Block-aware meta-analysis of animal diabetic kidney disease transcriptomes: analysis code and derived data

Analysis code and derived data for a block-aware, cross-model meta-analysis of animal diabetic kidney disease (DKD) transcriptomes (21 diabetic-versus-control contrasts; 14 independent rodent study blocks; mouse and rat; microarray and RNA-seq). A study-block-aware three-level random-effects meta-analysis identifies nine reproducible-stable genes — a tubular-redox module (down: SLC22A2/OCT2, TXN2, MSRB1; up: NFE2L1) — distinct from a genome-wide vascular–interferon axis, with a compartment-matched human DKD reference comparison. Contents: the executable R pipeline (code/) and derived data (data/) — per-contrast differential expression, meta-analysis and dependence-sensitivity results, evidence tiers with heterogeneity statistics, enrichment, figure source data, and supplementary tables SuppData_S1–S9, plus a SHA256 manifest. Primary datasets are public under their original accessions (GEO/ArrayExpress: E-MEXP-3165, GSE7253, GSE33744, GSE44375, GSE84663, GSE86300, GSE87359, GSE106841, GSE123853, GSE222776, GSE134804, GSE159059, GSE184836, GSE228960, GSE87899; human references GSE30528, GSE30529); only derived data and computed human summary statistics are redistributed. Code is MIT-licensed, data CC BY 4.0 (see LICENSE).