About & Data
Pseudogenomes for 1135 Arabidopsis thaliana strains were generated by combining reference and variant calls including deletions, with uncalled sites represented as Ns.
Data & Downloads
The data powering SNPst☆r is published as an
Annotated Research Context (ARC)
— Arabidopsis thaliana GWAS & Proteotype ARC (export 2026-06-05,
38.2 GB total).
Recommended: download the
complete ARC
(all datasets, ISA metadata and documentation in one bundle) from the repository:
Open the ARC repository »
Or download individual collections directly. All collections are gzipped JSON Lines (one document per line) except the SNP catalog, which is gzipped BSON (mongodump) and is restored with mongorestore, and the derived CDS/InterPro table, which is gzipped CSV. See the README and METADATA documents for details. The two derived Cys → Trp variant lists are plain CSV and open directly in a spreadsheet.
Climate-factor GWAS (study)
| Dataset |
Documents |
Size |
Format |
sha256 |
|
Climate-factor GWAS (raw p-values)
Raw GWAS p-values per SNP across ~196 climate factors. Climate-factor keys join to climate_factors_upd for human-readable descriptions.
|
3,042,994 |
250 MB |
jsonl.gz |
ce9831a2c6c5… |
↓ Download
|
Climate-factor GWAS (enriched p<1 subset)
Enriched subset of GWAS hits with p<1; climate-factor sub-documents carry pval, description and link inline.
|
185,450 |
12 MB |
jsonl.gz |
ccc12893ee6c… |
↓ Download
|
Climate-factor descriptions
Canonical descriptions of the ~196 climate factors (ID, Description, Link, Source, Units).
|
196 |
6.2 KB |
jsonl.gz |
c2e256803d60… |
↓ Download
|
Climate-factor key → human-id join
Join table mapping machine GWAS keys (actualID) to human-readable ID, Description, Link, Source and Units.
|
195 |
7.2 KB |
jsonl.gz |
97ba41c0969a… |
↓ Download
|
Ferrero-Serrano Á, Assmann SM (2019). Nat. Ecol. Evol. 3, 274-285. doi:10.1038/s41559-018-0754-5
AraGWAS Catalog (study)
| Dataset |
Documents |
Size |
Format |
sha256 |
|
AraGWAS significant hits
Curated significant SNP×phenotype associations from the AraGWAS Catalog (snp_id_c, chr, position, score, maf).
|
36,399 |
1.3 MB |
jsonl.gz |
ef77e7a85371… |
↓ Download
|
Togninalli M et al. (2018). Nucleic Acids Res. 46(D1), D1150-D1156. doi:10.1093/nar/gkx954
SNP, haplotype, proteotype & structure assays
| Dataset |
Documents |
Size |
Format |
sha256 |
|
SNP catalog
Full SNP catalog used by the website (transcript_id, snp_id, snp_dp, snp_location, ...). Gzipped BSON; restore with mongorestore. Large file (≈11 GB) — the complete ARC is the recommended way to obtain it.
|
20,010,332 |
11 GB |
bson.gz |
a0405c535f8e… |
↓ Download
|
Haplotypes
Per-transcript haplotypes across accessions (haplotype_id, haplotype_number, num_snps, num_accessions).
|
6,348,060 |
685 MB |
jsonl.gz |
b3ed98041a72… |
↓ Download
|
Transcript regions
Transcript genomic regions (transcript_id, chr, strand, start, end).
|
52,060 |
219 MB |
jsonl.gz |
bf148c32fad1… |
↓ Download
|
Proteotype haplotypes
Protein-level haplotype groups (transcript_id, protein_haplotype_id, protein_haplotype_number, num_accessions, accessions).
|
3,495,265 |
422 MB |
jsonl.gz |
488ebe820857… |
↓ Download
|
Proteotype structures
AlphaFold/PDB structural data per proteotype (transcript_id, proteotype_id, pdb_structure, filename, file_path). Very large file (≈26 GB) — the complete ARC is the recommended way to obtain it.
|
356,807 |
26 GB |
jsonl.gz |
1fd5cd88c2a9… |
↓ Download
|
ThermoMPNN stability predictions
ThermoMPNN ΔΔG stability predictions per protein (uniprot_id, protein_length, predictions, source_file, model_version).
|
41,597 |
1.3 GB |
jsonl.gz |
abba4ef25900… |
↓ Download
|
CDS variants with InterPro / domain annotation (derived)
One row per non-reference CDS SNP (snp_location = CDS, snp_is_ref = no): transcript_id, snp_id_c, aa_pos, and the InterPro entries and InterProScan signatures hit at that amino-acid position. Derived from snp_29_11_2024 by matching aa_pos against protein_domains_interproscan.regions_protein in regions_01_12_2025 — the same rule the SNP table applies. Multiple hits are separated by ';' and IDs stay aligned with their descriptions. Each row also carries domain_status, a reason code saying what a blank annotation means, and impact_annotation, the stored SnpEff putative-impact class (HIGH / MODERATE / LOW / MODIFIER) read straight off the SNP document - the same value the SNP table shows in its 'Impact annotation' column.
