在timeseries中按行在列中峰值的时间对行进行分组。

Question

我正在设计一个细胞周期时间序列中高度调控基因的热图。

example <- read.csv("example.csv", header = T)

example.m <- melt(example)

(e <- ggplot(example.m, aes(variable, Gene_ID)) + geom_tile(aes(fill = 
value), colour = "white") + scale_fill_gradient(low = "white", high = 
"steelblue"))

结果是这样的，

示例热图

我的值是日志转换的，我想知道是否有一种方法来排列我的行，使它们按它们在时间序列中的峰值位置排列(即，所有在0处表达最高的基因被聚在一起，在30处表达最高的基因被聚在一起，等等)。

我试着这样做

order <- arrange(example, X0, X30, X60, X90, X120, X150, X180, X210, X240)

然后用有序的数据帧绘制热图，但没有发生变化。

感谢您提供的任何帮助或建议。我真的很感激你的时间。

Answer 1

您应该能够添加这一行来设置Y轴example.m$Gene_ID <- factor(example.m$Gene_ID, levels = order$Gene_ID, labels = order$Gene_ID)的顺序

下面是包含一些示例数据的完整代码：

example <- data.frame(Gene_ID = paste0("TTHERM_", 1:9),
                      X0 = round(runif(9, min =0, max = 4.4999),0), 
                      X30 = round(runif(9, min =0, max = 4.4999),0), 
                      X60 = round(runif(9, min =0, max = 4.4999),0), 
                      X90 = round(runif(9, min =0, max = 4.4999),0), 
                      X120 = round(runif(9, min =0, max = 4.4999),0),
                      X150 = round(runif(9, min =0, max = 4.4999),0),
                      X180 = round(runif(9, min =0, max = 4.4999),0),
                      X210 = round(runif(9, min =0, max = 4.4999),0),
                      X240 = round(runif(9, min =0, max = 4.4999),0))

library(dplyr)
library(reshape2)
library(ggplot2)
example.m <- melt(example)

# This is your original plot
(e <- ggplot(example.m, aes(variable, Gene_ID)) + geom_tile(aes(fill = 
                                                                  value), colour = "white") + scale_fill_gradient(low = "white", high = 
                                                                                                                    "steelblue"))
# Your order command gives us the right order
order <- arrange(example, X0, X30, X60, X90, X120, X150, X180, X210, X240)

# This changes the order of the Y axis based on the sort order
example.m$Gene_ID <- factor(example.m$Gene_ID, levels = order$Gene_ID, labels = order$Gene_ID)

# This is the new plot
(e <- ggplot(example.m, aes(variable, Gene_ID)) + geom_tile(aes(fill = 
                                                                  value), colour = "white") + scale_fill_gradient(low = "white", high = 
                                                                                                                    "steelblue"))

原情节：

​

​

新情节：

​

​

这就是你想要的吗？