Genome-wide association studies (GWAS) have elucidated disease-associated loci, but interpreting results necessitates understanding the cellular context in which variants operate, because GWAS SNPs are enriched among regulatory elements which tend to act in very cell type specific manners. Integrating GWAS with functional data prioritizes disease-relevant cell types, contributing significantly to understanding disease mechanisms. Stratified LD score regression, a widely-implicated method for identifying functional enrichment from GWAS allows estimation of how much a set of genomic regions contributes to the overall heritability of a phenotype. A current gap in the literature is that although most GWAS loci are non-coding, research on integrating GWAS with scATAC-seq has been limited compared to transcriptomic data. Further, Sparsity of scATAC-seq data challenges cell type-specific peak identification, affecting S-LDSC accuracy. To address this, we developed a reproducible pipeline mapping genetic traits onto cell types using scATAC-seq, benchmarking for different levels of cell type peaks as input. During this process, we also found mouse-derived open chromatin profiles can serve as proxies for challenging-to-obtain human cell populations, facilitating disease heritability understanding.
In Amyotrophic Lateral Sclerosis (ALS), causal cell types enriched with GWAS heritability remain unidentified. Contrary to the assumption of motor neuron-centricity, ALS involves a network of cell types, with glial degeneration preceding motor neuron degeneration. Limited sequencing data on spinal cord, specifically on lower motor neurons contribute to this literature gap. Our comprehensive multiomic study combines ATAC and gene expression data in the human spinal cord, utilizing motor neuron enrichment strategies. This yielded 3694 cells across 19 clusters, identifying oligodendrocytes as a causal cell type in ALS pathogenesis using our pipeline. Beyond causal cell type identification, our study explores transcriptome-epigenome interactions and systematically identifies putative regulatory targets using multimodal single-cell datasets. This work advances disease mechanism understanding, providing a framework for unraveling intricate interactions across molecular levels.