeted clinical test may be covered and will be more informative.
- Expect to be able to download raw data, and expect third party interpretation sites to make confident claims from it that the original company would not make.
Common misreadings
Treating percentages as fixed facts
Ancestry estimates are recalculated when reference panels change, so the figures on your report can shift without any new information about you. People often remember the first version as the true one.
Reading a negative result as an all clear
A test that finds nothing means nothing was found among the positions examined. With a strong family history, that is a reason to ask what else could be tested rather than to stop looking.
Assuming genes settle the question
For most common conditions, genetic contribution is one input among many, and the same variant plays out differently in different circumstances. Certainty in either direction, doom or immunity, is usually a sign the report has been over read.
Skipping the family conversation
Testing a single person can surface information that affects siblings, parents and children who never consented to know it. Thinking through in advance who you would tell, and how, prevents the worst version of that conversation happening by accident.