Polar plot for survival and correlation connection

library(ggpolar)
#> Loading required package: ggplot2
library(survival)
library(ezcox)
#> Welcome to 'ezcox' package!
#> =======================================================================
#> You are using ezcox version 1.0.4
#> 
#> Project home : https://github.com/ShixiangWang/ezcox
#> Documentation: https://shixiangwang.github.io/ezcox
#> Cite as      : arXiv:2110.14232
#> =======================================================================
#>

data = survival::lung
head(data)
#>   inst time status age sex ph.ecog ph.karno pat.karno meal.cal wt.loss
#> 1    3  306      2  74   1       1       90       100     1175      NA
#> 2    3  455      2  68   1       0       90        90     1225      15
#> 3    3 1010      1  56   1       0       90        90       NA      15
#> 4    5  210      2  57   1       1       90        60     1150      11
#> 5    1  883      2  60   1       0      100        90       NA       0
#> 6   12 1022      1  74   1       1       50        80      513       0

Pick several variables.

vars = c("age", "sex", "ph.ecog", "ph.karno", "pat.karno", "meal.cal", "wt.loss")

Univariable Cox analysis

df_cox = ezcox(data, vars)
df_cox
#> # A tibble: 7 × 12
#>   Variable  is_control contrast_level ref_level n_contrast n_ref      beta    HR
#>   <chr>     <lgl>      <chr>          <chr>          <int> <int>     <dbl> <dbl>
#> 1 age       FALSE      age            age              228   228  0.0187   1.02 
#> 2 sex       FALSE      sex            sex              228   228 -0.531    0.588
#> 3 ph.ecog   FALSE      ph.ecog        ph.ecog          227   227  0.476    1.61 
#> 4 ph.karno  FALSE      ph.karno       ph.karno         227   227 -0.0164   0.984
#> 5 pat.karno FALSE      pat.karno      pat.karno        225   225 -0.0199   0.98 
#> 6 meal.cal  FALSE      meal.cal       meal.cal         181   181 -0.000124 1    
#> 7 wt.loss   FALSE      wt.loss        wt.loss          214   214  0.00132  1    
#> # ℹ 4 more variables: lower_95 <dbl>, upper_95 <dbl>, p.value <dbl>,
#> #   global.pval <dbl>

Correlation analysis

vars_comb = combn(vars, 2, simplify = FALSE)
cor_value = sapply(vars_comb, function(x) {
  cor(data[[x[1]]], data[[x[2]]], use = "pairwise")
})

df_cor = cbind(as.data.frame(t(sapply(vars_comb, function(x) x))), cor_value)
colnames(df_cor) = c("var1", "var2", "correlation")
df_cor$size = abs(df_cor$correlation)
df_cor$way = ifelse(df_cor$correlation > 0, "positive", "negative")
df_cor
#>         var1      var2 correlation       size      way
#> 1        age       sex -0.12216709 0.12216709 negative
#> 2        age   ph.ecog  0.19323604 0.19323604 positive
#> 3        age  ph.karno -0.20318207 0.20318207 negative
#> 4        age pat.karno -0.12616688 0.12616688 negative
#> 5        age  meal.cal -0.23141071 0.23141071 negative
#> 6        age   wt.loss  0.03814787 0.03814787 positive
#> 7        sex   ph.ecog -0.02060379 0.02060379 negative
#> 8        sex  ph.karno  0.01138505 0.01138505 positive
#> 9        sex pat.karno  0.04607145 0.04607145 positive
#> 10       sex  meal.cal -0.16835976 0.16835976 negative
#> 11       sex   wt.loss -0.12907708 0.12907708 negative
#> 12   ph.ecog  ph.karno -0.80726666 0.80726666 negative
#> 13   ph.ecog pat.karno -0.51122086 0.51122086 negative
#> 14   ph.ecog  meal.cal -0.09851018 0.09851018 negative
#> 15   ph.ecog   wt.loss  0.18758944 0.18758944 positive
#> 16  ph.karno pat.karno  0.52029737 0.52029737 positive
#> 17  ph.karno  meal.cal  0.04223324 0.04223324 positive
#> 18  ph.karno   wt.loss -0.17543452 0.17543452 negative
#> 19 pat.karno  meal.cal  0.16575874 0.16575874 positive
#> 20 pat.karno   wt.loss -0.17199064 0.17199064 negative
#> 21  meal.cal   wt.loss -0.10257242 0.10257242 negative

Visualization

df_cox$role = ifelse(
  df_cox$p.value > 0.05, "non-signf",
  ifelse(df_cox$HR < 1, "protector", "risker")
)
df_cox$`-log10(p)` = -log10(df_cox$p.value)

p = polar_init(df_cox, x = Variable, aes(color = role, size = `-log10(p)`))
p

p + 
  ggnewscale::new_scale("color") +
  polar_connect(df_cor, x1 = var1, x2= var2, size = size, color = way, alpha = 0.3) + 
  scale_size(range = c(0.1, 4))