xx 画图挑战赛
BioMooc.com
2021-12-26 20:13:32
重要信息:
- Web url of this page: R_plot_challenges
- R version: R4.1.0 on CentOS7.9
Good colors
1. single cell 40 colors
par(mar=c(0,0,0,0))
# Good colors
allcolour=c("#DC143C","#0000FF","#20B2AA","#FFA500","#9370DB","#98FB98","#F08080","#1E90FF",
"#7CFC00","#FFFF00","#808000","#FF00FF","#FA8072","#7B68EE","#9400D3","#800080",
"#A0522D","#D2B48C","#D2691E","#87CEEB","#40E0D0","#5F9EA0","#FF1493","#0000CD",
"#008B8B","#FFE4B5","#8A2BE2","#228B22","#E9967A","#4682B4","#32CD32","#F0E68C",
"#FFFFE0","#EE82EE","#FF6347","#6A5ACD","#9932CC","#8B008B","#8B4513","#DEB887")
#length(allcolour) #40
barplot( rep(1,length(allcolour)), col=allcolour, border = NA, space = 0, ann=F, axes=F )
2. Feature plot
par(mar=c(0,0,0,0))
library(RColorBrewer)
barplot( rep(1, 9), col= brewer.pal(9, "YlOrRd"), border = NA, axes=F, space=0 )
3. ggplot2 使用 色板
library(ggplot2)
bp <- ggplot(iris, aes(Species, Sepal.Length)) +
geom_boxplot(aes(fill = Species)) +
theme_minimal() +
theme(legend.position = "top")
bp+scale_fill_brewer(palette = "Dark2")
单个变量
1. barplot：自定义填充色、标签字体颜色
par(mar=c(0,0,0,0))
# view color
barplot(c(1,1), col=c("steelblue", "darkred"), axes=F, border = NA, space=0)
# 1. 模拟数据
dat.tmp=data.frame(
index.name=c(
"KIR+EOMES+ NK-like",
"Tn",
"Temra",
"KIR+TXK+ NK-like",
"NME1+ T",
"Tc17",
paste0("XX_prefix_", 1:4)
),
value=c( seq(0.1,2,0.5), seq(7,10,1), 15, 29 )/100
)
str(dat.tmp)## 'data.frame': 10 obs. of 2 variables:
## $ index.name: chr "KIR+EOMES+ NK-like" "Tn" "Temra" "KIR+TXK+ NK-like" ...
## $ value : num 0.001 0.006 0.011 0.016 0.07 0.08 0.09 0.1 0.15 0.29dat.tmp[, "index.name"]=factor(dat.tmp$index.name, levels = dat.tmp$index.name) #因子形式才能在ggplot2保持顺序
# str(dat.tmp)
#library("ggpubr")
set.seed(2021)
colors2=rnbinom(10,3,0.8)+1;
colors2=c("#BBBCBF", "#73709F", "#78B478")[colors2]; colors2## [1] "#BBBCBF" "#73709F" "#BBBCBF" "#73709F" "#73709F" "#73709F" "#BBBCBF"
## [8] "#78B478" "#78B478" "#73709F"library(ggplot2)
ggplot(dat.tmp, aes(index.name, value)) +
geom_col(fill=ifelse( dat.tmp$value<0.1 ,"steelblue", "darkred"),col=NA,width=0.8) +
geom_hline(yintercept=0.1, linetype="dashed", color="black", alpha=0.2) +
xlab("") + ylab(sprintf("pTrans Of %s", "Tex" ))+
theme_classic(base_size = 12)+
#scale_x_discrete(fill=colors2)+
theme(axis.text.x=element_text(angle=60, face="bold",
color=colors2, #Warning: 不提倡传入向量，以后会删除掉该功能
hjust=1,vjust=1))## Warning: Vectorized input to `element_text()` is not officially supported.
