Any two people share about 99.9% of their DNA.

DNA is a sequence written using four chemical bases, commonly represented by the letters A, C, G, and T. Most positions in human genome sequences are shared. Differences include substitutions at individual positions as well as insertions, deletions, duplicated stretches, and other structural changes. A comparison must decide how to represent all those possibilities.

The familiar 99.9% summary is associated with a simplified view emphasizing small sequence differences. It should not be described as the percentage of genes that are identical or the fraction of a person that is shared with someone else. Sequence similarity, gene content, ancestry, and biological traits are related but distinct concepts.

Including larger variants changes how much DNA is counted as different. NHGRI now explains that a person's genome is, on average, about 99.6% identical to a reference sequence when broader variation is considered. That reference-based summary does not simply replace every pairwise comparison with a new universal percentage; it answers a specifically defined comparison.

One copy of the human genome contains roughly three billion bases. Even a small percentage of that enormous total can represent millions of differing positions or bases. This is why very high overall similarity is compatible with extensive measurable variation among individuals. The percentage alone can hide the absolute scale of the data.

Reference genomes give researchers an organized sequence against which they can compare newly sequenced DNA. A reference is a practical coordinate system, not a genetically ideal person and not a complete representation of everyone. Differences from it are called variants, but being a variant does not automatically make a sequence harmful or unusual.

NHGRI's Human Genomic Variation fact sheet explicitly includes multiple variant types and gives the approximate 99.6% reference comparison. It also describes the human pangenome, a more comprehensive framework intended to represent variation across populations. These developments explain why older educational shorthand needs qualification rather than being repeated as a timeless exact constant.

Most variants do not cause disease, and only a small subset has established effects on health. The importance of a particular change depends on where it occurs and what it does, not just how many letters differ overall. Comparing genomes therefore involves both counting differences and investigating their biological consequences.

People commonly share the same genetic code, meaning the rules used to translate DNA information into proteins. That does not mean every DNA sequence is identical. Separating the code from the sequence avoids a common wording error and makes clear that the headline concerns similarity of genomic material rather than a percentage of shared biological instructions in an everyday sense.

A similarity percentage is meaningful only after the comparison rules are specified. Comparing aligned single DNA letters is not the same as counting missing segments, extra copies, and rearranged stretches. The NHGRI discussion uses broader accounting than the familiar headline shorthand. Neither figure is a measure of how similar two people's personalities, abilities, or lives must be; it describes selected features of their DNA sequences.