Is there a DNA test for weight loss? What trials found, and what a DNA diet test can describe

No. In the trials so far, matching a diet to someone’s genes hasn’t helped them lose more weight. The clearest test of the idea is DIETFITS: 609 adults were randomly assigned to a healthy low-fat or a healthy low-carbohydrate diet for a year, and the researchers checked a three-variant gene pattern that had been proposed to show which diet suits whom. Weight change was similar on both diets, and the gene pattern didn’t show who would do better on either.^[13]

Yes, but in many small ways rather than one big one. A 2015 study of up to 339,224 people found 97 regions of the genome linked with body mass index (BMI). Together they explained about 2.7% of the differences in BMI between people, though common variants as a whole may explain more than 20%.^[10]

The best-known gene is FTO. People with two copies of its common higher-BMI version weighed about 3 kg (6.6 lb) more, on average, than people with none.^[9] That is an average across thousands of people, not a forecast for you.

The variant didn’t change how people responded to weight programs, either. When eight trials of diet, activity or medicine programs were pooled (9,563 people in all), people who carried the FTO variant saw the same changes in weight and waist size as people who didn’t.^[11]

Not in the trials so far. Two of the best known tested the idea directly:

• Food4Me. Adults in seven European countries were randomly given general healthy-eating advice or personalized advice. Some personalized groups also had advice based on five diet-related gene variants. After six months, personalized advice changed diets more than general advice did, but adding the gene results made it no more effective. 1,269 people completed the trial.^[12]
• DIETFITS. The year-long low-fat and low-carbohydrate trial above: people whose diet matched their gene pattern lost no more weight than people whose diet didn’t.^[13]

In Food4Me, advice built on what people already ate did as well as advice that also used their genes. There was no evidence that adding gene results made the advice work better.^[12] That is why we describe a DNA result as a picture of some of your traits, not a plan.

Little you can act on. Some tests report FTO or other body-weight variants, but each is linked with a small average difference across large groups, not with what you will weigh.^[10]

What these tests mostly read is food. A nutrition DNA test, also sold as a nutrigenomics test, looks at a small set of variants that studies have linked with food-related traits. It reads the DNA you inherited, which stays the same through your life.

Tests like ours use genotyping: the lab checks selected variants, chosen in advance, at known places in your DNA. It doesn’t read whole genes. A result comes back in four steps:

The clearest links are the ones where one gene does most of the work. Most traits are shaped by many genes, each with a small effect, plus what you eat and how you live. Here are five traits these tests often report, and how strong the evidence is for each.

Partly. Babies make lactase, the enzyme that digests lactose, the sugar in milk. About 65% of people have a reduced ability to digest lactose after infancy.^[1]

Whether you keep making lactase is set by a stretch of DNA that controls the LCT gene. It sits inside a neighboring gene called MCM6. Some people inherit changes there that keep lactase production going into adulthood, which is called lactase persistence.^[1]

It varies a lot by ancestry. Lactase non-persistence affects 70 to 100 percent of people of East Asian descent, but only about 5 percent of people of Northern European descent.^[1] Different variants do the same job in different places, too: the best-known one is common in Europe, while at least three others are linked with lactase persistence in East Africa.^[3]

The closest thing is CYP1A2. It makes the enzyme that breaks down about 95% of the caffeine you take in.^[4]

In a study of 47,341 people of European descent, variants near CYP1A2, and near AHR, a gene that controls it, were linked with how much caffeine people habitually took in.^[5]

Genes are only part of the picture. Age, sex and hormones, smoking and diet all change how quickly caffeine is cleared.^[4] A caffeine result describes a tendency, not how a particular cup will affect you.

• It isn’t a diagnosis. A trait result can’t tell you whether a food causes your symptoms or whether you have a health condition.
• It isn’t a diet or a weight plan. Adding gene results to diet advice hasn’t made it work better in trials.^[12][13]
• It can’t measure your levels. Only a blood test shows your vitamin D level, for example.^[7]
• Your life counts too. What you eat now, how active you are, your sleep, age, hormones and any medicines all shape these traits.
• Genotyping reads selected variants. It can miss variants it doesn’t look for, and the variants that matter can differ between ancestries.^[3]
• Much of the research is in people of European descent. A link found in one group may not hold the same way in another.^[5][6]

• One sample, one report. GeneCore reads diet, fitness and everyday health traits from one saliva sample and sets them side by side.
• The diet part. Selected variants studied in carbohydrates and fats, appetite, fullness and taste, caffeine and lactose.
• Traits, not a meal plan. Each line gives your result and what research associates with it. It won’t tell you what to eat.
• Want diet only? GeneDiet reads the diet area on its own.

Our nutrition, supplement, athletic and wellbeing reports are for general wellness and information. The evidence behind these associations is less established than for our clinical panels, and these reports are not intended to diagnose, treat, cure or prevent any disease.

Want the background first? Read what nutrigenomics is, see how Nutrigenomics works at GeneMatrix, or find the test that fits you.