[翻译稿]我是如何创建人体解剖医学包-gganatogram的
原文来源 https://jespermaag.github.io/blog/2018/gganatogram/
本文为翻译版，不当之处请见谅！
GGANATOGREM
https://github.com/jespermaag/gganatogram
✎✎✎
希望可以为不同的生物创建解剖图像，但是目前只有人类男性可用。
在看到ggseg的twitter帖子之后，我有了这个包的想法。类似的工具对整个生物学都会有帮助。由于找不到任何类似的东西，我决定创建我的第一个R包。
该软件包使用ArrayExpress Expression Atlas图中的组织坐标。
https://www.ebi.ac.uk/gxa/home
https://github.com/ebi-gene-expression-group/anatomogram
【创建包】
下载所有svg
为了创建包，我首先必须从Expression Atlas中检索所有组织的坐标。
使用以下命令下载解剖图包。
npm install --save anatomogram
从svg中提取坐标
我使用python来提取homo_sapiens.mal.svg文件中每个组织的坐标，名称和转换。
以下代码获取svg，并将名称，坐标和转换写入文件，然后在R中处理。
from xml.dom import minidom
import os
import csv
organism="homo_sapiens.male"
doc = minidom.parse(organism + ".svg")
your_csv_file = open(organism + '_coords.tsv', 'w')
wr = csv.writer(your_csv_file, delimiter='\t')
for path in doc.getElementsByTagName('path'):
if "outline" in path.getAttribute('id') or "LAYER_OUTLINE" in path.getAttribute('id') :
wr.writerow([path.getAttribute('id') ,path.getAttribute('d'), str('matrix(1,0,0,1,0,0)')])
if path.getAttribute('id').startswith('UB'):
wr.writerow([path.getElementsByTagName('title')[0].firstChild.nodeValue, path.getAttribute('d'), str('matrix(1,0,0,1,0,0)')])
if path.parentNode.attributes['id'].value.startswith('UB'):
if "transform" not in list(path.parentNode.attributes.keys()):
wr.writerow([path.parentNode.attributes['id'].value, path.getAttribute('d'), str('matrix(1,0,0,1,0,0)')])
for path in doc.getElementsByTagName('g')[5:]:
if len(path.childNodes) >0 :
for node in path.childNodes:
if "text" not in node.nodeName:
print(node.nodeName)
print(node.attributes.keys())
if 'd' in list(node.attributes.keys()):
nodeVal = node.attributes['d'].value
wr.writerow([path.childNodes[1].attributes['id'].value, nodeVal, path.attributes['transform'].value])
your_csv_file.close()
处理R中的坐标，并创建一个包
首先，我创建了一个函数，将坐标提取到数据框中并转换数据。需要进行一些手动编辑才能获得正确的坐标，并删除一些不起作用的组织。
extractCoords <- function(coords, name, transMatrix) {
c <- strsplit(coords, " ")
c[[1]]
c[[1]][c(grep("M", c[[1]] )+1,grep("M", c[[1]] )+2)] <- NA
c[[1]] <- c[[1]][grep("[[:alpha:]]", c[[1]], invert=TRUE)]
anatCoord <- as.data.frame(lapply( c, function(u)
matrix(as.numeric(unlist(strsplit(u, ","))),ncol=2,byrow=TRUE) ))
anatCoord$X2[is.na(anatCoord$X1)] <- NA
anatCoord$X1[is.na(anatCoord$X2)] <- NA
anatCoord$id <- name
if (length(transMatrix[grep('matrix', transMatrix)])>0) {
transForm <- gsub('matrix\\(|\\)', '', transMatrix)
transForm <- as.numeric(strsplit(transForm, ",")[[1]])
anatCoord$x <- (anatCoord$X1* transForm[1]) + (anatCoord$X1* transForm[3]) + transForm[5]
anatCoord$y <- (anatCoord$X2* transForm[2]) + (anatCoord$X2* transForm[4]) + transForm[6]
} else if (grep('translate', transMatrix)) {
transForm <- gsub('translate\\(|\\)', '', transMatrix)
transForm <- as.numeric(strsplit(transForm, ",")[[1]])
if(name =='leukocyte' & transForm[1]==4.5230265) {
transForm <- c(103.63591+4.5230265,-47.577078+11.586659)
}
anatCoord$x <- anatCoord$X1 + transForm[1]
anatCoord$y <- anatCoord$X2 + transForm[2]
}
#anatCoord <- anatCoord[complete.cases(anatCoord),]
if (name == 'bronchus') {
if (max(anatCoord$x, na.rm=T) >100 ) {
anatCoord$x <- NA
anatCoord$y <- NA
}
}
if( any(anatCoord[complete.cases(anatCoord),]$x < -5)) {
anatCoord$x <- NA
anatCoord$y <- NA
}
if( any(anatCoord[complete.cases(anatCoord),]$x > 150)) {
anatCoord$x <- NA
anatCoord$y <- NA
}
return(anatCoord)
}
最后，用extractCoords函数处理了python输出。
hsMale <- read.table('homo_sapiens.male_coords.tsv', sep='\t', stringsAsFactors=F)
hgMale_list <- list()
for (i in 1:nrow(hsMale)) {
df <- extractCoords(hsMale$V2[i], hsMale$V1[i], hsMale$V3[i])
hgMale_list[[i]] <- extractCoords(hsMale$V2[i], hsMale$V1[i], hsMale$V3[i])
