遇见数据集

DEGORA benchmark reproduction data: differential-expression tables and locked gold panels for four human stimulus-response programs (v1)

收藏
Zenodo2026-06-18 更新2026-06-21 收录
官方服务:

资源简介:

Reproduction data for the DEGORA study ("DEGORA: turning heterogeneous published DEG summaries into an inspectable gene-evidence atlas"). DEGORA turns published RNA-seq and microarray differential-expression gene (DEG) tables into a local, source-traceable gene-evidence database that ranks genes by repeated, directionally consistent support across independent source units. This archive holds the 50 DEG tables that the seven Table 1 benchmark configurations use, plus the 5 locked gold-marker panels used only for recall evaluation (the gold genes are not inputs to the score). The DEG tables cover four human positive-control programs: type I interferon, ER stress / unfolded protein response (tunicamycin), heat shock / HSF1, and hypoxia / HIF1. Each program is provided as an RNA-seq corpus and, except for heat shock, also as a combined RNA-seq and microarray corpus, with 2 to 20 independent source units per corpus. The files keep the data/deg/raw/... and data/studies/gold/... paths that the DEGORA configs expect, and MANIFEST.csv lists every file with its topic, source GEO accession, byte size, and SHA-256. To reproduce a benchmark, install DEGORA (https://github.com/kangk1204/DEGORA), unpack degora_reproduction_data_v1.zip at the repository root (or run reproducibility/fetch_reproduction_data.sh), then run a config such as degora run reproducibility/datasets/01_ifn_rnaseq/config.xlsx. The seven configs are in the repository under reproducibility/datasets/, and the recall of each topic's locked gold panel reproduces the matching row of Table 1. All inputs come from public repositories, mainly NCBI GEO with a few journal and data-repository supplements. Each file's original accession is recorded in MANIFEST.csv, and the paper's supplementary file on data collection and configuration describes how every table was obtained. Users should also cite the original studies behind the accessions they use. The DEGORA software is released separately under the MIT license. Two files are included: degora_reproduction_data_v1.zip (the derived differential-expression tables and the locked gold-marker panels) and degora_microarray_probe_gene_matrices_v1.zip (the per-platform probe-to-gene expression matrices that are the input to the microarray Welch derivation; see the bundled README).

提供机构:
Zenodo
创建时间:
2026-06-18
二维码
社区交流群
二维码
科研交流群
商业服务