BioSample to run, via the ENA portal¶
Route id: biosample->run:ena_filereport
Provider: EBI ENA portal
Direction: SAMN/SAMEA/SAMD → SRR/ERR/DRR
curl -s 'https://www.ebi.ac.uk/ena/portal/api/filereport?accession=SAMN36028297&result=read_run&fields=run_accession&format=tsv&limit=0'
The route that makes the BioSample a usable join key rather than only a shared label. Two of this project's recoveries end here:
- GEO omitting the SRA relation — GSE135325 and GSE137444 record a BioSample per sample and nothing else, so the BioSample is the only way from a GSM to its reads.
- ArrayExpress studies with no secondary accession — E-MTAB-6505 is reachable only through the BioSamples its SDRF names.
Measured¶
Corpus: a seeded 3,000-BioSample draw from the 363,885 that the BioProject → BioSample census returned. Surveyed: 2026-09-15.
| Resolved | 3,000 (100%) |
| Empty | 0 |
| Failed | 0 |
| Unique runs | 4,827 |
1.6 runs per BioSample, and nothing empty. That total is not a surprise and
should not be read as a claim about BioSamples in general: the corpus was built
from what result=read_run returned, so every accession in it is one ENA
already associates with at least one run. It measures the round trip, not
coverage.
The honest coverage claim is the one on the BioProject → BioSample page, where the two routes disagree by 125,271 samples and the disagreement is explained by which archive registered the project.