如何使用ggplot2,reshape2和Hmisc在R中进行三角形热图？

我需要帮助在R中使用ggplot2,reshape2和Hmisc进行三角形热图,因为我需要在图上显示r和P值.

我试过cordata[lower.tri(c),]在很多地方插入它并没有帮助.我也试过使用不同的方法,但他们没有显示p值是一个rho,我需要!我试过在这里和谷歌上搜索"Hmisc + triangle + heatmap"并且发现没有任何效果.

这是原始数据,从excel表导入:df

# A tibble: 8 x 7
     Urine   Glucose    Soil         LB Gluconate   River    Colon
                               
1  3222500 377750000 7847250  410000000   3252500 3900000 29800000
2  3667500 187000000 3937500  612000000   5250000 4057500 11075000
3  8362500 196250000 6207500  491000000   2417500 2185000  9725000
4 75700000 513000000 2909750 1415000000   3990000 3405000       NA
5  4485000 141250000 7241000  658750000   3742500 3470000  6695000
6  1947500 235000000 3277500  528500000   7045000 1897500 25475000
7  4130000 202500000  111475  442750000   6142500 4590000  4590000
8  1957500 446250000 8250000  233250000   5832500 5320000  5320000

码:

library(readxl)
data1 <- read_excel("./pca-mean-data.xlsx", sheet = 1)
df <- data1[c(2,3,4,5,6,7,8,9,10,11)]
library(ggplot2)
library(reshape2)
library(Hmisc)
library(stats)
library(RColorBrewer)

abbreviateSTR <- function(value, prefix){  # format string more concisely
  lst = c()
  for (item in value) {
    if (is.nan(item) || is.na(item)) { # if item is NaN return empty string
      lst <- c(lst, '')
      next
    }
    item <- round(item, 2) # round to two digits
    if (item == 0) { # if rounding results in 0 clarify
      item = '<.01'
    }
    item <- as.character(item)
    item <- sub("(^[0])+", "", item)    # remove leading 0: 0.05 -> .05
    item <- sub("(^-[0])+", "-", item)  # remove leading -0: -0.05 -> -.05
    lst <- c(lst, paste(prefix, item, sep = ""))
  }
  return(lst)
}

d <- df

cormatrix = rcorr(as.matrix(d), type='pearson')
cordata = melt(cormatrix$r)
cordata$labelr = abbreviateSTR(melt(cormatrix$r)$value, 'r')
cordata$labelP = abbreviateSTR(melt(cormatrix$P)$value, 'P')
cordata$label = paste(cordata$labelr, "\n", 
                      cordata$labelP, sep = "")


hm.palette <- colorRampPalette(rev(brewer.pal(11, 'Spectral')), space='Lab')

txtsize <- par('din')[2] / 2
pdf(paste("heatmap-MEANDATA-pearson.pdf",sep=""))
ggplot(cordata, aes(x=Var1, y=Var2, fill=value)) + geom_tile() + 
  theme(axis.text.x = element_text(angle=90, hjust=TRUE)) +
  xlab("") + ylab("") + 
  geom_text(label=cordata$label, size=txtsize) + 
  scale_fill_gradient(colours = hm.palette(100))

dev.off()

我附上了我所拥有的一个示例图,我只需要切成两半!请尽可能帮助,我真的很感激!

这是一种使用某些dplyr函数来重塑和过滤数据的方法.在创建相关矩阵之后,我将melt两者结合起来df_cor$r并df_cor$P加入它们,使得将这些数据框组合在一起更简洁(更安全),然后制作标签.

然后我给每一行一对ID,这是一个组合Var1和Var2粘贴在一起的排序版本.因为我对它进行排序,(尿液,土壤)和(土壤,尿液)的行将具有相同的ID,而不考虑哪个是Var1哪个Var2.然后,按照此ID进行分组,我采用不同的观察结果,使用ID作为挑选重复项的唯一标准.这个长形数据的头部如下.

library(tidyverse)
library(Hmisc)
library(reshape2)

# ... function & df definitions removed

df_cor <- rcorr(as.matrix(df), type = "pearson")

df_long <- inner_join(
  melt(df_cor$r, value.name = "r"),
  melt(df_cor$P, value.name = "p"),
  by = c("Var1", "Var2")
) %>%
  mutate(r_lab = abbreviateSTR(r, "r"), p_lab = abbreviateSTR(p, "P")) %>%
  mutate(label = paste(r_lab, p_lab, sep = "\n")) %>%
  rowwise() %>%
  mutate(pair = sort(c(Var1, Var2)) %>% paste(collapse = ",")) %>%
  group_by(pair) %>%
  distinct(pair, .keep_all = T)

head(df_long)
#> # A tibble: 6 x 8
#> # Groups:   pair [6]
#>   Var1      Var2       r         p r_lab p_lab label         pair 
#>                           
#> 1 Urine     Urine  1     NA        r1    ""    "r1\n"        1,1  
#> 2 Glucose   Urine  0.627  0.0963   r.63  P.1   "r.63\nP.1"   1,2  
#> 3 Soil      Urine -0.288  0.489    r-.29 P.49  "r-.29\nP.49" 1,3  
#> 4 LB        Urine  0.936  0.000634 r.94  P<.01 "r.94\nP<.01" 1,4  
#> 5 Gluconate Urine -0.239  0.569    r-.24 P.57  "r-.24\nP.57" 1,5  
#> 6 River     Urine -0.102  0.811    r-.1  P.81  "r-.1\nP.81"  1,6

ggplot(df_long, aes(x = Var1, y = Var2, fill = r)) +
  geom_raster() +
  geom_text(aes(label = label)) +
  scale_fill_distiller(palette = "Spectral") +
  theme_minimal() +
  theme(panel.grid = element_blank())