其他
R语言学习 - 线图一步法
线图 - 一步绘制
绘图时通常会碰到两个头疼的问题：
有时需要绘制很多的图，唯一的不同就是输入文件，其它都不需要修改。如果用R脚本，需要反复替换文件名，繁琐又容易出错。 (R也有命令行参数，不熟，有经验的可以尝试下)
每次绘图都需要不断的调整参数，时间久了不用，就忘记参数怎么设置了；或者调整次数过多，有了很多版本，最后不知道用哪个了。
为了简化绘图、维持脚本的一致，我用bash对绘图命令做了一个封装，通过配置修改命令行参数，生成相应的绘图脚本，然后再绘制。
首先把测试数据存储到文件中方便调用。数据矩阵存储在line_data.xls和line_data_melt.xls文件中 (直接拷贝到文件中也可以，这里这么操作只是为了随文章提供个测试文件，方便使用。如果你手上有自己的数据，也可以拿来用)。
profile = "Pos;H3K27ac;CTCF;Enhancer;H3K4me3;polII
-5000;8.7;10.7;11.7;10;8.3
-4000;8.4;10.8;11.8;9.8;7.8
-3000;8.3;10.5;12.2;9.4;7
-2000;7.2;10.9;12.7;8.4;4.8
-1000;3.6;8.5;12.8;4.8;1.3
0;3.6;8.5;13.4;5.2;1.5
1000;7.1;10.9;12.4;8.1;4.9
2000;8.2;10.7;12.4;9.5;7.7
3000;8.4;10.4;12;9.8;7.9
4000;8.5;10.6;11.7;9.7;8.2
5000;8.5;10.6;11.7;10;8.2"
profile_text <- read.table(text=profile, header=T, row.names=1, quote="",sep=";")
# tab键分割，每列不加引号
write.table(profile_text, file="line_data.xls", sep="\t", row.names=T, col.names=T,quote=F)
# 如果看着第一行少了ID列不爽，可以填补下
system("sed -i '1 s/^/ID\t/' line_data.xls")profile = "Pos;variable;value;set
-5000;H3K27ac;8.71298;A
-4000;H3K27ac;8.43246;A
-3000;H3K27ac;8.25497;A
-2000;H3K27ac;7.16265;A
-1000;H3K27ac;3.55341;A
0;H3K27ac;3.5503;A
1000;H3K27ac;7.07502;A
2000;H3K27ac;8.24328;A
3000;H3K27ac;8.43869;A
4000;H3K27ac;8.48877;A
-5000;CTCF;10.6913;A
-4000;CTCF;10.7668;A
-3000;CTCF;10.5441;A
-2000;CTCF;10.8635;A
-1000;CTCF;8.45751;A
0;CTCF;8.50316;A
1000;CTCF;10.9143;A
2000;CTCF;10.7022;A
3000;CTCF;10.4101;A
4000;CTCF;10.5757;A
-5000;H3K27ac;8.71298;B
-4000;H3K27ac;8.43246;B
-3000;H3K27ac;8.25497;B
-2000;H3K27ac;7.16265;B
-1000;H3K27ac;3.55341;B
0;H3K27ac;3.5503;B
1000;H3K27ac;7.07502;B
2000;H3K27ac;8.24328;B
3000;H3K27ac;8.43869;B
4000;H3K27ac;8.48877;B
-5000;CTCF;10.6913;B
-4000;CTCF;10.7668;B
-3000;CTCF;10.5441;B
-2000;CTCF;10.8635;B
-1000;CTCF;8.45751;B
0;CTCF;8.50316;B
1000;CTCF;10.9143;B
2000;CTCF;10.7022;B
3000;CTCF;10.4101;B
4000;CTCF;10.5757;B"
profile_text <- read.table(text=profile, header=T, quote="",sep=";")
# tab键分割，每列不加引号
write.table(profile_text, file="line_data_melt.xls", sep="\t", row.names=T, col.names=T,quote=F)
# 如果看着第一行少了ID列不爽，可以填补下
system("sed -i '1 s/^/ID\t/' line_data_melt.xls")使用正常矩阵默认参数绘制个线图
# -f: 指定输入的矩阵文件，第一列为行名字，第一行为header
列数不限，列名字不限；行数不限，行名字默认为文本
# -A FALSE: 指定行名为数字
sp_lines.sh -f line_data.xls -A FALSE# -l: 设定图例的顺序
# -o TRUE: 局部拟合获得平滑曲线
# -A FALSE: 指定行名为数字
# -P: 设置legend位置，相对于原点的坐标
# -x, -y指定横纵轴标记
sp_lines.sh -f line_data.xls -l "'CTCF','Enhancer','polII','H3K4me3','H3K27ac'" -P 'c(0.8,0.3)' -o TRUE -A FALSE -x 'Up and down 5 kb of TSS' -y 'Relative density'# -A FALSE: 指定行名为数字
# -V 'c(-1000, 500)': 设置垂线的位置
# -D: 设置垂线的文本标记，参数为引号引起来的vector，注意引号的嵌套
# -I: 设置横轴的标记的位置
# -b: 设置横轴标记的文字
sp_lines.sh -f line_data.xls -A FALSE -V 'c(-1000,500)' -D "c('+1 kb','-0.5 kb')" -I "c(-5000,0,5000)" -b "c('-5 kb', 'TSS', '+5 kb')"使用melted矩阵默认参数绘制个线图 (除需要改变文件格式，指定-m TRUE -a xvariable外其它与正常矩阵一样)
# -f: 指定输入文件
# -m TRUE: 指定输入的矩阵为melted format, 三列，第一列为Pos (给-a)
# 第二列为variable (给-H，-H默认即为variable)
# 第三列为value，名字不可修改
# -A FALSE: 指定行名为数字
# -P 'c(0.8,0.2)': 设置legend位置，相对于原点的坐标
sp_lines.sh -f line_data_melt.xls -a Pos -m TRUE -A FALSE -P 'c(0.8,0.2)'完整的图
# -C: 自定义线的颜色
sp_lines.sh -f line_data_melt.xls -a Pos -m TRUE -A FALSE -P 'c(0.8,0.2)' -o TRUE -V 'c(-1000,500)' -D "c('+1 kb','-0.5 kb')" -I "c(-5000,0,4000)" -b "c('-5 kb', 'TSS', '+4 kb')" -x 'Up 5 kb and down 4 kb of TSS' -y 'Relative density' -C "'pink', 'blue'"外层引号与内层引号不能相同
凡参数值中包括了
空格，括号，逗号等都用引号括起来作为一个整体。
完整参数列表如下：
