Fig. 2

Exome-wide, gene set, and single-cell polygenic burden test of PTVs in EM cases.a Exome-wide burden analysis of PTVs, D-mis, B-mis, and synonymous variants within rare variants (MAF < 0.5%). Significance of association is displayed with the P-values from the two-sided Firth logistic regression test; errors bars indicated 95% confidence intervals (CIs) of the corresponding odds ratios. Multiple test correction P = 0.01. The graphs display the means and standard deviations. b Burden analysis of genes involved in retrograde vesicle-medicated transport, Golgi to ER. PTVs, D-mis, B-mis, and synonymous variants within rare variants are displayed. Significance of association is displayed with the P-values from the two-sided Firth logistic regression test; errors bars indicated 95% confidence intervals (CIs) of the corresponding odds ratios. Multiple test correction P = 0.01. c Uniform manifold approximation and projection (UMAP) embedding plot shows the cellular component of a scRNA-seq dataset for human embryonic eyes. RPCs retinal progenitor cells, NPCs neurogenic RPCs, MGCs Müller glial cells, HCs horizontal cells, BCs bipolar cells, ACs amacrine cells, RGCs retinal ganglion cells, RPE retinal pigment epithelia, Endothelium endothelial cells. d Burden analysis of genes highly expressed in each cell types (testing for enrichment in rare variants of the 10% most specific genes in cell type). PTVs and synonymous variants within rare variants are displayed. The sample size for EM is as follows: Ncases = 449, Ncontrols = 9,606. The black solid line represents the null (OR = 1). Each point shows the point estimate of OR from logistic regression. Bars show 95% confidence intervals (CI). e Overview of scPBS approach. scPBS takes a case-control genotype matrix and a scRNA-seq data set as input and outputs individual-cell-level polygenic burden score for association with the disease. f Per-cell rvTRS calculated by scPBS and Seurat for EM are shown in tSNE coordinates. g Per-cell types rvTRS. Boxplots show the median with interquartile range (IQR) (25%–75%); whiskers extend 1.53 the IQR. The number of cells in each cell types is shown in parentheses. Source data are provided as a Source Data file.