Selecting the highest Pearson correlation coefficient of two matrices

Question

[jivarajivaraj]

Hi,

I have two matrices of normalised gene expression values from single cell data from two time points. In rows I have the same genes but cells in each matrix is not same(different time point). In figure 2 C of this paper https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5668937/ authors say they have calculated Pearson correlation and give one Pearson correlation coefficient for each comparison as the highest correlation coefficient.

I am using many Pearson calculation like corrplot.mixed(cor(cbind(h_160,h_140)), order="hclust", tl.col="black") that gives me one Pearson correlation coefficient for each pair of cells (A matrix of Pearson correlation coefficient). So how I could select the highest Pearson correlation coefficient for these two matrices??

Answer 1

[Kevin Blighe]

The maximum would just be:

cormatrix <- cor(cbind(h_160,h_140))
max(cormatrix, na.rm=TRUE)

Note that I have 'parallellised' the cor() function in R, here (see section 2 cor function, parallellised): R functions edited for parallel processing

In addition, it looks like they are just doing a simple network plot in Figure 2C. I also use Pearson or Spearmann correlation coefficients in my tutorial, here: Network plot from expression data in R using igraph

Kevin