Spatial Fc-Engagement Programs in Lecanemab-Treated Microglia Oppose

This repository contains the complete R analysis script and all
publication figures (Figures 1–5) accompanying the manuscript "Spatial
Fc-Engagement Programs in Lecanemab-Treated Microglia Oppose
Alzheimer's Disease Transcriptomic Signatures: An Integrative Spatial
Transcriptomic Analysis."

The single integrated script
(lecanemab_spatial_integrated_analysis.R) reproduces the full
analytical pipeline: (1) ingestion of spatial transcriptomic (.h5ad)
and scRNA-seq (10x MTX) data from GEO accessions GSE297667 and
GSE297665; (2) pseudobulk differential expression across five
plaque-distance zones (0–25, 25–50, 50–100, 100–250, >250 µm) using
edgeR quasi-likelihood GLMs, followed by Reactome pathway enrichment
using fgsea; (3) microglial identity and state scoring for
compositional control; (4) pathway-level reversal analysis comparing
lecanemab-associated spatial signatures against Alzheimer's disease
signatures from four independent human bulk transcriptomic datasets
(GSE33000, GSE44770, GSE5281, GSE48350), aggregated by
inverse-variance weighted fixed-effect meta-analysis; (5)
leave-one-section-out robustness analysis of spatial module
trajectories; and (6) generation of all publication figures.

All input data are publicly available from NCBI GEO. The GSE48350
dataset required NES sign-correction (contrast coded Control vs AD in
source matrix); this is documented in the script and manuscript
Methods. Scripts are written in R (≥ 4.4.0) and depend on
Bioconductor packages (edgeR, limma, fgsea, SingleCellExperiment,
zellkonverter, GEOquery) and CRAN packages (msigdbr, mgcv, data.table,
ggplot2, patchwork). Full instructions are provided in README.md.