Datasets:
File size: 1,748 Bytes
4545e2b e31d86a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 | ---
license: mit
task_categories:
- other
tags:
- transcriptomics
- microarray
- illumina
- whole-blood
- population-cohort
---
# Data Card: SHIP-TREND Whole-Blood Microarray Dataset (GSE36382)
## Summary
Expression + feature metadata for SHIP-TREND,
a population-based whole-blood gene expression dataset from the SHIP-TREND (Study of Health in Pomerania).
991 samples, all self-reported as White European ethnicity from North-Eastern Germany.
This is used in the SepsisStratifier work as a healthy control set.
## Source accession
| Accession | Platform | N (samples) | Retrieval source |
|---|---|---|---|
| [GSE36382](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE36382) | Illumina HumanHT-12 V3.0 (GPL6947) | 991 | NCBI GEO, via `GEOquery::getGEO()` |
## Files
- `feature_metadata.parquet` — one row per probe (`IlluminaID`): annotation from `illuminaHumanv3.db` (`SYMBOL`, `ENTREZID`, `ENSEMBL`, `GENENAME`, `UNIPROT`) plus reannotation QC fields from the package's internal `ExtraInfo` table (`ProbeQuality`, `CodingZone`, `GenomicLocation`, `SecondMatches`, `OtherGenomicMatches`, `RepeatMask`, `OverlappingSNP`, `ProbeSequence`) — same annotation pipeline used for DILGOM, since both are on the v3 platform.
- `expression/` — partitioned parquet dataset (by `accession`, for schema consistency with other datasets in this collection), long format: `IlluminaID`, `sample_id`, `accession`, `value`.
- **No `sample_metadata.parquet` for this dataset**. Phenotype data is restricted and only available upon valid request.
## Expression value processing
Per the accession description, expression values are **quantile normalized and log2-transformed**.
## Provenance / reproducibility
See [pull_GSE26282.R] |