CODEWITHSUNDEEP
rheatmap
JORIS
JORIS

Reputation: 59

R- How to plot a heatmap that shows significant correlations?

I have a big heatmap. I want to plot it more clearly by showing only variables with a spearman correlation greater 0.5 and lower -0.5. This results in a correlation matrix with blank spaces for greater correlations. Therefore the heatmap has blank spaces as well. I want a heatmap that shows correlations in the descripted way (with * when p.value below 0.05) but blank spaces are not colored white but in the appropiate color.

This code generates a heatmap with blank spaces, how to put the right correlations back into the matrix, that feeds the heatmap? How to avoid blank spaces in the heatmap?

library(tidyverse)
library(ggplot2)  
library(gplots)

#data.frame
df <- data.frame(var.1 = c('gucci','prada','lacoste','pseudo','gucci','prada','lacoste','pseudo'),var.2 = c('carat.1','carat.1','carat.1','carat.1','carat.2','carat.2','carat.2','carat.2'),spearman = c(-0.5,0.5,-1,0.02,-0.5,0.5,-1,0.02),p.value = c(0.05,0.5,1,0.03,0.05,0.5,1,0.03))

#only spearman greater 0.5 and lower -0.5
df.spe <- df[df$spearman >= 0.5 | df$spearman <= -0.5,]

#heatmap matrix
sub <- df.spe %>%  dplyr::select(.,everything(),-starts_with('p.adj.log')) %>% pivot_wider(.,names_from = 'var.1',values_from = 'estimate') %>%
    column_to_rownames(.,'var.2') %>% data.matrix(.)

sub2 <- df.spe  %>% dplyr::select(.,everything(),-starts_with('estimate')) %>% pivot_wider(.,names_from = 'var.1',values_from = 'p.adj.log') %>%
    column_to_rownames(.,'var.2') %>% data.matrix(.)
sub2 <- ifelse(sub2 >= -log10(0.05),'*','')

#heatmap
heatmap.2(sub,cexRow = .35,cexCol = .35,trace = 'none',key.title = 'Spearman correlation',col = my_palette,keysize = .5,key.par = list(cex=.4) ,notecol = 'black',srtCol = 30,cellnote = sub2)

Thanks ;)

Upvotes: 0

Views: 2579

Answers (1)

user12728748
user12728748

Reputation: 8506

You have not provided reproducible data and it is not clear to me what exactly you are trying to achieve.

If I understand correctly, you seem to generate "blank spaces", yet don't want that, and instead just want to highlight significant cases with a star symbol.

Have a look at my reproducible example below and see if that helps you achieve what you want.

Going from there, you could for example set lower.tri (and diag, if you like) of the correlation and p-value matrices to NA and not cluster rows and columns of the heatmap if you want to just keep triangular matrix and blank out the rest.

library(Hmisc)         # for correlations and p-values
library(RColorBrewer)  # for color palette
library(gplots)

# define a color palette
my_palette <- colorRampPalette(rev(brewer.pal(n = 7, name = "RdYlBu")))(100)

# generate reproducible matrix to calculate correlations on
set.seed(23)
mat <- matrix(stats::runif(100, 3, 14), nrow = 10, ncol = 10,
              dimnames = list(paste0("Brand", 1:10), paste0("Val", 1:10)))
modmat <- sample(1:10, 4)
mat[modmat, 1:5] <- mat[modmat,1:5] + stats::runif(20, 4, 6)
mat[modmat, 6:10] <- 14-mat[modmat, 1:5] # for negative correlation

# calculate spearman-rank correlation
cor.mat <- rcorr(t(mat), type = "spearman")

# only keep comparisons that have some abs. correlation >= .5 (optional)
keep <- rownames(cor.mat$r)[rowSums(abs(cor.mat$r)>=0.5) > 1]
cor.mat <- lapply(cor.mat, function(x) x[keep, keep])

# set diagonal to 1, since it is not interesting and should not be marked
diag(cor.mat$P) <- 1

# plot heatmap and mark cells with abs(r) >= .5 and p < 0.05
heatmap.2(cor.mat$r, 
          # cexRow = .35, cexCol = .35, 
          trace = 'none',
          key.title = 'Spearman correlation',
          # keysize = .5, key.par = list(cex=.4), 
          notecol = 'black', srtCol = 30, 
          col = my_palette,
          cellnote = ifelse(cor.mat$P < 0.05 & abs(cor.mat$r)>=0.5, "*", ""))

Created on 2021-02-22 by the reprex package (v1.0.0)

Upvotes: 2

Related Questions