R/CorrPlot.R

In cosmonet-package/COSMONET: Screening-network Cox methods for survival analysis

# This script contains the utility function corrPlot() to perform the correlation analysis between PIs.
#
## Input: df1 and df2 -- two data frames containing the survival and prognostic information 
#         groupRisk1, groupRisk2 -- character giving the risk information
#         method -- method for computing correlation coefficient. Allowed values are one of "pearson", "kendall", or "spearman".
#         name1 and name1 -- character specifying labels of x and y coordinates
## Output: Create a scatter plot inclusing the correlation coefficient, p-value and linear regression line

CorrPlot <- function(df1, df2, groupRisk1, groupRisk2, method=c("pearson", "kendall", "spearman"), name1, name2, set){
  
  df <- merge(df1, df2, by.x = 1, by.y = 1)
  df$condition <- ""
  df$condition[df[,groupRisk1]==2 & df[,groupRisk2]==2] <- "High Risk"
  df$condition[df[,groupRisk1]==1 & df[,groupRisk2]==1] <- "Low Risk"
  df$condition[df[,groupRisk1]==1 & df[,groupRisk2]==2 | df[,groupRisk1]==2 & df[,groupRisk2]==1] <- "Not  identified"
  df$condition <- factor(df$condition, labels = c("High Risk","Not  identified","Low Risk"),
                         levels = c("High Risk","Not  identified","Low Risk"))
  
  library(ggpubr)
  corrPlot <- ggscatter(df, x = "PI.x", y = "PI.y",
                    color = "condition", shape = "condition", size = 4, palette = c("red","grey","blue"),        # Points color, shape and size
                    add = "reg.line",                                                                            # Add regression line
                    add.params = list(color = "blue", fill = "lightgray"),                                       # Customize reg. line
                    font.label = c(14, "plain"), 
                    # conf.int = TRUE,                                                                           # Add confidence interval
                    cor.coef = TRUE,                                                                             # Add correlation coefficient
                    cor.coeff.args = list(method = method, label.x.npc = "left", label.y.npc = "top", label.sep = "\n"),
                    cor.method = method, 
                    cor.coef.size = 5) + 
    labs(x=paste0("PI_",name1), y=paste0("PI_",name2), color = set) 
  
  corrPlot <- corrPlot + font("xlab", size = 14, color = "black", face = "plain") +
    font("ylab",size = 14, color = "black", face = "plain") 
  
  corrPlot <-ggpar(corrPlot, font.legend = c(14, "plain", "black"), legend.title = list(color = "", shape = ""))
    
  return(corrPlot)
}