Buzz
In science, replication holds significant value. A single experiment might be a coincidence, but repeated results from multiple trials bring us closer to the truth. So, when a DNA app named BabyGlimpse claimed it could predict what my children might look like, I was excited. With three kids, I had the chance to verify its predictions three times over.
BabyGlimpse is one of many DNA testing services that emerged last year in collaboration with Helix. After sending in a saliva sample to Helix, they sequence your DNA and store your data. You can then allow partner companies to access specific data relevant to their services. Insitome tells you how closely you're related to Neanderthals. Vinome analyzes your taste preferences to recommend wines. And BabyGlimpse examines a few genetic markers from both you and your partner to predict what your potential child might look like.
Initially, this may seem overly romantic, akin to the digital age's version of scribbling your name alongside your crush’s. However, it costs $260 (or $99 if both of you have already submitted a saliva sample to Helix), and it requires sharing some of your DNA data with your partner. It's limited to basic traits, no Alzheimer’s genes or similar serious conditions are involved—but it’s always worth considering just how much information you’re comfortable sharing.
As with any DNA test, there are important disclaimers. BabyGlimpse analyzes your ethnic background, which could uncover information that challenges family stories and lead to awkward discussions. Additionally, the process only works if you submit your saliva to Helix, so you’ll need to be at ease with their privacy policies.
How It Works
Helix appears to
Even though Helix analyzes large portions of your genome, the insights from partner companies like BabyGlimpse function similarly to genotyping services like 23andMe. They sift through your DNA for specific single-letter variations that have been linked in scientific studies to certain traits. For instance, Europeans with two cytosines (“CC”) at a specific location, rs4988235, are often lactose intolerant. However, there's some ambiguity here: one could be lactose intolerant without this genetic variation, and this marker doesn’t predict anything if your ancestors weren't European.
So, even for basic traits like lactose intolerance, it’s important to approach the results cautiously. BabyGlimpse suggests that both my husband and I are likely lactose-tolerant and can consume all the milk we desire, and our children will probably share this trait. (Although they drink milk fine, they are still too young to be certain. Even lactose-intolerant kids often don’t show symptoms right away.)
But the primary reason people send their spit isn’t to find out if their kids will be lactose intolerant—it’s about exploring the potential traits of future offspring. The company’s promotional images focus on predicting hair and eye color, promising you a glimpse into what your baby could look like.
The Insight I Gained
Before receiving my husband's results, I had the chance to check my own. The app predicted that I had a 68 percent chance of having brown hair. That turned out to be correct—while I was blonde as a child, my hair is now brown. However, the app's prediction for my eye color was off; it guessed blue, with only a 27 percent probability of my actual hazel/green eyes.
When my husband's results arrived, they indicated that he has both brown hair and brown eyes (which is accurate). The app also suggested that our future children would likely share these traits. That’s a win: all three of my children fit this description.
There isn’t a single DNA marker that determines hair or eye color, as many genes contribute to the creation of the brown pigment and its transport to the correct areas. BabyGlimpse computes a score based on several distinct DNA markers.
But what I appreciate most is how they provide context for the uncertainty. Few traits are purely genetic; environmental factors like diet, upbringing, prenatal experiences, or even unknown influences can play a role. There’s also an element of chance involved. Regarding hair color, BabyGlimpse states that genetics account for 86 percent of this trait, with the markers they use explaining 70 to 80 percent of that. For eye color, the app claims 96 percent genetic influence, with markers covering 80 to 90 percent of that. One of the studies they reference suggests that this level of prediction can apply to individuals worldwide, making it a more reliable predictor than the lactose intolerance marker, which primarily pertains to Europeans.
My husband tends to tan easily and has a slightly darker complexion than I do, but our results categorized him as having 'very light' skin, while the kids and I were labeled as having 'light' skin. I’m not sure what the distinction is supposed to mean, but alright then.
The height prediction didn’t seem to match up. BabyGlimpse claimed I was taller than average and my husband shorter. According to this calculator, we’re both above average, with him being taller than me. Regardless, our children are expected to be tall, and that prediction seems fairly accurate.
The hair curliness prediction missed the mark, and it also lacked specific details. Instead of offering a breakdown of probabilities for straight, wavy, and curly hair, it simply stated that we and our children would all have 'wavy' hair. That’s true for me and one of the kids, but my husband and one of my sons have very curly hair, and the third child’s hair is too short to tell. BabyGlimpse acknowledges that the known DNA variations only account for a 'relatively low proportion (~20-50%)' of the variability in hair curliness. So, this test isn’t particularly helpful.
There’s a prediction for male pattern baldness, which isn’t exactly a trait we’re concerned about for babies. To be fair, all of these traits are about a person’s appearance as an adult. Many newborns have blonde hair, blue or gray eyes, and are very short. So, this doesn’t exactly help you visualize your baby but rather gives you an idea of what to expect in 20 years or more, long after you’ve forgotten about your BabyGlimpse results.
Things take a strange turn with the 'just for fun' category. Based on one measly marker tested in elite athletes, BabyGlimpse labels our children’s 'muscle fiber type' as 'normal muscle composition. Gonna have to work hard (like most of us).' Call me skeptical (who, me?), but I don’t think this is a particularly insightful result, especially when they admit—or perhaps guess—that this marker accounts for less than 5 percent of the genetics involved.
BabyGlimpse suggests our children are likely to have a bit of a sweet tooth, based on the interpretation of just one marker. Additionally, their personalities are predicted to be either 'strategist' or 'mixed warrior-strategist' based on another tiny variation. According to SNPedia, which I’m using for the links for these markers, 23andme categorizes this as 'worrier' versus 'warrior.' Regardless, I really don’t believe my kid’s personality can be determined by a single nucleotide.
To their credit, BabyGlimpse keeps personality traits in the 'just for fun' section and makes it clear that the marker they test explains only 'around 1%' of the known genetic variation, which itself only accounts for 60 percent of the total variation. But honestly, when you know a result explains just 0.6 percent of a trait, what’s the point?
What We Took Away From This
BabyGlimpse is accurate in positioning itself as purely for entertainment. I can't think of a single significant decision I’d make based on these results, and perhaps that’s a good thing. But without it, what would we have known? We probably would have guessed our kids would have brown hair, brown eyes, and light skin, and still been left wondering if they'd grow up to be 'warriors,' 'strategists,' or elite athletes.
So, we learned nothing substantial. But it was kind of fun? Now we can blame our kids' DNA (which ultimately means blaming ourselves) when they inherit our sweet tooth. Then again, we were probably doing that anyway.