## Results may be unexpected or may change in future versions of ggplot2.
# https://www.jianshu.com/p/1bacaedb58b0
# 看来未来要用原生R画图来实现标签文字染色
# 原生R的绘图
par(mar=c(8,4,0.5,0.5))
posX=barplot(dat.tmp$value, col=ifelse( dat.tmp$value<0.1 ,"steelblue", "darkred"),
ylim=c(0, 0.3), border = NA, #条状图无描边
space=0.1,
yaxs ="r", #y轴上下留白
yaxt="n") #不要y轴
box(bty="l", lwd=2)
#y轴
axis(side=2, at=seq(0.0,0.3, 0.1), labels = seq(0,0.3, 0.1), tck=-0.02,
las=2, mgp=c(1.5,0.5,0))
cell.ident=c("Tex")
mtext( sprintf("pTrans of %s",cell.ident[1]), side = 2, line = 2 )
#x轴
axis(side=1, at=posX, label=NA, tck=-0.02)
text(x=posX, y=par("usr")[3]-0.02,
labels = dat.tmp$index.name,
col = colors2,
adj=1, srt=60, xpd=T)
2个变量
两个变量的情况:
- 两个平等的维度: 散点图
- 一个是数值，另一个是分类变量: boxplot, violin plot
1. 堆叠密度图 Stacked density plot
使用 ggplot2 内置数据 diamonds ，画出各种价位的钻石中，每种切割效果的比例。
library(ggplot2)
library(dplyr)head(diamonds)## # A tibble: 6 × 10
## carat cut color clarity depth table price x y z
## <dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
## 1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
## 2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
## 3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
## 4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63
## 5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
## 6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48# v0 只有密度曲线
ggplot(data=diamonds, aes(x=price, color=cut)) +
geom_density()+
theme_classic()
# v1: 使用fill是概率密度图，但是依旧互相覆盖
diamonds %>% select( c("price", "cut") ) %>%
#head(n=100) %>%
ggplot( aes( x=price, y=..density.., fill=cut ) ) +
geom_density( alpha=0.2 ) #半透明
# v2: 堆叠，且使用比例
diamonds %>% select( c("price", "cut") ) %>%
#head(n=1000) %>%
ggplot( aes( x=price, y=..density.., fill=cut ) ) +
geom_density(position = "fill") #使用 fill 密度曲线标准化为1，且互相堆叠
# v3: 修饰主题
# 差不多就这样了。
ggplot(diamonds, aes( x=price, y=..density.., fill=cut ) ) +
geom_density(position = "fill")+
theme_classic()+
scale_y_continuous(expand = c(0,0))+ #这个可以去掉与X轴间隙
scale_x_continuous(expand = c(0,0)) #这个可以去掉与Y轴间隙
v3: 看答案
# Stacked density plot: https://www.r-graph-gallery.com/135-stacked-density-graph.html#stacked
ggplot(data=diamonds, aes(x=price, group=cut, fill=cut)) +
geom_density(adjust=1.5, position="fill") +
theme_minimal()
# 这里使用的是 group 参数，如果不加呢？貌似没变化
ggplot(data=diamonds, aes(x=price, fill=cut)) +
geom_density(adjust=1.5, position="fill") +
theme_minimal()
# adjust 参数什么意思？取bin的宽度是默认宽度bw的多少倍
# 默认是1倍，越小越精细，越能刻画局部变化
ggplot(data=diamonds, aes(x=price, fill=cut)) +
geom_density(adjust=0.1, position="fill") +
theme_minimal()
# v4：添加自定义颜色，和几个分类一一对应
ggplot(data=diamonds, aes(x=price, fill=factor(cut, levels = c("Fair", "Good", "Very Good", "Premium", "Ideal") ) )) +
geom_density( position="fill") +
theme_minimal() +
#scale_fill_manual("Cut effect", values=c( "red", "blue", "#FFC125", "#AB82FF", "#CD8500" ) )
scale_fill_manual("Cut effect", values=c( "#FF3030", "#009ACD", "#FFC125", "#AB82FF", "#CD8500" ) )
# v5 如何分面画密度曲线？
ggplot(data=diamonds, aes(x=price, fill=cut)) +
geom_density() +
facet_wrap(~cut)+ #按照cut分面
theme_minimal()+
theme(
legend.position = "none", #不要图例
panel.spacing = unit(0.5, "lines"), #每个小图的间隔
axis.text.x=element_text(angle=60), #x坐标轴文字 旋转60度
#axis.ticks.x=element_line(color="red") #x坐标轴刻度 隐藏
)
2. 水平条状图 horizontal barplot(单个，批量)
目的：展示不同气缸数(cyl)的自动/手动档车型的比例。
# opar=par(no.readonly = T)
par(mar=c(1,2,1,1))
df0=table(mtcars$am, mtcars$cyl)
df0##
## 4 6 8
## 0 3 4 12
## 1 8 3 2# 转为百分比，按列
df0=apply(df0, 2, function(x){x/sum(x)})
df0##
## 4 6 8
## 0 0.2727273 0.5714286 0.8571429
## 1 0.7272727 0.4285714 0.1428571# v1 画图
colors = c("#83C57E", "#3A5F88")
barplot(df0, col=colors) #指定颜色
# v2
barplot(df0,
horiz = T, #水平放置
col=colors)
# v3 修饰
barplot(df0,
horiz = T, #水平放置
las=2, #文字和坐标一致
space=0, #bar间距为0
border="white", #白边
xaxt="n", #去掉x坐标轴和轴标签
col=colors)
精细画法
#' horizontal barplot
#'
#' @param df1 color by
#' @param colors Fill color of each bars, by row in df from row1 to row2, ...
