The DNA diet
does not work.
You have seen the promise. Match your diet to your genes and lose two or three times more weight. It is the engine of this entire industry.
It has been tested properly, twice. At Stanford, 609 adults were sorted into low-fat and low-carb genotype patterns and followed for a year. The genotype did not predict which diet worked better for anyone. The difference between groups was under a kilogram and the interaction was not significant.
An independent team repeated it in 2023, putting people on diets that either matched or mismatched a ten-variant profile, with assessors blinded to both. The matched group did better by about half a kilogram, which is to say by chance.
The original claim came from a conference talk that was never published as a paper.
The researcher behind that talk is the one who ran the Stanford trial that found nothing. We would rather tell you that up front than build your report on it.
The diet by genotype interaction. In plain terms, no effect.
Genotype-matched diets beat mismatched by 0.6 kg over twelve weeks. Indistinguishable from chance.
Across 18 randomised trials, being told your DNA-based risk changed smoking, diet and activity by nothing measurable. It did not change motivation either.
941 variants together. A report showing you four of them is showing you a rounding error.
Five results that
change a decision.
Short list on purpose. These are the variants where the biology is understood, the evidence is replicated, and knowing the answer changes something specific you would otherwise guess at.
Your dose is not everyone's dose
In a double-blind trial over a 10 km time trial, fast metabolisers went 6.8 per cent faster on 4 mg per kg. Slow metabolisers went 13.7 per cent slower on exactly the same dose. More is not better for everyone, and the cut-off time before bed differs too.
Caveat we will give you: one strong endurance trial, and a strength study that found no genotype effect. Treat it as a dose to test, not a prescription.
The flush response is a real risk signal
If you go red after one drink, you likely clear acetaldehyde poorly. That carries a materially higher oesophageal cancer risk at any given intake. The variant is common across East Asia and rare in Europe, which matters in Bangkok.
This is a reason to drink less, which is the one piece of genetic advice we give without hedging.
A result that sends you to a doctor
Two copies of the main haemochromatosis variant is a reason to get iron studies done. Around 1 in 10 men with it go on to serious liver disease if the iron loading is never found. The other 9 do not, which is the point.
We refer this. We do not manage it on a gym floor.
A reason to test, not a dose
Certain variants make low vitamin D around 2.5 times more likely. In Bangkok, where most people assume the sun handles it, that is worth knowing. It tells you to measure and keep measuring.
It predicts your starting point, not how much you need. Blood work answers that better than DNA does.
Usually confirmation, occasionally a relief
Lactase persistence is one of the cleanest gene to trait links there is. Most adults already know how dairy treats them, so this rarely surprises anyone. When it does explain years of guessing, it is worth the line on the report.
How fast you respond is itself inherited
481 adults trained for 20 weeks. Average gain in aerobic capacity was solid, but some gained almost nothing and others more than a litre a minute. Around 47 per cent of that difference in response was heritable.
No test on the market predicts which one you are. That is precisely why we measure you every two weeks instead of comparing you to an average.
The panel is
shorter on purpose.
Plenty of tests report dozens of variants. Length is easy. Deciding what not to report is the harder and more useful job.
There is also a reliability problem nobody in this industry volunteers. When 49 samples were sent for clinical confirmation of variants flagged in consumer DNA data, 40 per cent of them turned out to be false positives.
And almost all of these scores were built in European populations. For a Thai or wider Asian client, several are of reduced or simply unknown accuracy. We say so rather than printing a number that looks precise.
The American College of Medical Genetics formally advises against testing it. The association behind it has been disproven by meta-analysis. It is the most overstated variant in this industry and we do not report it.
ACTN3 is a real finding about populations of top-level athletes and tells you nothing about one person. Around 18 per cent of Europeans carry none of the protein at all and are perfectly capable. We do not programme from it.
The tendon and ligament genetics literature is small candidate studies in athlete groups. There is no replicated finding strong enough to tell a paying client their injury risk. We will not invent one.
Variants such as APOE carry real clinical, insurance and personal weight, and belong with a physician and proper counselling. Anything medical sits with Orba Health, who prescribe. We coach.
The consensus position across sport and exercise genetics is that no child should be genetically tested to define training or identify talent. We hold that line, including for families who ask.
One input among
several, and not the loudest.
Your genes
Tested once, relevant for life. Which is also why the report is worth keeping short and honest rather than impressive.
Your blood work
Drawn and interpreted by Orba Health, our medical partner. This is where most of the actionable health information lives.
Your own numbers
InBody body composition every two weeks, strength, movement and recovery. The data that actually moves, and the data your plan is built on.
Genotype shifts probabilities. Behaviour shifts outcomes, and the gap between those two is where coaching happens. The best illustration is the headline obesity variant: it is worth around 3 kg, and being physically active cuts its effect by roughly a quarter.