sc-eQTL_BWB

Code for the study: Single-cell eQTL analysis identifies cell type-specific regulation of gene expression in peripheral blood mononuclear cells response to lipopolysaccharide in BWB.

All the statistical analyses were performed by in-house R/Python script or published tools/packages.

1.1 Single-cell RNA-seq analysis and cell type identification

Ambient RNA and potential doublets were detected by DecontX and Scrublet, respectively, with default settings. Single-cell analysis was performed by Seurat; cell type was identified by the classical marker.

1.2 Covariates preparing for sc-eQTL mapping

Five principal component analyses (PCA) of samples were then carried out by Plink based on the LD-pruned genotypes; Ten probabilistic Estimation of Expression Residuals (PEER) in each of the cell types were estimated using the PEER software package.

1.3 sc-eQTL mapping

A linear regression model implemented by TensorQTL was used. A linear mixed model with sparse genetic relationship matrix (GRM) implemented by fastGWA was employed.

1.4 Identification of content-specific eQTL

An interaction model was employed to identify the response and TF interaction eQTLs. The mixed linear model was used to perform dynamic eQTL.

1.5 Estimation of shared signals

MASH was applied to estimate the shared eQTL signals between each pair of eQTL summary statistics.

1.6 Colocalization analysis

COLOC was used to identify the shared genetic variants between sceQTL and clinical diseases.

1.7 Replication of the identified sc-eQTL

To validate the identified eQTL, we checked their effect size in the public GTEx dataset.

1.8 Overlapping with epigenome

The publicly available dataset scATAC-seq from the steady state and stimulated by LPS was employed.