The team found that snATAC-seq alone had limited power to define robust subclusters within major cell types. Because the chromatin accessibility measurements were sparse—consistent with a relatively shallow sequencing depth of roughly 30,000 reads per nucleus—the researchers could not recover strong, statistically significant, cell-type-specific snATAC-seq peaks that directly tracked clinical phenotypes in all three groups.

In contrast, when chromatin accessibility was evaluated across the transcriptomically defined subclusters, a clearer signal emerged. Tens to hundreds of peaks differed significantly between RNA-derived cell subclusters, using differential accessibility testing (including an FDR-adjusted threshold). This pattern aligns with prior work suggesting that regulatory differences are more detectable when anchored to expression-defined cell states.

Motif enrichment analysis of these differential peaks highlighted candidate transcription-factor binding sites overrepresented in each cluster. The researchers then cross-referenced these enriched motifs with transcription-factor expression in the corresponding cell types, narrowing the list to a smaller set of factors potentially shaping each cluster’s identity.

Several putative drivers stood out: HSF1 and HSF2 for the SERPINH1+ astrocyte cluster, EGR1 for the GPNMB+ microglia cluster, and PATZ1 and ELF5 for the WIF1+ COX8A+ astrocyte cluster. Across oligodendrote-derived factors, candidate drivers included WT1 in astrocyte factor 8 and RORA and ZNF449 in microglia factor 14, among others.

Importantly, some regulatory candidates appeared shared across multiple states rather than acting uniquely in one lineage. For example, SREBF1 emerged in both an RORB+ CUX2+ CCDC68+ L23 glutamatergic cluster and in WIF1+ USH1C+ AC astrocytes, and it also surfaced in an oligodendrocyte factor, suggesting coordinated regulation across cell programs.

To connect accessibility patterns to genetic risk, the study evaluated chromatin occupancy around ~70 AD-associated GWAS loci. Most loci showed subtle cell-type effects, while a notable region-specific signature appeared in GABAergic neurons. Across the three population groups, only a locus containing KANSL1 showed consistent differences in occupancy.

Overall, the work supports a model in which transcriptome-defined cell states provide the scaffolding needed to detect chromatin regulatory differences, and where shared transcription-factor programs may help unify AD-associated cell signatures across populations.

Subject of Research: Alzheimer’s disease cell-type signatures and regulatory drivers using matched snRNA-seq and snATAC-seq
Article Title: Cell-type signatures of Alzheimer’s disease shared across population groups
Article References: Luquez, T., Algoo, J., Chiu, R. et al. Nature (2026). https://doi.org/10.1038/s41586-026-10793-0
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41586-026-10793-0
Keywords: single-nucleus ATAC-seq, Alzheimer’s disease, GWAS loci, transcription factors, chromatin accessibility, cell-type signatures, astrocytes, microglia, oligodendrocytes