names(hgMale_list)[i] <- paste0(hsMale$V1[i],'-', i)
}
names(hgMale_list) <- gsub('-.*', '', names(hgMale_list))
之后，将结果列表用作gganatogram包的基础。可以使用以下说明从github安装该软件包。
【安装】
使用devtools从github安装。
# install from Github
devtools::install_github("jespermaag/gganatogram")
【用法】
这个包需要ggplot2和ggpolypath
library(ggplot2)
library(ggpolypath)
library(gganatogram)
library(dplyr)
并且需要使用函数gganatogram，您需要拥有一个包含器官组织，颜色和数值的数据框。
organPlot <- data.frame(organ = c("heart", "leukocyte", "nerve", "brain", "liver", "stomach", "colon"),
type = c("circulation", "circulation", "nervous system", "nervous system", "digestion", "digestion", "digestion"),
colour = c("red", "red", "purple", "purple", "orange", "orange", "orange"),
value = c(10, 5, 1, 8, 2, 5, 5),
stringsAsFactors=F)
head(organPlot)
## organ type colour value
## 1 heart circulation red 10
## 2 leukocyte circulation red 5
## 3 nerve nervous system purple 1
## 4 brain nervous system purple 8
## 5 liver digestion orange 2
## 6 stomach digestion orange 5
使用函数gganatogram，根据颜色填充器官。
gganatogram(data=organPlot, fillOutline='#a6bddb', organism='human', sex='male', fill="colour")
然后，我们可以使用ggplot主题和函数来调整图上细节
gganatogram(data=organPlot, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +
theme_void()
我们还可以使用hgMale_key绘制所有可用组织，这是一个可用的对象
hgMale_key$organ
## [1] "bone marrow" "frontal cortex"
## [3] "prefrontal cortex" "gastroesophageal junction"
## [5] "caecum" "ileum"
## [7] "rectum" "nose"
## [9] "tongue" "penis"
## [11] "nasal pharynx" "spinal cord"
## [13] "throat" "diaphragm"
## [15] "liver" "stomach"
## [17] "spleen" "duodenum"
## [19] "gall bladder" "pancreas"
## [21] "colon" "small intestine"
## [23] "appendix" "urinary bladder"
## [25] "bone" "cartilage"
## [27] "esophagus" "skin"
## [29] "brain" "heart"
## [31] "lymph_node" "skeletal_muscle"
## [33] "leukocyte" "temporal_lobe"
## [35] "atrial_appendage" "coronary_artery"
## [37] "hippocampus" "vas_deferens"
## [39] "seminal_vesicle" "epididymis"
## [41] "tonsil" "lung"
## [43] "trachea" "bronchus"
## [45] "nerve" "kidney"
gganatogram(data=hgMale_key, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +theme_void()
要跳过图表的轮廓，请使用outline = F
organPlot %>%
dplyr::filter(type %in% c('circulation', 'nervous system')) %>%
gganatogram(outline=F, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +
theme_void()
我们可以根据给予每个器官的值来填充组织
gganatogram(data=organPlot, fillOutline='#a6bddb', organism='human', sex='male', fill="value") +
theme_void() +
scale_fill_gradient(low = "white", high = "red")
我们也可以使用facet_wrap来比较组。
首先创建两个数据框以及设置类型列中的不同数值和条件。
compareGroups <- rbind(data.frame(organ = c("heart", "leukocyte", "nerve", "brain", "liver", "stomach", "colon"),
colour = c("red", "red", "purple", "purple", "orange", "orange", "orange"),
value = c(10, 5, 1, 8, 2, 5, 5),
type = rep('Normal', 7),
stringsAsFactors=F),
data.frame(organ = c("heart", "leukocyte", "nerve", "brain", "liver", "stomach", "colon"),
colour = c("red", "red", "purple", "purple", "orange", "orange", "orange"),
value = c(5, 5, 10, 8, 2, 5, 5),
type = rep('Cancer', 7),
stringsAsFactors=F))
gganatogram(data=compareGroups, fillOutline='#a6bddb', organism='human', sex='male', fill="value") +
theme_void() +
facet_wrap(~type) +
scale_fill_gradient(low = "white", high = "red")
gganatogram(data=hgMale_key, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +
theme_void() +
facet_wrap(~type)
gganatogram(data=hgMale_key, outline=F, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +
theme_void() +
facet_wrap(~type, scale='free')
organtype <- organPlot
organtype %>%
mutate(type=organ) %>%
gganatogram( outline=F, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +
theme_void() +
facet_wrap(~type, scale='free')
／End.
点击文末的阅读原文链接直达土豆的简书博客！