Written and maintained by CASRAI Editorial Board
Last updated
GTEx stands for the Genotype-Tissue Expression project, an NIH Common Fund programme that links inherited genetic variation to gene expression across human tissues. This page is for researchers who need to know what the GTEx Portal holds, which release to use (V7, V8, V10 or V11), and where the programmatic API lives. Facts below were read from the NHGRI project page and the GTEx Portal API v2 on 2026-11-07.
What is GTEx?
GTEx is a resource of gene-expression and genotype data from post-mortem donors, built to help explain how inherited gene changes contribute to common diseases. According to the National Human Genome Research Institute (NHGRI), NIH launched it in September 2010 as an NIH Common Fund programme, managed by the NIH Office of the Director with NHGRI, NIMH, NCI and other institutes. The programme builds a database and a tissue bank, and the GTEx Portal hosts the latest release together with an eQTL browser. An eQTL is a genetic variant associated with differences in the expression level of a gene; see the CASRAI guide to genomics for the wider field.
Who runs it and what it covers
NHGRI states that GTEx is funded through the NIH Common Fund and managed by the NIH Office of the Director in partnership with NHGRI, the National Institute of Mental Health, the National Cancer Institute and other NIH institutes. Its scope is human post-mortem donors: the final analyses described by NHGRI cover genotype data from about 948 donors and about 17,382 RNA-seq samples across 54 tissue sites and 2 cell lines.
What does GTEx stand for?
GTEx stands for Genotype-Tissue Expression, the title NHGRI uses for the project. The abbreviation is shared with unrelated uses, so search for “GTEx portal” or “Genotype-Tissue Expression” when you want the research resource.
How many individuals are in GTEx?
It depends on the release and on what is being counted. NHGRI reports 961 donors enrolled by December 2015 and, for the final 2020 analyses, about 948 post-mortem donors, 54 tissue sites plus 2 cell lines, and about 17,382 RNA-seq samples. The portal’s own dataset metadata gives subject counts per release, and the number of subjects used for eQTL analysis is smaller than the number of subjects in the release.
How does GTEx work?
- Tissue collection. NHGRI describes two donor routes: organ and tissue donors (or their families) who allow leftover tissue that cannot be transplanted to go to research, and surgical donors who agree in advance that tissue removed for medical reasons may be used. Transplant use takes priority, and tissues are collected and stored through a National Cancer Institute biobank on GTEx’s behalf.
- Matched genotype and expression data. The project pairs donors’ genotype data with RNA-seq data from their tissues, which is why the release tables below report both RNA-seq samples and RNA-seq plus genotype samples.
- eQTL analysis. Using those paired data, GTEx identifies expression quantitative trait loci (eQTLs), genetic variants associated with gene-expression differences. NHGRI notes that eQTLs vary between people and tissues, and some affect a gene in one tissue while others act across many.
- Release through the portal. Results are published as numbered releases on the GTEx Portal, which hosts an eQTL browser for exploring relationships between genes and the DNA variants affecting them.
GTEx releases compared (V7, V8, V10, V11)
Choosing a release matters because the genome build and annotation differ. The table is taken directly from the portal’s dataset metadata endpoint (read 2026-11-07).
| Release | Subjects | Tissues | RNA-seq samples | RNA-seq + genotype samples | eQTL subjects / tissues | Genome build | GENCODE | dbSNP |
|---|---|---|---|---|---|---|---|---|
| GTEx Analysis v7 | 714 | 53 | 11,688 | 10,294 | 620 / 48 | GRCh37/hg19 | v19 | 147 |
| GTEx Analysis v8 | 948 | 54 | 17,382 | 15,201 | 838 / 49 | GRCh38/hg38 | v26 | 151 |
| GTEx Analysis v10 | 946 | 54 | 19,788 | 15,253 | 943 / 50 | GRCh38/hg38 | v39 | 155 |
| GTEx Analysis v11 | 946 | 54 | 19,788 | 15,253 | 943 / 50 | GRCh38/hg38 | v47 | 155 |
Notes from the same metadata: V7 changed expression units from RPKM to TPM; V8 renamed the tissue site “Cells – Transformed fibroblasts” to “Cells – Cultured fibroblasts”; V10 is described as the most complete analysed dataset for adult GTEx and adds a new data type, small RNA-seq. The portal also lists a separate Kid’s First harmonization for CFDE (456 subjects, 5 tissues).
