Skip to main content
. 2021 Mar 4;17(3):e1009080. doi: 10.1371/journal.pgen.1009080

Fig 3. Benchmark evaluation for cell-type-specific AEI detection assuming one cell type with AEI.

Fig 3

We evaluated the performance of BSCET assuming the major cell type for each SNP, i.e., cell type with the highest mean expression in scRNA-seq data [12], had cell-type-specific AEI using true cell type proportions. (A) Type I error rate and power evaluated at the cell type resolution under significance level α = 0.05 (dashed line). The number and percentage above each bar indicates number of SNPs, out of 11,300 tested SNPs, with cell-type-specific AEI of each cell type. (B) Detection power separated by the cell type and true level of AEI. The number above each bar indicates the total number of SNPs with cell-type-specific AEI within the category. The solid line indicates the overall power, i.e., across all cell types, at each level of AEI.