Single-Cell Sequencing Reveals Lineage-Specific Dynamic Genetic Regulation of Gene Expression During Human Cardiomyocyte Differentiation

Reem Elorbany, Joshua M Popp, Katherine Rhodes, Benjamin J Strober, Kenneth Barr, Guanghao Qi, Yoav Gilad, and Alexis Battle

A workflowr project.

Preprint available online at bioRxiv

Preprocessing of Single Cell Data

• Quality control filtering of single cell data is done at code/create_seurat.R
  • The (cells x genes) matrix of UMI counts produced by this file is available online at GEO Series GSE175634 as a .mtx file (GSE175634_cell_counts.mtx.gz, cell and gene indices at GSE175634_cell_indices.tsv.gz and GSE175634_gene_indices.tsv.gz respectively)
• Normalization of the single cell counts data is done at code/normalize_seurat.R
  • The (cells x genes) matrix of corrected UMI counts produced by this file is available as a .mtx file (GSE175634_cell_counts_sctransform.mtx.gz, same cell and gene indices as the uncorrected counts)
  • The (cells x highly variable genes) matrix of pearson residuals produced by this file is available as a .tsv file (GSE175634_pearson_residuals_sctransform.mtx.gz)

Trajectory Inference

• Trajectory inference is performed at analysis/ti.ipynb

Dynamic eQTL Calling

• Pseudobulk aggregation, normalization, and computation of PC covariates for both pseudobulk and bulk data are done at code/preprocessing.R)
• Dynamic eQTL calling is done at code/eqtl_dynamic.R for linear eQTLs, code/eqtl_dynamic_nonlinear.R for nonlinear dynamic eQTLs, and multiple testing correction is done at code/mtc_dynamic.R for linear eQTLs, code/mtc_dynamic_ttest.R for nonlinear dynamic eQTLs

Cell Type Interaction eQTL Calling

• Cell type interaction eQTL calling is performed at code/eqtl_interaction.R, with multiple testing at code/mtc_interaction.R

Further Analyses

• Overlap of dynamic eQTLs with GWAS hits is done at code/gwas.R