gganatogram|gganatogram 人体解剖医学包 gganatogram人体解剖医学包

原文来源 https://jespermaag.github.io/blog/2018/gganatogram/
本文为翻译版，不当之处请见谅！
gganatogram https://github.com/jespermaag/gganatogram

文章图片
image.png
【gganatogram|gganatogram 人体解剖医学包】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，我想到了这个包的想法。类似的工具对整个生物学都有帮助。由于找不到任何类似的东西，决定创建我的第一个R包。
该软件包使用ArrayExpress Expression Atlas中图中的组织坐标。
https://www.ebi.ac.uk/gxa/home
https://github.com/ebi-gene-expression-group/anatomogram
生成包下载所有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。此代码获取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
我创建了一个函数来将坐标提取到数据框中并转换数据。需要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)] <- NAc[[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$id <- nameif (length(transMatrix[grep('matrix', transMatrix)])>0) { transForm <- gsub('matrix\$|\$', '', transMatrix) transForm <- as.numeric(strsplit(transForm, ",")[[1]])anatCoord$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))

然后将结果列表用作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)

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 = https://www.it610.com/article/c(10, 5, 1, 8, 2, 5, 5), stringsAsFactors=F)head(organPlot)

##organtype colour value ## 1heartcirculationred10 ## 2 leukocytecirculationred5 ## 3nerve nervous system purple1 ## 4brain nervous system purple8 ## 5liverdigestion orange2 ## 6stomachdigestion orange5

使用函数gganatogram，根据颜色填充器官。

gganatogram(data=https://www.it610.com/article/organPlot, fillOutline='#a6bddb', organism='human', sex='male', fill="colour")

文章图片
image.png 我们可以使用ggplot主题和函数来调整图

gganatogram(data=https://www.it610.com/article/organPlot, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") + theme_void()

文章图片
image.png 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=https://www.it610.com/article/hgMale_key, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") +theme_void()

文章图片
image.png 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()

文章图片
image.png 我们可以根据给予每个器官的值来填充组织

gganatogram(data=https://www.it610.com/article/organPlot, fillOutline='#a6bddb', organism='human', sex='male', fill="value") + theme_void() + scale_fill_gradient(low = "white", high = "red")

文章图片
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 = https://www.it610.com/article/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 = https://www.it610.com/article/c(5, 5, 10, 8, 2, 5, 5), type = rep('Cancer', 7), stringsAsFactors=F))

gganatogram(data=https://www.it610.com/article/compareGroups, fillOutline='#a6bddb', organism='human', sex='male', fill="value") + theme_void() + facet_wrap(~type) + scale_fill_gradient(low = "white", high = "red")

文章图片
image.png

gganatogram(data=https://www.it610.com/article/hgMale_key, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") + theme_void() + facet_wrap(~type)

文章图片
image.png

gganatogram(data=https://www.it610.com/article/hgMale_key, outline=F, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") + theme_void() + facet_wrap(~type, scale='free')

文章图片
image.png

organtype <- organPlot organtype %>% mutate(type=organ) %>% gganatogram( outline=F, fillOutline='#a6bddb', organism='human', sex='male', fill="colour") + theme_void() + facet_wrap(~type, scale='free')