Griswold et al. Miami cohort RNA-seq expression matrices for "Sex and insulin resistance biomarker modelling in a new large‑scale Alzheimer's disease transcriptomic resource"

This dataset contains bulk RNA‑seq expression matrices and sample‑level phenotype data from the Miami Alzheimer’s disease cohort used in the study “Sex and insulin resistance biomarker modelling in a new large‑scale Alzheimer’s disease transcriptomic resource”. FASTQ files were aligned with STAR using two strategies: a standard gene‑level alignment to the reference genome and an alternative alignment restricted to genes represented on the Affymetrix GeneTitan array to enable cross‑platform replication. Gene‑level counts were obtained with featureCounts, low‑count genes were removed, and filtered matrices were size‑factor normalised with DESeq2 and log2‑transformed. Combat‑Seq was applied to filtered counts using total count deciles as pseudo‑batches, and additional neutrophil‑adjusted matrices were generated by correcting for estimated neutrophil fractions derived from immune cell deconvolution. The dataset includes two filtered raw count matrices, four corresponding processed matrices (including neutrophil‑adjusted versions), and a phenotype table linking each sample to diagnosis and key covariates, with consistent sample identifiers across all files. Please note that there are other systematic differences in the case vs. control data that have not been fully explored through ComBat correction.