#' @param labels text labeled for each bar
#' @param v position of vertical line
#' @param adj.x the position of text label
#' @param scale whether scale to 1 by Column
#'
#' @return
#' @export
#'
#' @examples
barplot_h=function(df1, colors, labels=NULL, v=NULL, adj.x=-0.04, scale=T){
# columns of the df, means number of bars
len=ncol(df1)
df1=df1[,seq(len,1,-1)] # invert by columns index
# normalization by row
if(scale){
df1=apply(df1, 2, function(x){
x/sum(x)
})
}
# 主图
posY=barplot( df1, #color by row, a column a bar
col=colors,
horiz = TRUE, #水平放置
ann=F, xaxt="n", yaxt="n", #去掉x、y轴及文字
#las=2,
border = "white", space=0 )
#添加y轴文字
if(is.null(labels)){
labels=seq(len,1,-1)
}else{
labels=labels[seq(len,1,-1)]
}
text(adj.x, #文字的x坐标
posY, #文字的y坐标
labels=labels, #文字内容
col="black", #颜色
xpd=T) #可显示在主图外部
# 竖直虚线，可选
if(!is.null(v)){
abline(v=v, lty=3, lwd=2, xpd=T)
}
}
# 传入一个矩阵，一列一个柱子，不同行染色不同
# test1: 可以输入一个table(para1, para2)返回的表格，参数1颜色，参数2指定几个bar
df0=table(mtcars$am, mtcars$cyl);df0 # 数据：不同气缸数cyl的自动/手动档车型的比例##
## 4 6 8
## 0 3 4 12
## 1 8 3 2# 4 6 8
# 0 3 4 12 (No)
# 1 8 3 2 (Yes)
par(mfrow=c(5,1), mar=c(1,2,0.5,0))
colors = c("#83C57E", "#3A5F88")
barplot_h(df0, colors)
barplot_h(df0, colors, scale=F, adj.x = -0.3)
barplot_h(df0, colors, labels = c("dog", 'cat', 'goose'), adj.x = -0.07 )
barplot_h(df0, colors, labels = colnames(df0) )
barplot_h(df0, colors, v=0.2 )
# test2 模拟数据
getDf=function(n=7,adjust=10, seed=NULL){
if(!is.null(seed)) set.seed(seed)
t1=rnorm(n, 100, 30)
t1=t1/ max(t1+adjust)
return(t(data.frame(
normal=t1,
tumor=1-t1
)))
}
# test
getDf()## [,1] [,2] [,3] [,4] [,5] [,6] [,7]
## normal 0.6688673 0.92943432 0.93564975 0.4615631 0.6076559 0.4309296 0.8767301
## tumor 0.3311327 0.07056568 0.06435025 0.5384369 0.3923441 0.5690704 0.1232699colors2 = c("#5AC9FF", "#0182D5", "#FE8C72", "#C90C12")
# plot
par(mar=c(0,0, 0, 0), oma=c(0,0,0,0))
mat=matrix(c(1:12),nrow=6,byrow=F); mat #分割画布为6行2列## [,1] [,2]
## [1,] 1 7
## [2,] 2 8
## [3,] 3 9
## [4,] 4 10
## [5,] 5 11
## [6,] 6 12#layout(mat)
layout(mat, widths = c(1,3),heights =c(5,7,6,2,7,7)) #设置左右和上下比例
# Left, text
for(i in 1:6){
plot(c(1), type="n", ann=F, axes=F)
#box(which = "figure", col="red")
text(1.4, 1, cex=1.1, adj=1,
label=c("","Alveolar", "Endothelial", "Epithelial", "Fibroblast", "B cell")[i])
if(1==i){ text(1.4, 1,adj=1, labels=expression( italic("Fake_Data") ), col="red", cex=1.2) }
}
#Right
# legend
plot(c(1), type="n", axes=F)
legend(0.6, 1.4, legend = c("Non-malignant", "Tumor"), fill=colors2[1:2],
#adj = 0, #文字相对于小图
#xjust=0.5, #整个图例的x坐标相对于指定点
border = NA,
ncol = 2,
x.intersp=0.4, #小图和文字间距
text.width = 0.4, #各小图之间水平间距
cex=1.2,
pt.cex=0.5,
inset=.02,
bty="n")
text(1, 0.8, labels="Anticipated fraction\ntumor versus non-malignant",
