我想为海洋生物学课程制作一个简单的系统发育树作为一个教育的例子。我有一个具有分类等级的物种清单:
Group <- c("Benthos","Benthos","Benthos","Benthos","Benthos","Benthos","Zooplankton","Zooplankton","Zooplankton","Zooplankton",
"Zooplankton","Zooplankton","Fish","Fish","Fish","Fish","Fish","Fish","Phytoplankton","Phytoplankton","Phytoplankton","Phytoplankton")
Domain <- rep("Eukaryota", length(Group))
Kingdom <- c(rep("Animalia", 18), rep("Chromalveolata", 4))
Phylum <- c("Annelida","Annelida","Arthropoda","Arthropoda","Porifera","Sipunculida","Arthropoda","Arthropoda","Arthropoda",
"Arthropoda","Echinoidermata","Chorfata","Chordata","Chordata","Chordata","Chordata","Chordata","Chordata","Heterokontophyta",
"Heterokontophyta","Heterokontophyta","Dinoflagellata")
Class <- c("Polychaeta","Polychaeta","Malacostraca","Malacostraca","Demospongiae","NA","Malacostraca","Malacostraca",
"Malacostraca","Maxillopoda","Ophiuroidea","Actinopterygii","Chondrichthyes","Chondrichthyes","Chondrichthyes","Actinopterygii",
"Actinopterygii","Actinopterygii","Bacillariophyceae","Bacillariophyceae","Prymnesiophyceae","NA")
Order <- c("NA","NA","Amphipoda","Cumacea","NA","NA","Amphipoda","Decapoda","Euphausiacea","Calanioda","NA","Gadiformes",
"NA","NA","NA","NA","Gadiformes","Gadiformes","NA","NA","NA","NA")
Species <- c("Nephtys sp.","Nereis sp.","Gammarus sp.","Diastylis sp.","Axinella sp.","Ph. Sipunculida","Themisto abyssorum","Decapod larvae (Zoea)",
"Thysanoessa sp.","Centropages typicus","Ophiuroidea larvae","Gadus morhua eggs / larvae","Etmopterus spinax","Amblyraja radiata",
"Chimaera monstrosa","Clupea harengus","Melanogrammus aeglefinus","Gadus morhua","Thalassiosira sp.","Cylindrotheca closterium",
"Phaeocystis pouchetii","Ph. Dinoflagellata")
dat <- data.frame(Group, Domain, Kingdom, Phylum, Class, Order, Species)
dat
我想获得树状图(聚类分析)并使用Domain作为第一个切割点,Kindom作为第二个切割点,Phylum作为第三个切割点,等等。缺失值应该被忽略(没有切割点,而是直线) 。组应该用作标签的着色类别。
我有点不确定如何从这个数据帧中建立距离矩阵。 R有很多系统发育树包,他们似乎想要新的数据/ DNA /其他高级信息。因此,对此的帮助将不胜感激。
答案 0 :(得分:6)
回答我自己的问题可能有点蹩脚,但我找到了一个更简单的解决方案。也许有一天会帮助别人。
library(ape)
taxa <- as.phylo(~Kingdom/Phylum/Class/Order/Species, data = dat)
col.grp <- merge(data.frame(Species = taxa$tip.label), dat[c("Species", "Group")], by = "Species", sort = F)
cols <- ifelse(col.grp$Group == "Benthos", "burlywood4", ifelse(col.grp$Group == "Zooplankton", "blueviolet", ifelse(col.grp$Group == "Fish", "dodgerblue", ifelse(col.grp$Group == "Phytoplankton", "darkolivegreen2", ""))))
plot(taxa, type = "cladogram", tip.col = cols)
请注意,所有列都必须是因子。这演示了使用R的工作流程。虽然代码本身只是几行,但需要一周时间才能找到一些东西=)
答案 1 :(得分:3)
如果你想手工绘制树 (这可能不是最好的方法), 你可以从以下开始 (这不是一个完整的答案: 颜色缺失, 并且边缘太长)。 这假定数据已经排序。
# Data: remove Group
dat <- data.frame(Domain, Kingdom, Phylum, Class, Order, Species)
# Start a new plot
par(mar=c(0,0,0,0))
plot(NA, xlim=c(0,ncol(dat)+1), ylim=c(0,nrow(dat)+1),
type="n", axes=FALSE, xlab="", ylab="", main="")
# Compute the position of each node and find all the edges to draw
positions <- NULL
links <- NULL
for(k in 1:ncol(dat)) {
y <- tapply(1:nrow(dat), dat[,k], mean)
y <- y[ names(y) != "NA" ]
positions <- rbind( positions, data.frame(
name = names(y),
x = k,
y = y
))
}
links <- apply( dat, 1, function(u) {
u <- u[ !is.na(u) & u != "NA" ]
cbind(u[-length(u)],u[-1])
} )
links <- do.call(rbind, links)
rownames(links) <- NULL
links <- unique(links[ order(links[,1], links[,2]), ])
# Draw the edges
for(i in 1:nrow(links)) {
from <- positions[links[i,1],]
to <- positions[links[i,2],]
lines( c(from$x, from$x, to$x), c(from$y, to$y, to$y) )
}
# Add the text
text(positions$x, positions$y, label=positions$name)