← All research data

SeqDesk · original analysis

Surging data volume vs lagging contextual metadata

One chart for the FAIR argument: how many samples the ENA takes in each year, against the share of them that carry a valid MIxS environment field (which applies mainly to environmental samples).

5.0M
samples · 2025
866×
more samples than 2008
12%
carry a MIxS environment
0.9%
did in 2008
1.0K10K100K1.0M10M0%25%50%75%100%2008201020122014201620182020202220242025sample submission year (first public)12345
Samples deposited / year (log)Share with valid MIxS environment context (right axis)

Left axis is logarithmic — the volume line climbs across four orders of magnitude while the context line never leaves the floor · hover to read both.

The copper line is how many samples the ENA took in each year — from 5.7K in 2008 to 5.0M in 2025, a 866× rise (note the logarithmic axis). The orange line is the share of those samples that carry a valid MIxS environment, and it stays low, ending at just 12%. Sequence production rose by orders of magnitude over this period, while the share carrying this standardized contextual field — one signal among many that support findability and reuse — rose only modestly. A MIxS environment applies mainly to environmental samples, so read this as one field’s trend, not a verdict on every record. Each year is the whole cohort that went public that year — read the long arc, not single-year wobbles.

Timeline — the standards behind the curve
  • 12008-05MIGS standard — The Genomic Standards Consortium's “Minimum Information about a Genome Sequence” (Field et al., Nat Biotechnol) — the first community checklist for genome & metagenome context.
  • 22011-05MIxS specification — MIMARKS & MIxS (Yilmaz et al., Nat Biotechnol) define the minimum contextual fields — geography, environment, collection — this analysis scores against.
  • 32016-03-15FAIR Principles — Wilkinson et al. (Sci Data) coin the FAIR principles, putting rich, machine-readable metadata at the centre of data stewardship.
  • 42020-03-11COVID-19 sequencing surge — The WHO declared a pandemic and SARS-CoV-2 submissions exploded; these viral-genome batches dominate recent cohorts, so post-2020 swings reflect what was submitted as much as practice.
  • 52023-03-03INSDC spatiotemporal standards — INSDC tightened collection-date & lat/lon reporting and standardised the missing-value vocabulary — the very placeholders the metadata tracker counts as empty.
How this is tracked
CohortEvery public sample in the ENA, grouped by the year it was first made public (partial current year excluded)
VolumeWhole-archive sample counts per first-public year — exact, not a sample
“MIxS environment”A valid, non-placeholder environment context (broad-scale / local / medium), the rarest contextual field
Update cadenceWeekly automated re-count · latest snapshot 2026-06-29
CompanionPairs with sequencing metadata completeness (the same data, field by field) and reporting-standard adoption
Methodscripts/check-metadata-completeness.mjs
Samples vs MIxS environment, by submission year
YearSamplesWith a MIxS environment
20255.0M11.9%
20245.0M11.7%
20236.6M12.8%
20227.1M9.2%
20216.8M8.2%
20203.2M11.7%
20192.6M14.1%
20181.9M11.9%
20171.6M10.6%
20161.2M13.1%
2015931K9.6%
2014544K11.8%
2013434K4.2%
2012467K1.4%
2011238K0.6%
2010154K0.3%
200932K0.3%
20085.7K0.9%

The headline of the metadata story: produced volume against described context, re-counted weekly straight from the EMBL-EBI / European Nucleotide Archive (Portal API /count, whole-archive). Only aggregate percentages and counts are published. · back to all research data

Report errorpmu15@helmholtz-hzi.de