Build 37 vs 38 — why the same rsID can move

Genome positions are not absolute — they are measured against a reference sequence that has been updated. It affects your raw data file more than you might think.

What a "build" is

The human reference genome is a standard sequence that all positions are measured against. Build 37 (GRCh37, from ~2009) was the standard for many years; Build 38 (GRCh38, from ~2013) corrects errors and fills gaps — see NCBI's reference genome documentation. The positions in a raw data file are coordinates against whichever version was exported.

Why the coordinates differ

Between builds some regions have been inserted, removed or remapped — a SNP at position 11856378 in Build 37 can sit at a different coordinate in Build 38. Same biological point, same rsID — but a different position number in the file.

rsID is your stable reference

That is why the rsID (e.g. rs1801133) is the reliable identifier: it points to the same genetic variant regardless of which build the coordinates come from (in rare cases old rsIDs are merged in dbSNP — then both point to the same variant record). DNAHelper matches on rsID rather than raw coordinate — so a Build 37 export (like FamilyTreeDNA's) and a Build 38 file are interpreted consistently. See the FTDNA guide for their Build 37 export.

Practical effect for you

  • Different files, same rsID, different coordinates — normal, not an error.
  • Tools that match on coordinate can misread a Build 37 file if they assume Build 38 — DNAHelper avoids this through rsID mapping.
  • The chip set matters more for your results than the build version — different providers measure different markers.

When it matters

Mainly if you cross-reference positions against research databases manually, or compare old FTDNA exports against newer ones. For everyday analysis in DNAHelper you do not need to think about build — but if you ever wonder why a position number "does not match", this is the answer.

Analyze my file locally

Also: What is a raw DNA file? · rsID lookup — free tool · FTDNA guide · Local vs cloud analysis · English overview