ct@ehbio:~ $sp_lines.sh
***CREATED BY Chen Tong (chentong_biology@163.com)***
Usage:
/MPATHB/self/s-plot/sp_lines.sh options
Function:
This script is used to draw a line or multiple lines using ggplot2.
You can specify whether or not smooth your line or lines.
Two types of input files are supported, normal matrix or melted matrix format. Column separator for both types of input files is **tab**.
Here is an example of normal matrix format. The first column will be treated as X-axis variables and other columns represents each type of lines. The number of columns is unlimited and names of columns is unlimited.
**Set** column is not needed. If given, <facet_plot> (multiple plots in one page) could be displayed.
------------------------------------------------------------
Pos H3K27ac CTCF Enhancer H3K4me3 polII
-5000 8.71298 10.69130 11.7359 10.02510 8.26866
-4000 8.43246 10.76680 11.8442 9.76927 7.78358
-3000 8.25497 10.54410 12.2470 9.40346 6.96859
-2000 7.16265 10.86350 12.6889 8.35070 4.84365
-1000 3.55341 8.45751 12.8372 4.84680 1.26110
0 3.55030 8.50316 13.4152 5.17401 1.50022
1000 7.07502 10.91430 12.3588 8.13909 4.88096
2000 8.24328 10.70220 12.3888 9.47255 7.67968
3000 8.43869 10.41010 11.9760 9.80665 7.94148
4000 8.48877 10.57570 11.6562 9.71986 8.17849
------------------------------------------------------
------------With SET------------------------------------------
Pos H3K27ac CTCF Enhancer H3K4me3 polII Set
-5000 8.71298 10.69130 11.7359 10.02510 8.26866 1
-4000 8.43246 10.76680 11.8442 9.76927 7.78358 1
-3000 8.25497 10.54410 12.2470 9.40346 6.96859 1
-2000 7.16265 10.86350 12.6889 8.35070 4.84365 1
-1000 3.55341 8.45751 12.8372 4.84680 1.26110 1
0 3.55030 8.50316 13.4152 5.17401 1.50022 1
1000 7.07502 10.91430 12.3588 8.13909 4.88096 1
2000 8.24328 10.70220 12.3888 9.47255 7.67968 1
3000 8.43869 10.41010 11.9760 9.80665 7.94148 1
4000 8.48877 10.57570 11.6562 9.71986 8.17849 1
-5000 8.71298 10.69130 11.7359 10.02510 8.26866 2
-4000 8.43246 10.76680 11.8442 9.76927 7.78358 2
-3000 8.25497 10.54410 12.2470 9.40346 6.96859 2
-2000 7.16265 10.86350 12.6889 8.35070 4.84365 2
-1000 3.55341 8.45751 12.8372 4.84680 1.26110 2
0 3.55030 8.50316 13.4152 5.17401 1.50022 2
1000 7.07502 10.91430 12.3588 8.13909 4.88096 2
2000 8.24328 10.70220 12.3888 9.47255 7.67968 2
3000 8.43869 10.41010 11.9760 9.80665 7.94148 2
4000 8.48877 10.57570 11.6562 9.71986 8.17849 2
-------------------------------------------------------------
For matrix format, example command lines include:
* Attribute of X-axis value (first column of matrix) is <number>
s-plot lines -f matrix.file -A FALSE
* Attribute of X-axis value (first column of matrix) is <text>
s-plot lines -f matrix.file
* Attribute of X-axis value (first column of matrix) is numbers, change legned order (default alphabet order)
s-plot lines -f matrix.file -l "'polII', 'CTCF', 'Enhancer', 'H3K27ac', 'H3K4me3'"
* Attribute of X-axis value (first column of matrix) is numbers, change legned order (default alphabet order), smooth lines to look better (Pay attention to whether this will change the data trend)
s-plot lines -f matrix.file -l "'polII', 'CTCF', 'Enhancer', 'H3K27ac', 'H3K4me3'" -o TRUE
* Attribute of X-axis value (first column of matrix) is numbers, with <Set> (Set is column name) column
s-plot lines -f matrix.file -F "+facet_grid(Set ~ ., scale='free_y')"
FILEFORMAT when -m is true
#The name "value" shoud **not** be altered.
