Fig. 4

Evaluations of scCorr performance in gene-gene correlations among CD4+ T cells in dataset 1. Gene-gene relationships were quantified by Pearson correlation. In A and B, gene-gene correlation was separately performed in nonclustering single cells (X-axis) and in scCorr-partitioned cell clusters (K = 100) (Y-axis). Only the top 100 significantly correlated gene pairs are shown. Blue dots indicate agreement, and red dots indicate opposite directions of gene-gene correlation between the two methods. Correlated genes are shown as –log₁₀ P- (A) and R- (B) values. scCorr detected a significant correlation between MAPKAPK2 and ATF4 genes (C), while the conventional noncluster single-cell-based method showed no significant correlation between the two genes (D). Gene-gene correlation varied in different numbers of clusters