Practical consequence: V7 uses GRCh37/hg19 while V8, V10 and V11 use GRCh38/hg38, so variant coordinates must be converted before combining V7 results with later releases or with other hg38 datasets. The portal’s V7 record also carries the dbGaP accession phs000424.v7.p2; for what dbGaP is, see dbGaP.
Using the portal and the API
The GTEx Portal (gtexportal.org) is the project’s public front end, and it also offers a mobile site. Programmatic access is through the GTEx Portal V2 API (version 2.0.0, published by the GTEx Project), which the service description says “provides access to the data powering the GTEx portal.”
- Open the API service root at
https://gtexportal.org/api/v2/for the service description. - Read the interactive documentation at
https://gtexportal.org/api/v2/docs. - Use the OpenAPI specification at
https://gtexportal.org/api/v2/openapi.jsonto generate a client in your language of choice. - Query the dataset metadata endpoint,
/api/v2/metadata/dataset?format=json, to see the releases and their counts before choosing adatasetIdsuch asgtex_v10. - Record the release identifier and the genome build with every result, so analyses can be reproduced.
Strengths and limits
- Strength: one programme covers 54 tissue sites with matched genotype and RNA-seq data, so expression and genetic variation can be studied together (NHGRI, portal metadata).
- Strength: a documented JSON API with an OpenAPI specification supports scripted, reproducible retrieval.
- Limit: donors are post-mortem, so the data describe tissue collected after death rather than living-person samples (NHGRI).
- Limit: counts differ by release and by analysis (for example 946 subjects but 943 for eQTL in V10), so quote the release name with any number.
- Limit: the V7 to V8 genome-build change complicates cross-release comparisons.
Related resources on CASRAI
GTEx is a subject-specific resource, so it sits alongside other data repositories in a research data plan. See institutional versus subject repositories for how such resources differ from a university repository, and biorepository for tissue banks, which NHGRI describes as part of the programme alongside the database.
FAQ
What does GTEx stand for?
Genotype-Tissue Expression, the title NHGRI uses for the NIH Common Fund project.
How many individuals are in GTEx?
It depends on the release: NHGRI reports 961 donors enrolled by December 2015 and about 948 post-mortem donors in the final analyses; the portal lists 714 subjects for v7, 948 for v8 and 946 for v10 and v11.
How does GTEx work?
Donated post-mortem tissues are collected, matched genotype and RNA-seq data are generated, eQTLs are identified, and the results are released through the GTEx Portal and its eQTL browser.
What is the GTEx portal?
The GTEx Portal at gtexportal.org hosts the latest GTEx release and an eQTL browser, per NHGRI.
Which GTEx version is the newest?
As of 2026-11-07 the portal’s dataset metadata lists v11 (GENCODE v47) after v10 (GENCODE v39), both on GRCh38/hg38.
Is there a GTEx API?
Yes. The GTEx Portal V2 API is documented at gtexportal.org/api/v2/docs, with an OpenAPI specification at gtexportal.org/api/v2/openapi.json.
Sources
- NHGRI, Genotype-Tissue Expression Project (GTEx): genome.gov/Funded-Programs-Projects/Genotype-Tissue-Expression-Project (read 2026-11-07).
- GTEx Portal V2 API service description and dataset metadata: gtexportal.org/api/v2/ and gtexportal.org/api/v2/metadata/dataset?format=json (read 2026-11-07).