#variable can be altered using -H
#Actually this format is the melted result of last format.
--------------------------------------------------------------
Pos variable value
-5000 H3K27ac 8.71298
-4000 H3K27ac 8.43246
-3000 H3K27ac 8.25497
-2000 H3K27ac 7.16265
-1000 H3K27ac 3.55341
0 H3K27ac 3.55030
1000 H3K27ac 7.07502
2000 H3K27ac 8.24328
3000 H3K27ac 8.43869
4000 H3K27ac 8.48877
-5000 CTCF 10.69130
-4000 CTCF 10.76680
-3000 CTCF 10.54410
-2000 CTCF 10.86350
-1000 CTCF 8.45751
0 CTCF 8.50316
1000 CTCF 10.91430
2000 CTCF 10.70220
3000 CTCF 10.41010
4000 CTCF 10.57570
-------------------------------------------------------------
* Attribute of X-axis value (melt format) is <number>
s-plot lines -f matrix.file -m TRUE -a Pos -A FALSE
* Attribute of X-axis value (first column of matrix) is <text>
s-plot lines -f matrix.file -m TRUE -a Pos
* If the name of the second column is <type> not <variable>, one should specify with <-H>.
s-plot lines -f matrix.file -A FALSE -m TRUE -a Pos -H type
* Attribute of X-axis value (first column of matrix) is numbers, change legned order (default alphabet order)
s-plot lines -f matrix.file -m TRUE -a Pos -l "'polII', 'CTCF', 'Enhancer', 'H3K27ac', 'H3K4me3'"
* Attribute of X-axis value (first column of matrix) is numbers, change legned order (default alphabet order), smooth lines to look better (Pay attention to whether this will change the data trend)
s-plot lines -f matrix.file -m TRUE -a Pos -l "'polII', 'CTCF', 'Enhancer', 'H3K27ac', 'H3K4me3'" -o TRUE
* Attribute of X-axis value (first column of matrix) is numbers, with <Set> (Set is column name) column
s-plot lines -f matrix.file -F "+facet_grid(Set ~ ., scale='free_y')"
OPTIONS:
-f Data file (with header line, the first column would be be treated as rownames for
normal matrix. No rownames for melted format. Columns are tab seperated)
[NECESSARY]
-m When true, it will skip melt preprocesses. But the format must be
the same as listed before.
[Default FALSE, accept TRUE]
-a Name for x-axis variable
[Only needed when <-m> is <TRUE>.
For the melted data, 'Pos' should be given here.
For normal matrix, default the first column will be used,
program will assign an value 'xvariable' to represent it.
]]
-A Are x-axis variables numbers.
[Default <TRUE>, meaning X-axis label is <text>.
<FALSE> means X-axis label is <numerical>.]
-H Name for legend variable.
[Default variable, this should only be set when -m is TRUE]
-J Name for color variable.
[Default same as -H, this should only be set when -m is TRUE]
-l Set orders of legend variable.
[Default column order for normal matrix, accept a string like
"'CTCF','H3K27ac','Enhancer'" to set your own order.
Pay attention to the usage of two types of quotes.
***When -m is TRUE, default order would be alphabet order.*********
]
-P Legend position[Default right. Accept
top, bottom, left, none, or 'c(0.08,0.8)'.]
-L Levels for x-axis variable, suitable when x-axis is not treated as numerical.
[Default the order of first column for normal matrix.
Accept a string like "'g','a','j','x','s','c','o','u'" to set your own oder.
