Opposite correlation coefficient using corrplot and stat_cor()

Hi corrplot team,

Thanks for such a great package to visualize the correlation matrix. Recently, I am using this package to analyze my single-cell data. However, I found the results I got were opposite to that of stat_cor().

I am sure I used the same object and the same cluster. But when I made the single cell data matrix and used the following code for the significant filter.

genes <- c("TNFRSF4", "TNFRSF18", "IL2RA")
C3 <- as.data.frame(Treg3@assays[["RNA"]]@data[genes, ])
C31 <- as.data.frame(t(C3))
C32<-cor(C31)

cor.mtest <- function(mat, ...) {
    mat <- as.matrix(mat)
    n <- ncol(mat)
    p.mat<- matrix(NA, n, n)
    diag(p.mat) <- 0
    for (i in 1:(n - 1)) {
        for (j in (i + 1):n) {
            tmp <- cor.test(mat[, i], mat[, j], ...)
            p.mat[i, j] <- p.mat[j, i] <- tmp$p.value
        }
    }
  colnames(p.mat) <- rownames(p.mat) <- colnames(mat)
  p.mat
}
# matrix of the p-value of the correlation
p.mat <- cor.mtest(C32)

corrplot(C32, type="upper", order="hclust", tl.col="black", tl.srt=45, title="eTreg",
         p.mat = p.mat, sig.level = 0.05, insig = "blank", mar=c(0,0,1,0))

I got TNFRSF4 and TNFRSF18 are positively correlated but negatively correlated with IL2RA (see the pale pink dots).

But when I used the Seurat embedded FeatureScatter() with stat_cor(), I found all three are positively correlated (R is person correlation coefficient, while p is p-value)

FeatureScatter( object = Treg3, feature1  = "IL2RA", feature2 = "TNFRSF4", cols = "khaki")+
stat_cor(method = "pearson")

I also checked my data with before and after treatment and found they are definitely positively correlated.

Can you help me fix it? Whether it is because of the cor.mtest() function?

Thanks, Yale

taiyun / corrplot

Opposite correlation coefficient using corrplot and stat_cor() #163