在 geom_histogram 上绘制 geom_point

Question

我有一个数据框：

test <- structure(list(Sample_ID = c("S1","S2", "S3", "S4", "S1", "S2", "S3", "S4"), 
                       CN_Region = c("A", "A", "A", "A", "B", "B", "B", "B"),
                       MedianLog2Ratio = c(-0.2, -0.2, -0.25, -0.25, -0.25, -0.2, -0.1, -0.3), 
             CN_truth = c("2", "2", "2", NA, "2", "2", "2", "1")), class = c("data.table","data.frame"))

当我绘制hist 时，它工作正常： hist(test$MedianLog2Ratio)

我想使用 ggplot 绘制每个区域的直方图，并使用与 SampleID 关联的 geom_points CN_truth 叠加：

g <- ggplot(test, aes(x = MedianLog2Ratio)) + geom_histogram()
g + geom_point(aes(colour = factor(CN_truth))

情节应该看起来像这样松散的（当然它会有更少的数据块和更少的数据）： 其中图例指的是CN_truth，标题是CN_Region

Answer 1

一种方法如下：

test <- data.frame(Sample_ID = c("S1","S2", "S3", "S4", "S1", "S2", "S3", "S4"), 
                       CN_Region = c("A", "A", "A", "A", "B", "B", "B", "B"),
                       MedianLog2Ratio = c(-0.2, -0.2, -0.25, -0.25, -0.25, -0.2, -0.1, -0.3), 
                       CN_truth = c("2", "2", "2", NA, "2", "2", "2", "1"))
test <- transform(test, freqmlr = ave(seq(nrow(test)), MedianLog2Ratio, FUN=length))



g <- ggplot(test, aes(x = MedianLog2Ratio)) + geom_histogram(color="black", fill="white")+
     geom_point(aes(x=MedianLog2Ratio, y=freqmlr, colour=factor(CN_truth)))+
        xlab('MedianLog2Ratio') +
        ylab('Freq')+
        labs(colour='CN_truth')
g 

如果您想从图例中删除 NA（例如 this 一个），有很多帖子适合您。另请注意，如果 x 轴有许多具有相同值的点，您可以在每个直方图中稍微移动它们以使它们可见。例如通过添加随机十进制值：

g <- ggplot(test, aes(x = MedianLog2Ratio)) + geom_histogram(color="black", fill="white")+
  geom_point(aes(x=(MedianLog2Ratio+runif(nrow(test), 0.0, 0.010)), y=freqmlr, 
                 colour=CN_truth ))+
  xlab('MedianLog2Ratio') +
  ylab('Freq')+
  labs(colour='CN_truth')
g  + scale_colour_manual(values = c("red", "blue"), limits = c("1", "2"))