cex=1.2, adj=0.5)
par(mar=c(0, 1, 0.5, 0))
# plot
barplot_h( getDf(7, 0, 1) , colors2, v=0.3)
barplot_h( getDf(6, 40, 3) , colors2, v=0.3)
barplot_h( getDf(2, 60, 4) , colors2, v=0.3)
barplot_h( getDf(7, 80, 5) , colors2, v=0.3)
barplot_h( getDf(7, 120, 7) , colors2, v=0.3)
3. 右侧带层次聚类的水平 barplot
目的: 随机从 iris 数据集的三个子类中各抽取3个样本，画每个样本中三个指标的百分比 水平barplot，并在右侧添加层次聚类效果。
- 有关聚类的R包可以参考cluster包和ape包
- 要用到 ape 包: http://ape-package.ird.fr/
- R语言最全谱系图 http://rstudio-pubs-static.s3.amazonaws.com/1876_df0bf890dd54461f98719b461d987c3d.html
# 模拟数据：
set.seed(2021)
df0=iris[c(sample(1:50, 3), sample(51:100, 3), sample(101:150, 3)),]
rownames(df0)=paste0(df0$Species,"_", 1:9)
#head(df0)
#table(df0$Species)
#归一化, by observation
df1= data.frame( t(apply(df0[,1:4], 1, function(x){x/sum(x)} )) )
head(df1)## Sepal.Length Sepal.Width Petal.Length Petal.Width
## setosa_1 0.4742268 0.3505155 0.1443299 0.03092784
## setosa_2 0.4900000 0.3600000 0.1400000 0.01000000
## setosa_3 0.5052632 0.3157895 0.1473684 0.03157895
## versicolor_4 0.4000000 0.2142857 0.2928571 0.09285714
## versicolor_5 0.4041096 0.2054795 0.2876712 0.10273973
## versicolor_6 0.3986014 0.1958042 0.3146853 0.09090909# step1. 求距离
out.dist= dist(df1); out.dist## setosa_1 setosa_2 setosa_3 versicolor_4 versicolor_5
## setosa_2 0.02820412
## setosa_3 0.04667768 0.05203344
## versicolor_4 0.22352571 0.24405587 0.21518674
## versicolor_5 0.22727858 0.24831676 0.21713839 0.01479858
## versicolor_6 0.24954658 0.26902641 0.23934955 0.02870168 0.03152262
## virginica_7 0.27489956 0.29535002 0.26607942 0.05142526 0.05071139
## virginica_8 0.27331825 0.29452500 0.26606333 0.05140095 0.05043403
## virginica_9 0.27149932 0.29358160 0.26042938 0.05880821 0.04849519
## versicolor_6 virginica_7 virginica_8
## setosa_2
## setosa_3
## versicolor_4
## versicolor_5
## versicolor_6
## virginica_7 0.02938451
## virginica_8 0.03556237 0.01234903
## virginica_9 0.05025099 0.03885795 0.03618163# step2. 聚类
out.hclust=hclust( out.dist, method="ward.D2" ); out.hclust##
## Call:
## hclust(d = out.dist, method = "ward.D2")
##
## Cluster method : ward.D2
## Distance : euclidean
## Number of objects: 9# step3. 画图，原生R不能旋转树
par(mar=c(0,0,0,0))
plot(out.hclust,
hang=0.1, #最外围长度
ann=F, axes=F,
cex = 0.7,
col = "darkred")
step4. 精修图
# 使用 ape包绘制右侧聚类数，使用上文2定义的 barplot_h()函数绘制barplot
# plot 253*310
# 对矩阵的观测值按照聚类结果调整顺序，用来画条形图
df1=df1[ rev(out.hclust$order),]
df1## Sepal.Length Sepal.Width Petal.Length Petal.Width