|
3,822,870 |
31 MB |
csv.gz |
da432aabe878… |
↓ Download
|
Cys → Trp missense variants (derived)
Every record shown as "Cys -> Trp" in the SNP table's "Amino acid substitution" column, across all transcripts (all 1135 accessions, default cutoffs). One row per transcript isoform, with the SNP table's own columns: SNP_ID, base frequencies, amino-acid and mRNA position, SNP location and effect, substitution, chemical-property effect, ThermoMPNN ΔΔG, impact annotation, InterPro ID and description, AraGWAS top hits and climate-factor GWAS hits. Derived from snp_29_11_2024 by selecting snp_aa_seq = "Cys -> Trp" with snp_aa_change = nonsynonymous.
|
2,500 |
451 KB |
csv |
0523bfe10ff6… |
↓ Download
|
Cys → Trp missense variants, splice-region subset (derived)
Companion to the table above: records stored as "Cys -> Trp" in snp_29_11_2024 whose snp_aa_change is missense_variant&splice_region. The SNP table leaves the "Amino acid substitution" cell empty for these, so they are not reachable by filtering that column. Same columns, plus a Note column.
|
22 |
7.5 KB |
csv |
1b3172c70724… |
↓ Download
|
Derived from the 1001 Genomes panel — 1001 Genomes Consortium (2016). Cell 166, 481-491. doi:10.1016/j.cell.2016.05.063
Cross-assay lookup tables
| Dataset |
Documents |
Size |
Format |
sha256 |
|
Accession data
Accession metadata (acc_id, name, cs_number, country, group).
|
1,135 |
18.5 KB |
jsonl.gz |
5d58eb9473b8… |
↓ Download
|
Accession locations
Geolocations per accession (accession_id, name, lat, lng, country).
|
7,406 |
87 KB |
jsonl.gz |
9aaaed3d36aa… |
↓ Download
|
Transcript → UniProt mapping
Mapping between transcripts and UniProt entries (uniprot_id, isoform_id, gene_name).
|
48,321 |
778 KB |
jsonl.gz |
0ed018b664b0… |
↓ Download
|
Accession metadata derived from the 1001 Genomes panel (Cell 2016, 166, 481-491).
Programmatic data access (API)
The full Arabidopsis thaliana / 1001 Genomes GWAS dataset can be accessed programmatically through the public AraGWAS REST API:
https://aragwas.1001genomes.org/api/.
If you use this resource, please cite snpStar.
The Collaborative Research Centre CRC 1664 is funded by Deutsche Forschungsgemeinschaft (DFG).
Data release, coverage and known limitations
This section states what the data behind SNPStar is, where it is
incomplete, and which questions this release cannot answer. It is
written for people who intend to use the numbers, not only read
them. Where a figure is a sample rather than a full count, it says
so.
Release
The portal serves SNPStar data release 2026-09, covering the
1,135 Arabidopsis thaliana accessions of the 1001
Genomes panel, built from the GMI-MPI v3.1 pseudogenomes.
The SNP table holds 20,104,947 rows. Frequencies shown in the
SNP table are folded minor-allele frequencies over all 1,135
accessions, computed per position at query time; the table
itself applies no frequency cutoff, so any threshold you see quoted
below belongs to a specific analysis layer and not to the table.
Credits
-
Pseudogenomes (download FASTA files from the 1001 Genomes Data Center) were generated by Alexander Platzer.
-
GWAS analysis was performed by Francesco Pezzini with the software GEMMA .
Every climate factor was used as a phenotype and for every climate factor, one analysis run was carried out. In a data preprocessing step ecotypes related to missing values were filtered out from the phenotype vector. Minor allele frequency (MAF) and linkage disequilibrium (LD) thresholds in GEMMA were set to 1% and 0.8 respectively. This means only SNPs with a MAF of 1% or high were analysed. The obtained P-values were corrected for multiple comparisons using the Benjamini-Yekutieli method. The climate factor data was taken from the paper by Ferrero-Serrano and Assmann.
-
The de-novo GWAS are computed with TASSEL 5.0 (https://tassel.bitbucket.io/).
-
The analysis of the SNP data was performed by Lasse Feldhahn .
-
This tool was built on the basis of the 1001 Genomes Tool Pseudogenomes ( Github page) developed by Joffrey Fitz.
-
The additional web app and database development was done by Benjamin Schmidt. The webapp uses bootstrap, selectize, jquery, popper the calipho-sib feature-viewer and sequence-viewer and the webix datatable .
-
Additional features and enhancements, including the public API and API integration, were developed by Muhammad Tahir.
References
