ggplot facet_wrap 选定的data.frame列？答案

【问题标题】：ggplot facet_wrap selected columns of data.frame?ggplot facet_wrap 选定的data.frame列？
【发布时间】：2017-01-17 01:57:34
【问题描述】：

我有一个data.frame X，其中包含点/样本坐标X1 和X2：

> head(X)
                               X1        X2 Cluster Timepoint Transcripts       MEF       ESC
Drop_6_6A_0_TACCTAATCTAC 169.3437  20.18623       2     Day 0       49688 0.4366071 0.3260743
Drop_6_6A_0_TCAGCTTGTCAC 155.8880 -16.69927       3     Day 0       47365 0.4554254 0.3350818
Drop_6_6A_0_TCGCAATAAGAT 168.4270  36.50967       2     Day 0       44881 0.4114934 0.2595030
Drop_6_6A_0_AATCTACCAATC 164.3964 -27.17404       3     Day 0       44640 0.4748225 0.3525822
Drop_6_6A_0_GGATTAAGTTCA 162.2900 -24.10504       3     Day 0       36822 0.4723676 0.3391785
Drop_6_6A_0_TGATCTAGTGTC 155.4231 -19.18974       3     Day 0       35889 0.4664174 0.3408899

我想将选定的标记作为列添加到X，并根据关联的表达式值调整散点图上的点大小。

NANOG  = t(data['NANOG',rownames(X)])
SAL4 = t(data['SAL4',rownames(X)])
COL5A2 = t(data['COL5A2',rownames(X)])
ESRRB = t(data['ESRRB',rownames(X)])
ELN    = t(data['ELN',rownames(X)])
POU5f1 = t(data['POU5F1',rownames(X)])
PTN    = t(data['PTN',rownames(X)])
CXCL5  = t(data['CXCL5',rownames(X)])
Z = cbind(X, NANOG, SAL4, POU5f1, ESRRB, COL5A2, ELN, PTN, CXCL5)

绑定此数据后，新的data.frame Z 如下所示：

> head(Z)
                               X1        X2 Cluster Timepoint Transcripts       MEF       ESC     NANOG NA POU5F1 ESRRB   COL5A2 ELN      PTN     CXCL5
Drop_6_6A_0_TACCTAATCTAC 169.3437  20.18623       2     Day 0       49688 0.4366071 0.3260743 0.0000000 NA      0     0 5.113106   0 1.004522 0.2645434
Drop_6_6A_0_TCAGCTTGTCAC 155.8880 -16.69927       3     Day 0       47365 0.4554254 0.3350818 0.2763494 NA      0     0 3.068572   0 1.309109 1.0395819
Drop_6_6A_0_TCGCAATAAGAT 168.4270  36.50967       2     Day 0       44881 0.4114934 0.2595030 0.0000000 NA      0     0 5.264248   0 0.000000 0.0000000
Drop_6_6A_0_AATCTACCAATC 164.3964 -27.17404       3     Day 0       44640 0.4748225 0.3525822 0.0000000 NA      0     0 3.554919   0 1.592698 0.2916205
Drop_6_6A_0_GGATTAAGTTCA 162.2900 -24.10504       3     Day 0       36822 0.4723676 0.3391785 0.0000000 NA      0     0 3.838676   0 1.536569 1.9954283
Drop_6_6A_0_TGATCTAGTGTC 155.4231 -19.18974       3     Day 0       35889 0.4664174 0.3408899 0.0000000 NA      0     0 4.029014   0 6.187616 0.0000000

现在，我可以绘制单个散点图，点的大小与相应的表达式值（如下所示）一致，但我不确定如何在一个 facet_wrap 图中执行此操作。

library(gridExtra)
g = arrangeGrob(
  ggplot(Z, aes(X1, X2, color=NANOG)) + ggtitle("NANOG") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='red') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.nanog.expression.no.noise.pdf"), height=pdf_height, width=pdf_width+5),
  ggplot(Z, aes(X1, X2, color=SAL4)) + ggtitle("SAL4") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='red') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.SAL4.expression.no.noise.pdf"), height=pdf_height, width=pdf_width+5),
  ggplot(Z, aes(X1, X2, color=POU5f1)) + ggtitle("POU5F1") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='red') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.pou5f1.expression.pdf"), height=pdf_height, width=pdf_width+5), 
  ggplot(Z, aes(X1, X2, color=ESRRB)) + ggtitle("ESRRB") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='red') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.ESRRB.expression.pdf"), height=pdf_height, width=pdf_width+5), 
  ggplot(Z, aes(X1, X2, color=COL5A2)) + ggtitle("COL5A2") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='green') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.col5a2.expression.pdf"), height=pdf_height, width=pdf_width+5), 
  ggplot(Z, aes(X1, X2, color=ELN)) + ggtitle("ELN") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='green') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.eln.expression.pdf"), height=pdf_height, width=pdf_width+5), 
  ggplot(Z, aes(X1, X2, color=PTN)) + ggtitle("PTN") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='green') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.ptn.expression.pdf"), height=pdf_height, width=pdf_width+5), 
  ggplot(Z, aes(X1, X2, color=CXCL5)) + ggtitle("CXCL5") + 
    geom_point() + 
    xlab(paste0("TSNE1")) + 
    ylab(paste0("TSNE2")) +
    theme_bw() + theme(axis.line = element_line(colour = "black"), panel.grid.minor = element_blank(), panel.background = element_blank()) + scale_colour_gradient(low='light blue', high='green') + 
    ggsave(paste0(outdir, timepoint, ".tsne.",lab,".density.clustered.all.genes.TSNE1.TSNE2.cxcl5.expression.pdf"), height=pdf_height, width=pdf_width+5), 
  nrow=2, ncol=4
)

上面的代码按预期工作，但非常冗长，不足以容纳大量的选定标记，比如 100 个。我假设我必须以某种方式融化 Z data.frame？任何帮助将不胜感激。

【问题讨论】：

请提供一个可重现的例子。
我的博文可能会帮助你解决这个问题：drsimonj.svbtle.com/plot-some-variables-against-many-others

标签： r ggplot2 facet-wrap

【解决方案1】：

正如 OP 所建议的，一种方法是融合原始数据帧Z：

library(reshape2)
d <- melt(Z, id = 1:5, measure = 6:ncol(Z))

其中id 可以是 id 变量的整数（列索引）或字符串（列名）向量，measure 是一个向量，它给出了各种度量的位置（本例中的标记）。然后拨打ggplot：

library(ggplot2)
ggplot(d, aes(x = X1, y = X2, size = value)) +
  geom_point() +
  facet_wrap(~ variable)

根据需要添加标签和其他装饰。使用 OP 从Z 提取的输出：

【讨论】：