Why are nutrients missing from food composition tables?

Because analysing each nutrient in a laboratory is expensive and needs its own method. The agencies that publish the tables, such as the USDA or ANSES-CIQUAL, analyse what people look up most and leave the rest blank, for example the amino acids or choline of many foods.

How does vitaloop estimate missing nutrients?

With a proprietary algorithm. It only estimates what the source table does not publish, the estimated parts never add up to more than the known total (amino acids never exceed protein, for example), and every nutrient is compared with laboratory values before it appears in the app.

What does ≈ in front of a value mean in vitaloop?

That it is vitaloop's estimate: the table publishes no value for that nutrient in that food. Tap the value to see its range. A value without ≈ comes from the table.

Which nutrients does vitaloop estimate?

In the app today: the 18 amino acids in protein, choline, and some minerals, such as magnesium, phosphorus or sodium, for foods whose table leaves them out. It is free for catalogue foods and your meals.

Are the estimates reliable?

They are the best figure available without a lab test, not a lab test. That is why they carry ≈ and a range, and a nutrient whose estimates fail the comparison with laboratory values is not shown.

Missing nutrients in food composition tables: how we estimate them

Published 4 November 2026 · Vitaloop Labs

Food composition tables have gaps. Many give a food's protein but not its amino acids, or its main vitamins but not its choline. vitaloop fills those gaps with a proprietary algorithm, marks every estimate with ≈ and shows you its range. It never replaces a measured value, and what it estimates never exceeds the known total.

vitaloop is free by default, with no ads. Without paying you log food, drinks and supplements, and see the calories, macros and micronutrients of every food, estimates included. Pro adds the AI: logging by photo or voice, the weekly report, patterns, and estimated micronutrients for packaged products.

Why the tables have gaps

Analysing a food in a laboratory is expensive, and every nutrient needs its own method. The agencies that publish the tables, such as the USDA or France's ANSES-CIQUAL, analyse what people look up most and leave the rest blank. So a food can have measured values for energy, protein and iron, and none at all for amino acids or choline.

Estimating a missing value is not our idea. The FAO handbook on food composition data separates analysed values from imputed ones, estimated from similar foods, and asks that every imputed value be labelled as such (Greenfield and Southgate, FAO, 2003). The USDA and CIQUAL already publish some values they have not measured, and vitaloop marks its own with ≈. How we calculate covers how much of each vitaloop value is measured.

Which nutrients we estimate

All 18 amino acids in protein, in the Amino acid profile section of every food and every meal (see what an amino acid profile is).
Choline, at the end of Vitamins. EFSA sets an adequate intake of 400 mg a day for adults (EFSA, 2016), yet few tables publish it.
Some minerals, such as magnesium, phosphorus or sodium, for foods whose table leaves them out.

More nutrients are already estimated and will appear in the app once they pass the same checks.

How we do it

The estimates come from a proprietary algorithm built by vitaloop. We won't go into how it works inside, but here are the rules it follows, because they are what lets you trust the result:

A measured value always wins. We only estimate what the source table does not publish. If the table has a value, that is the one you see.
The parts fit inside the total.If the table gives a food's protein, its estimated amino acids never add up to more than that protein. The same goes for sugars and for each type of fat.
Checked first, shown second. Before a nutrient appears in the app, we compare its estimates with laboratory values the algorithm has never seen. If it fails, it stays out of the app, even though we have the numbers.
No data, no zero. If we cannot estimate a nutrient reliably, we leave it blank. A zero would suggest the food contains none.

How to read a value with ≈

A value with ≈ in front, such as ≈ 0.65 g, is our estimate. Tap it to see the range we expect the real value to fall in. The narrower the range, the more confident the estimate.

For a meal, the total adds up the measured and estimated values of each food. If a good part of that total comes from estimates, it carries ≈ too. And if a food in the meal has no value at all, the app tells you how much of the meal the figure covers.

What about packaged products?

When you scan a product, every plan shows what its label declares. With Pro, vitaloop adds the vitamins and minerals the label leaves out, marked as estimates. More in which vitamins and minerals a scanned product has.

What an estimate is not

It is not a lab test. It is the best figure we can give without one, which is why it is marked.
A table value is an average. It sums up the samples the source analysed, and your food can differ from it with the variety or the season.

If a value looks off, email us at support@vitaloopapp.com with the name of the food.

Sources

Greenfield H, Southgate DAT. Food composition data: production, management and use. 2nd ed. Rome: FAO, 2003, ch. 1, pp. 7–8.
EFSA NDA Panel. Dietary Reference Values for choline. EFSA Journal 2016;14(8):4484.

vitaloop is a personal nutrition and data-logging app developed by Vitaloop Labs, SL. It is not a medical device as defined under EU Regulation 2017/745 (EU MDR) and does not constitute professional medical advice. Personalized suggestions are based on your personal data and should not be used as a substitute for advice from a qualified healthcare professional. Nutritional data may contain inaccuracies; always verify critical information on product labels.