All articles
Cancer Genetics· 3 min read

Reference Genome Builds: Why GRCh37 vs GRCh38 Appears on Reports

A genetic variant is described by its position relative to a reference genome. Two builds are in common use, GRCh37 and GRCh38, and a variant has different coordinates in each. Getting this wrong causes confusion when comparing reports or searching databases.

Quick Answer

A genetic variant is described by its position relative to a reference genome. Two builds are in common use, GRCh37 and GRCh38, and a variant has different coordinates in each. Getting this wrong causes confusion when comparing reports or searching databases.

Reference Genome Builds: Why GRCh37 vs GRCh38 Appears on Reports: mechanism and interpretation mapThree connected stages summarise the article's mechanism, measured effect and interpretation boundary.BRCA1 · TP531What a Reference Build IsMechanism2Why Coordinates DifferObserved consequence3Transcript Versions Matter TooInterpret in contextGene or pathway evidence → measured phenotype → assay-aware conclusion
Mechanism map: the article’s main biological stages are separated from the final interpretation so a pathway relationship is not mistaken for a clinical conclusion.

What a Reference Build Is

The reference genome is a consensus sequence against which a person's DNA reads are aligned and compared. It is periodically updated by the Genome Reference Consortium as gaps are filled and errors corrected.

GRCh37 (also called hg19) was released in 2009; GRCh38 (hg38) in 2013, with continued minor patches. GRCh38 corrects thousands of errors, closes gaps and better represents regions that vary between individuals.

Why Coordinates Differ

Because sequence was added and rearranged between builds, the same variant sits at a different numerical position in GRCh37 and GRCh38. A report should state which build it used.

Variant nomenclature at the protein or transcript level (for example a specific amino-acid change) is more stable, which is one reason clinical reports emphasise that description alongside genomic coordinates.

Explore:BRCA1

Transcript Versions Matter Too

A coding variant is numbered relative to a chosen reference transcript for the gene. Genes can have several transcripts, and using a different one changes the coding position and sometimes the predicted consequence.

Laboratories increasingly use standardised reference transcript sets to reduce this ambiguity, but older reports may use different transcripts.

Explore:TP53

Practical Consequences

When comparing a new report to an older one, or looking a variant up in a database, the build and transcript must match or coordinates must be converted (lifted over). Liftover is usually reliable but can fail in rearranged regions.

Population frequency resources are build-specific, so a frequency looked up in the wrong build can be misleading.

Interpretation Notes

Check the genome build and reference transcript on any variant report before comparing it with another source.

The protein-level description of a variant is the most portable identifier and is what most curated clinical databases key on.

Key Takeaways

  • ·GRCh37 and GRCh38 are successive reference builds; a variant has different coordinates in each.
  • ·Reports should state the build and the reference transcript used.
  • ·Protein-level variant descriptions are more stable than genomic coordinates.
  • ·Build mismatches cause errors when comparing reports or looking up population frequencies.

Put these genes in pathway context

Frequently asked questions

What is the key idea in Reference Genome Builds: Why GRCh37 vs GRCh38 Appears on Reports?

A genetic variant is described by its position relative to a reference genome. Two builds are in common use, GRCh37 and GRCh38, and a variant has different coordinates in each. Getting this wrong causes confusion when comparing reports or searching databases.

What should be kept with the result or mechanism?

Reports should state the build and the reference transcript used. Protein-level variant descriptions are more stable than genomic coordinates. Build mismatches cause errors when comparing reports or looking up population frequencies.

References

  1. 1Clinical Validation of Genome Reference Consortium Human Build 38 in a Laboratory Utilizing Next-Generation Sequencing. Clinical Chemistry, 2022. PubMed
  2. 2100,000 Genomes Pilot on Rare-Disease Diagnosis in Health Care - Preliminary Report. New England Journal of Medicine, 2021. PubMed
  3. 3Guidelines for Validation of Next-Generation Sequencing-Based Oncology Panels. Journal of Molecular Diagnostics, 2017. PubMed
  4. 4Interpreting variants in genes affected by clonal hematopoiesis in population data. Human Genetics, 2023. PubMed

Continue Reading

Choose your next research step

Move from this explanation into a gene profile, a pathway map, or the next evidence update.

BRCA1 has 100+ trials currently recruiting on ClinicalTrials.gov. The GeneAnalyses digest summarises the new and changed ones each day.