NCBI GEO

ncbi.nlm.nih.gov · Storage · 2 records · JSON

NCBI SRA — with anonymous read via AWS Open Data Storage★ qualifiesanonymousno cardR
NCBI GEO Storagecaveatanonymousno cardRwrite-gated

NCBI SRA — with anonymous read via AWS Open Data Storage

★ qualifiesanonymousno cardR

www.ncbi.nlm.nih.gov/sra · academic · storage/ncbi-sra-with-anonymous-read-via-aws-open-data · JSON · compare

Free tier	No per-account quota published. Submission is genuinely bulk: FTP, Aspera (`ascp -QT -l 100m -k1 -d`), or Aspermera; operator guidance for 'tens or more terabytes' is to contact SRA support directly. unit mixed; metered on unstated
Price	none stated
Per-object cap	not stated
Retention	per-object-cap: none on deposit; individual runs of hundreds of GB are normal. Retention: permanent, no expiry
Egress	Free. Read side is mirrored to AWS Open Data (`sra-pub-run-odp.s3.amazonaws.com`) with anonymous, ranged, CORS-open access — no requester-pays, no auth.
Identity	anonymous derived from the record's prose, not from an identity field, and no signup was executed: few minutes from logging in'); Aspera is the working path. Anonymous S3 mirror can lag NCBI's own copy. Deposit is per-submission and effectively unbounded; the AWS mirror means the fetch
Binding	none no constraint in the record binds at this scale; the record carries no limit of a kind the vocabulary names
Interface	s3, sftp sftp
Tooling	Aspera Connect / ascp, sra-tools (prefetch/fasterq-dump), and the AWS Open Data mirror which behaves like any public S3 bucket.
Composes with	Deposit is per-submission and effectively unbounded; the AWS mirror means the fetch workload does not touch NCBI's own edge at all, so read capacity is Amazon's, not NCBI's. Each public run is an independently addressable anonymous object.
Reliability	Historically flaky FTP uploads ('frequent drop-outs, you only have a few minutes from logging in'); Aspera is the working path. Anonymous S3 mirror can lag NCBI's own copy.
Notes	No per-account quota published. Submission is genuinely bulk: FTP, Aspera (`ascp -QT -l 100m -k1 -d`), or Aspermera; operator guidance for 'tens or more terabytes' is to contact SRA support directly.
Evidence rung	reported One page read moves it to documented.
Checked	2026-10-21 (inherited); stale on 2026-11-04

Evidence (3)

date	source	what it proved
2026-10-21	www.ncbi.nlm.nih.gov/sra/docs/submitfiles/	FTP + Aspera ascend commands, 'recommend it for all medium to large submissions'
2026-10-21	www.biostars.org/p/9572863/	'Uploads in theory would only be limited by bandwidth'; >10s of TB goes through SRA support
2026-10-21	sra-pub-run-odp.s3.amazonaws.com/sra/SRR000001/SRR000001	probe: GET -r 0-511 -> 206, Content-Range: bytes 0-511/312527083 (298 MB object), real AWS headers

Unknown or withheld values (4)

$.free_tier.notes process
$.per_object_cap.notes not-extracted
$.per_object_cap.stated not-extracted
$.retention.notes not-extracted

NCBI GEO Storage

caveatanonymousno cardRwrite-gated

www.ncbi.nlm.nih.gov/geo · academic · storage/ncbi-geo · JSON · compare

Free tier	No published volume cap; raw sequence data is redirected to SRA. Notable for retention rather than size. unit mixed; metered on unstated
Price	none stated
Per-object cap	not stated
Retention	not stated
Egress	Free and anonymous.
Identity	anonymous derived from the record's prose, not from an identity field, and no signup was executed: r up to four years before release, with reviewer tokens for anonymous read-only access. per-object-cap: not stated (defer to SRA for raw reads). Retention: 'the maximum allowable limit is f
Limits	write-gated approval step (~5 business days for accession) source per-object-cap: not stated (defer to SRA for raw reads). Retention: 'the maximum allowable limit is four years' for private hold, then mandatory public release; records can be edited after release. Binding constraint is the human curation/approval step (~5 business days for accession).
Binding	write-gated the write is a queue with a person or a rule in it, so the upload rate is set by somebody else
Interface	rest, ftp rest
Tooling	GEO FTP/HTTPS plus the GEOmetadb / GEOquery R packages.
Composes with	Not a volume play. The interesting property is private-retention: records may stay private for up to four years before release, with reviewer tokens for anonymous read-only access.
Reliability	Records cannot stay private once an accession is cited in a manuscript or preprint — GEO staff release unilaterally.
Notes	No published volume cap; raw sequence data is redirected to SRA. Notable for retention rather than size.
Evidence rung	reported One page read moves it to documented.
Checked	2026-10-21 (inherited); stale on 2026-11-04

Evidence (2)

date	source	what it proved
2026-10-21	www.ncbi.nlm.nih.gov/geo/info/faq.html	'the maximum allowable limit is four years', forced release if accession cited
2026-10-21	www.ncbi.nlm.nih.gov/geo/info/submission.html	5 business days for accession, private-until-published policy

Unknown or withheld values (5)

$.free_tier.notes process
$.per_object_cap.notes not-extracted
$.per_object_cap.stated not-extracted
$.retention.notes not-extracted
$.retention.stated not-extracted