Run to experiment, via the ENA portal

Route id: run->experiment:ena_filereport Provider: EBI ENA portal Direction: SRR/ERR/DRR → SRX/ERX/DRX

curl -s 'https://www.ebi.ac.uk/ena/portal/api/filereport?accession=SRR25056225&result=read_run&fields=experiment_accession&format=tsv'

Walking up the hierarchy, so that a bare run accession — the thing a collaborator actually pastes into an email — is a valid entry point to the whole graph.

Measured

Corpus: sample:3000@reprocessed-srr, a seeded draw from the 262,690 runs the pipeline has reprocessed. Surveyed: 2026-09-15.

Resolved	2,977
Empty	23
Failed	0
Unique experiments	2,888

2,977 runs to 2,888 experiments — runs share experiments, as they should: a 10x library sequenced across lanes is one experiment and several runs.

The 23 empties are runs that no longer exist

They are not a defect in this route. Every one of the 23 was checked against NCBI directly, and all 23 are absent from esearch db=sra as well:

curl -s 'https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=sra&term=SRR28779487&retmode=json'
# "count": "0"

Every one of these runs was reprocessed successfully at some point — that is how it got into the corpus — and is now unresolvable in either archive. Withdrawn, suppressed, or moved to controlled access; the archives do not distinguish, and nothing announces any of them.

0.77% of a random sample of already-processed runs. That is the attrition rate of public sequencing data over the corpus's lifetime, and it is the reason fetch cache diff and the archive-state monitor exist: a resolution that worked last year is not evidence that it works today, and the only way to know is to keep asking and record the dates.