# EoE Phase-1 Dataset Prioritization Rationale

**Inventory:** 56 EoE-relevant datasets (57 retrieved; GSE168883 excluded — environmental
enteropathy, keyword co-mention only). Sources: GEO, CELLxGENE, PRIDE, ArrayExpress.

## Tiering rubric

Datasets were scored on **relevance** (human, esophageal/EoE, clear disease contrast),
**tractability** (open access, processable matrices), and **analytical role**:

- **Tier 1 — standardized DE + cross-study meta-analysis (12 datasets).**
  Human *in-vivo* esophageal biopsy expression (bulk RNA-seq or microarray) with an
  interpretable EoE-vs-control (or treatment-response) design and adequate sample size
  (n≥12). These feed Steps 3–5 (QC → per-study DE → meta-analysis).
- **Tier 2 — single-cell (6 datasets).** Esophageal/gut single-cell resources for
  cell-population-resolved analysis (Step 6). Anchored by the GSE126250 T-cell atlas, the
  GSE218607 EoE epithelial landscape, and the CELLxGENE oesophagus atlas (~1.6M cells).
- **Tier 3 — complementary omics & mechanistic (38 datasets).** Methylation, miRNA/ncRNA
  (incl. plasma-EV biomarkers), proteomics, blood/peripheral expression, and *in-vitro*/mouse
  functional-genomics studies. These provide orthogonal validation (Step 8) rather than primary DE.

## Inclusion / exclusion notes

- *In-vitro* organoid/cell-line and mouse-model expression series (e.g. GSE271128 OVA mouse,
  GSE256022 amniotic-fluid organoid, GSE75142 eosinophil activation, GSE264591) are placed in
  Tier 3 — informative for mechanism but not comparable in a human case-control meta-analysis.
- Genotype/enhancer functional-genomics (GSE232341/GSE232337) are Tier 3 (not expression DE).
- Blood/plasma expression (GSE173895, GSE184182, E-MTAB-16277) is Tier 3 — systemic signal,
  distinct compartment from esophageal tissue.

## Tier 1 datasets (primary DE + meta-analysis)

| Accession | n | Assay | Rationale |
|---|---|---|---|
| GSE250595 | 68 | bulk RNA-seq | Familial EoE cohort, esophageal biopsy (large n) |
| GSE148381 | 51 | bulk RNA-seq | Open case-control esophageal expression, adequate n |
| GSE278888 | 28 | bulk RNA-seq | Open case-control esophageal expression, adequate n |
| GSE197702 | 27 | bulk RNA-seq | Open case-control esophageal expression, adequate n |
| GSE228083 | 25 | microarray | Open case-control esophageal expression, adequate n |
| GSE303169 | 25 | bulk RNA-seq | PPI-responder EoE vs control biopsies; treatment context |
| GSE234973 | 23 | bulk RNA-seq | Open case-control esophageal expression, adequate n |
| GSE156651 | 20 | bulk RNA-seq | IFN signature, adult+pediatric EoE vs control biopsy |
| GSE246323 | 20 | bulk RNA-seq | Acute mucosal response to food injection, EoE biopsy |
| GSE113341 | 20 | bulk RNA-seq | Open case-control esophageal expression, adequate n |
| GSE41687 | 18 | bulk RNA-seq | Open case-control esophageal expression, adequate n |
| GSE58640 | 16 | bulk RNA-seq | EoE transcriptome by RNA-seq (foundational cohort) |

## Tier 2 datasets (single-cell)

| Accession | cells/samples | Rationale |
|---|---|---|
| CxG:f11cb29c (Gut Cell Atlas / Oesophagus) | 1596200 | Single-cell esophageal — cell-population resolution |
| GSE126250 | 1089 | Single-cell esophageal — cell-population resolution |
| GSE218607 | 31 | EoE epithelial single-cell landscape |
| GSE249276 | 12 | Single-cell esophageal — cell-population resolution |
| GSE201153 | 10 | Single-cell esophageal — cell-population resolution |
| GSE227593 | 8 | Single-cell esophageal — cell-population resolution |

## Tier 3 (complementary — 38 datasets)

By assay: Other (1), bulk RNA-seq (15), methylation (2), miRNA/ncRNA (7), microarray (5), proteomics (5), single-cell RNA-seq (3).
Key resources: GSE186888 (methylation), GSE275710 (plasma-EV miRNA biomarkers),
PXD062507 / PXD052250 / PXD031509 (proteomics: epithelial barrier, Periostin, IL-20 signaling).

*Full machine-readable tiering in `dataset_tiers.csv`.*