This will only be considered when -A is TRUE.
***When -m is used, this default order would be alphabet order.*********
]
-o Smooth lines or not.
[Default FALSE means no smooth. Accept TRUE to smooth lines.]
-O The smooth method you want to use.
[smoothing method (function) to use, eg. lm, glm, gam, loess,rlm.
For datasets with n < 1000 default is 'loess'.
For datasets with 1000 or more observations defaults to 'gam'.
]
-V Add vertical lines.[Default FALSE, accept a series of
numbers in following format "c(1,2,3,4,5)" or other
R code that can generate a vector.]
-D Add labels to vlines.
[Default same as -V.
Accept a series of numbers in following format "c(1,2,3,4,5)" or other R code
that can generate a vector as labels.
Or one can give '1' to disallow labels]
-j Add horizontal lines.[Default FALSE, accept a series of
numbers in following format "c(1,2,3,4,5)" or other
R code that can generate a vector]
-d Add labels to hline.
[Default same as -j
Accept a series of numbers in following format "c(1,2,3,4,5)" or other R code
that can generate a vector as labels.
Or one can give '1' to disallow labels]
-I Manually set the position of xtics.
[Default FALSE, accept a series of
numbers in following format "c(1,2,3,4,5)" or other R code
that can generate a vector to set the position of xtics]
-b Manually set the value of xtics when -I is specified.
[Default the content of -I when -I is specified,
accept a series of numbers in following format "c(1,2,3,4,5)" or other R code
that can generate a vector to set the position of xtics]
-X Display xtics. [Default TRUE]
-Y Display ytics. [Default TRUE]
-R Rotation angle for x-axis labels (anti clockwise)
[Default 0]
-B line size. [Default 1. Accept a number.]
-t Title of picture[Default empty title]
-x xlab of picture[Default empty xlab]
-y ylab of picture[Default empty ylab]
-c Manually set colors for each line.[Default FALSE, meaning using ggplot2 default.]
-C Color for each line.
When -c is TRUE, one has two options:
1. Supplying a function to generate colors,
like "rainbow(11)" or "rainbow(11, alpha=0.6)",
rainbow is an R color palletes,
11 is the number of colors you want to get,
0.6 is the alpha value.
The R palletes include <heat.colors>, <terrain.colors>,
<topo.colors>, <cm.colors>.
2. Supplying a list of colors in given format,
the number of colors should be equal to the number of
bars like "'red','pink','blue','cyan','green','yellow'" or
"rgb(255/255,0/255,0/255),rgb(255/255,0/255,255/255),
rgb(0/255,0/255,255/255),rgb(0/255,255/255,255/255),
rgb(0/255,255/255,0/255),rgb(255/255,255/255,0/255)"
One can use R fucntion <colors()> to list all available colors.
-s Scale y axis
[Default null. Accept TRUE. This function is depleted.
But if the supplied number after -S is not 0, this parameter will be set to TRUE]
-F The formula for facets.[Default no facets,
"+facet_grid(level ~ .)" means divide by levels of 'level' vertically.
"+facet_grid(. ~ level)" means divide by levels of 'level' horizontally.
"+facet_grid(lev1 ~ lev2)" means divide by lev1 vertically and lev2 horizontally.
"+facet_wrap(~level, ncol=2)" means wrap horizontally with 2 columns.
#Pay attention to the single quote for parameters in function for scale.
Example: "+facet_wrap(~Size,ncol=6,scale='free')"
Example: "+facet_grid(Size ~ .,scale='free_y')"
]
-G If facet is given, you may want to specifize the order of
variable in your facet, default alphabetical order.
[Accept sth like (one level one sentence, separate by';')
'data$size <- factor(data$size, levels=c("l1", "l2",...,"l10"), ordered=T)' ]
-v If scale is TRUE, give the following 'scale_y_log10()'[default], 'coord_trans(y="log10")',
or other legal command for ggplot2 or simply 'log2'.]
-S A number to add if scale is used.
[Default 0. If a non-zero number is given, -s would be set to TRUE.]
-p Other legal R codes for gggplot2 could be given here.
[Begin with '+' ]
-w The width of output picture (cm).[Default 20]
-u The height of output picture (cm).[Default 12]
-E The type of output figures.[Default pdf, accept
eps/ps, tex (pictex), png, jpeg, tiff, bmp, svg and wmf)]
-r The resolution of output picture.[Default 300 ppi]
-z Is there a header. Must be TRUE. [Default TRUE]
-e Execute or not[Default TRUE]
-i Install depended packages[Default FALSE]R绘图学习
为了推广，也为了激起大家的热情，如果想要sp_lines.sh脚本的，还需要劳烦大家动动手，转发此文章到朋友圈，并留言索取。
也希望大家能一起开发，完善功能。