Which public datasets do biologists use most? And where they put their own, by preprints
70,035 of the 112,324 preprints read so far name a public dataset they used; UniProt most of all, then Protein Data Bank, Gene Expression Omnibus, GenBank. Updated daily (as of Nov 16)
Counts preprints first posted Apr 2025–Nov 2026; Nov 2026 is still being read.
UniProt is the public dataset preprints name most.
The public databases and data collections preprints say they used, ranked by how many preprints name each.
| # | Dataset | Preprints |
|---|---|---|
| 1 | UniProtProtein sequence database · uniprot.org | 8,938 |
| 2 | Protein Data Bank3D structure archive · rcsb.org | 8,392 |
| 3 | Gene Expression OmnibusGene expression data archive · ncbi.nlm.nih.gov/geo | 5,921 |
| 4 | GenBankNucleotide sequence database · ncbi.nlm.nih.gov/genbank | 4,558 |
| 5 | KEGGPathway and gene database · kegg.jp | 3,657 |
| 6 | EnsemblGenome annotation browser · ensembl.org | 3,308 |
| 7 | MSigDBGene set collection · gsea-msigdb.org/gsea/msigdb | 2,963 |
| 8 | Gene OntologyGene function vocabulary · geneontology.org | 2,573 |
| 9 | TCGACancer genomics consortium data · portal.gdc.cancer.gov | 2,559 |
| 10 | RefSeqCurated reference sequences · ncbi.nlm.nih.gov/refseq | 2,401 |
| 11 | STRINGProtein interaction networks · string-db.org | 2,274 |
| 12 | 1000 GenomesHuman genetic variation catalogue · internationalgenome.org | 2,104 |
| 13 | Sequence Read ArchiveRaw sequencing read archive · ncbi.nlm.nih.gov/sra | 2,034 |
| 14 | UK BiobankLarge UK health cohort · ukbiobank.ac.uk | 1,988 |
| 15 | GENCODEGene annotation reference · gencodegenes.org | 1,935 |
| 16 | GTExGene expression across tissues · gtexportal.org | 1,784 |
| 17 | ReactomeCurated pathway database · reactome.org | 1,541 |
| 18 | ENCODERegulatory element encyclopedia · encodeproject.org | 1,460 |
| 19 | SILVARibosomal RNA database · arb-silva.de | 1,360 |
| 20 | gnomADPopulation variant frequencies · gnomad.broadinstitute.org | 1,357 |
| 21 | AlphaFold Protein Structure DatabasePredicted protein structures · alphafold.ebi.ac.uk | 1,208 |
| 22 | UCSC Genome BrowserGenome browser and tracks · genome.ucsc.edu | 1,192 |
| 23 | Human Protein AtlasProtein expression atlas · proteinatlas.org | 1,010 |
| 24 | PfamProtein family database · ebi.ac.uk/interpro/entry/pfam | 982 |
| 25 | ClinVarClinical variant interpretations · ncbi.nlm.nih.gov/clinvar | 954 |
| 26 | NCBI nrNon-redundant protein database · blast.ncbi.nlm.nih.gov | 949 |
| 27 | Genome Taxonomy DatabaseReference databases | 901 |
| 28 | JASPARReference databases | 779 |
| 29 | GWAS CatalogPublished trait associations · ebi.ac.uk/gwas | 770 |
| 30 | Allen Mouse Brain AtlasMouse brain expression atlas · mouse.brain-map.org | 735 |
The leaders in each kind of dataset
Reference databases
| Preprints | |
|---|---|
| UniProt | 8,938 |
| Protein sequence database · uniprot.org | |
| Protein Data Bank | 8,392 |
| 3D structure archive · rcsb.org | |
| Gene Expression Omnibus | 5,921 |
| Gene expression data archive · ncbi.nlm.nih.gov/geo | |
Consortium data
| Preprints | |
|---|---|
| TCGA | 2,559 |
| Cancer genomics consortium data · portal.gdc.cancer.gov | |
| 1000 Genomes | 2,104 |
| Human genetic variation catalogue · internationalgenome.org | |
| GTEx | 1,784 |
| Gene expression across tissues · gtexportal.org | |
Biobanks and cohorts
| Preprints | |
|---|---|
| UK Biobank | 1,988 |
| Large UK health cohort · ukbiobank.ac.uk | |
| Human Connectome Project | 699 |
| All of Us | 457 |
| US health research cohort · researchallofus.org | |
Brain and cell atlases
| Preprints | |
|---|---|
| Human Protein Atlas | 1,010 |
| Protein expression atlas · proteinatlas.org | |
| Allen Mouse Brain Atlas | 735 |
| Mouse brain expression atlas · mouse.brain-map.org | |
| Allen Brain Atlas | 512 |
| Brain gene expression atlas · brain-map.org | |
Chemical and drug data
| Preprints | |
|---|---|
| PubChem | 711 |
| Chemical compound database · pubchem.ncbi.nlm.nih.gov | |
| ChEMBL | 396 |
| Bioactive molecule database · ebi.ac.uk/chembl | |
| DrugBank | 349 |
| Drug and target database · go.drugbank.com | |
Preprints that use AI models lean on protein data.
AlphaFold Protein Structure Database, Protein Data Bank and UniProt are each named 2.5 to 4.4 times as often in preprints that use an AI model as in all preprints read.
- AlphaFold Protein Structure Database 772 preprintsPredicted protein structures · alphafold.ebi.ac.uk4.4×
- Protein Data Bank 4,215 preprints3D structure archive · rcsb.org3.4×
- UniProt 3,337 preprintsProtein sequence database · uniprot.org2.5×
- gnomAD 506 preprintsPopulation variant frequencies · gnomad.broadinstitute.org2.5×
Zenodo is where labs deposit most; for single-cell and spatial data, it's GEO.
35,955 preprints say where they put their data or code.
- ZenodoOpen repository for any research · zenodo.org6,391
- GEONCBI gene expression archive · ncbi.nlm.nih.gov/geo5,835
- GitHubCode hosting platform · github.com4,702
- SRANCBI raw sequencing reads archive · ncbi.nlm.nih.gov/sra3,756
- PRIDEProteomics data repository · ebi.ac.uk/pride2,020
- PDBProtein structure archive · wwpdb.org2,015
- GEONCBI gene expression archive · ncbi.nlm.nih.gov/geo2,237
- ZenodoOpen repository for any research · zenodo.org460
- SRANCBI raw sequencing reads archive · ncbi.nlm.nih.gov/sra304
- PRIDEProteomics data repository · ebi.ac.uk/pride154
- EGAControlled-access genetic data archive · ega-archive.org124
- ArrayExpressFunctional genomics data archive · ebi.ac.uk/biostudies/arrayexpress123
How does Tnuki know so much?
- Preprints analyzed dailyNew bioRxiv and medRxiv preprints are analyzed daily.
Whole paper readWe assess the full text, tables, and supplementary files.- Only real uses countedWe read to make sure the lab actually used the product, not just cited.
How we count
An AI model reads each bioRxiv and medRxiv preprint and lists what its authors used — instruments, kits, services, software, reagents — and how: used in this work, someone else's data, or only mentioned. A preprint counts here once, when the model says the lab used it, if it was first posted in Apr 2025 or later. Names are merged by a second model (versions and spellings into one tool, brands under their owner). We show counts and shares, and the titles of the latest preprints, each linking to the preprint on bioRxiv or medRxiv; we show no other text from a preprint and no author names. Updated daily, most recently on Nov 16, 2026.