## virginica_8 0.3734177 0.1898734 0.3227848 0.11392405
## virginica_7 0.3809524 0.1845238 0.3273810 0.10714286
## virginica_9 0.3895349 0.1744186 0.3023256 0.13372093
## versicolor_5 0.4041096 0.2054795 0.2876712 0.10273973
## versicolor_4 0.4000000 0.2142857 0.2928571 0.09285714
## versicolor_6 0.3986014 0.1958042 0.3146853 0.09090909
## setosa_2 0.4900000 0.3600000 0.1400000 0.01000000
## setosa_1 0.4742268 0.3505155 0.1443299 0.03092784
## setosa_3 0.5052632 0.3157895 0.1473684 0.03157895# set colors
colors2=c("#A5CDE1", "#1F78B3", "#349F48", "#F47E20")
# pdf("bar_plot_with_tree.pdf", width=2.53, height=3.1)
par(mar=c(0,0, 0, 0), oma=c(0,0,0,0))
mat=matrix(c(1,1,2,3),nrow=2,byrow=T); mat #分割画布为2行2列## [,1] [,2]
## [1,] 1 1
## [2,] 2 3#layout(mat)
layout(mat, widths = c(16,4),heights =c(1,6)) #设置左右和上下比例
#1 up pannel: legend
# 警告: 通常 Rmarkdown图和pdf图是一致的。而R终端直接显示的图则略有错位
plot(c(1), type="n", axes=F)
legend(0.7, 1.4,
legend = c("Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width"),
#c("BP c-kit+Gr1", "BM Gr1", "PB Gr1", "SP Gr1"),
fill=colors2,
#adj = 0, #文字相对于小图
#xjust=0.5, #整个图例的x坐标相对于指定点
border = NA,
ncol = 2,
x.intersp=0.2, #小图和文字间距
y.intersp=1.2, #小图上下间距
text.width = 0.25, #各小图之间水平间距
cex=0.9,
pt.cex=0.5,
inset=.02,
bty="n")
#2 down left
#par("mar") #5.1 4.1 4.1 2.1
par(mar=c(3,2.5, 0, 0))
barplot_h( t(df1), colors=colors2,
labels = substring(rownames(df1), nchar(rownames(df1))-3),
adj.x = -0.12)
axis(side=1, at=seq(0,1,0.25), labels = seq(0,1,0.25), mgp=c(1,0.5,0))
box(bty="l", col="black", lwd=2)
#3 down right
par(mar=c(3.8, 0.1, 0.8, 1))
library(ape)
# ape::plot.phylo() #F1 查看更多帮助
plot(as.phylo(out.hclust),
direction ="leftwards",
edge.width = 2, #线条粗细
use.edge.length=F, #是否使用各类的距离？
show.tip.label=F, #不显示文字
cex = 2) #文字放大倍数
特殊需求
旋转 ggplot2 图形
library(ggplot2)
bp <- ggplot(iris, aes(Species, Sepal.Length)) +
geom_boxplot(aes(fill = Species)) +
theme_minimal() +
theme(legend.position = "top")
bp2=bp+scale_fill_brewer(palette = "Dark2")
# bp2
library(grid)
grid.newpage()
vp1=viewport(x=1, y=0.4, width=0.8, height=0.7, just=c("right", "bottom"), angle=30)
pushViewport(vp1);
# grid.rect()
print(bp2, newpage=F)## Warning in grid.Call.graphics(C_setviewport, vp, TRUE): cannot clip to rotated
## viewport
## Warning in grid.Call.graphics(C_setviewport, vp, TRUE): cannot clip to rotated
## viewport
ggplot2 图例放左侧，图例上移
g1=ggplot(mtcars, aes(wt, mpg, color=gear) )+
geom_point(size=1)+
scale_color_gradientn(name="Z Score", colors=brewer.pal(9, "YlOrRd"))+
theme_classic(base_size = 12)+
theme(
legend.position = c("left"),
legend.text=element_text(size=8),
legend.title=element_text(size=8),
legend.key.width = unit(4, "mm"), #control legend width and height
legend.key.height = unit(4, "mm"),
legend.margin=margin(t = 0, r = 0, b = 70, l = 0, unit = "pt"), #move legend up a litter
);
print(g1)
# the end of this file ==
# The file is under updating now and then.## [1] "0.18 minutes elapsed."