Murphy · Cell reports methods 2025 · computational tool development and secondary bioinformatic analysis · n=>300 postmortem brain ChIP-seq samples

CHAS infers cell type-specific signatures in bulk brain histone acetylation studies of neurological and psychiatric disorders.

Cited 2 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Computational method development and secondary bioinformatic analysis of postmortem tissue datasets (graded by non-clinical design analogy).

PubMed 40300607 · doi:10.1016/j.crmeth.2025.101032 · record verified 2026-08-29

What was done

The authors developed CHAS (cell type-specific histone acetylation score), a computational method to infer cell type-specific regulatory signatures from bulk brain H3K27ac profiles. They evaluated the tool by reanalyzing over 300 bulk postmortem brain H3K27ac chromatin immunoprecipitation sequencing (ChIP-seq) samples from published studies of Alzheimer's disease, Parkinson's disease, autism spectrum disorder, schizophrenia, and bipolar disorder.

What was found

CHAS detected cell type-specific biological pathways and recapitulated known disease-associated shifts in cellular composition. Across most examined disorders, genetic risk variants and epigenetic dysregulation localized to distinct cell populations; specifically, Alzheimer's disease genetic risk was enriched exclusively in microglia, whereas epigenetic alterations predominantly localized to oligodendrocyte-specific H3K27ac regions. The abstract reports no quantitative performance statistics or correlation metrics.

Why it matters

This tool allows researchers to extract cell-type-resolved epigenetic and pathway insights from existing bulk postmortem brain histone acetylation datasets without requiring single-cell epigenomic sequencing.

Limits

The abstract reports no numerical validation metrics (such as deconvolution accuracy, sensitivity, or false-discovery rates) and does not detail validation against ground-truth single-cell H3K27ac data. Analysis relies entirely on retrospective postmortem tissue datasets across five specific conditions, which may carry confounding technical and biological variability.

Cited by