在 circlize 包中，该ChordDiagram()函数只允许一个“from”列、一个“to”列和一个可选的“value”列。但是，在您的情况下，实际上我们可以对原始数据框进行一些转换，以将其修改为三列数据框。
在您的示例中，您想将北美的 Acanthosaura_armata 与欧洲的 Acanthosaura_armata 区分开来，一种解决方案是合并区域名称和物种名称Acanthosaura_armata|North_America以形成唯一标识符。接下来我将演示如何通过 circlize 包来可视化这个数据集。
读入数据。注意我用下划线替换了空格。
df = read.table(textConnection(
"import_region export_region species flow
North_America Europe Acanthosaura_armata 0.0104
Southeast_Asia Europe Acanthosaura_armata 0.0022
Indonesia Europe Acanthosaura_armata 0.1971
Indonesia Europe Acrochordus_granulatus 0.7846
Southeast_Asia Europe Acrochordus_granulatus 0.1101
Indonesia Europe Acrochordus_javanicus 2.00E-04
Southeast_Asia Europe Acrochordus_javanicus 0.0015
Indonesia North_America Acrochordus_javanicus 0.0024
East_Asia Europe Acrochordus_javanicus 0.0028
Indonesia Europe Ahaetulla_prasina 4.00E-04
Southeast_Asia Europe Ahaetulla_prasina 4.00E-04
Southeast_Asia East_Asia Amyda_cartilaginea 0.0027
Indonesia East_Asia Amyda_cartilaginea 5.00E-04
Indonesia Europe Amyda_cartilaginea 0.004
Indonesia Southeast_Asia Amyda_cartilaginea 0.0334
Europe North_America Amyda_cartilaginea 4.00E-04
Indonesia North_America Amyda_cartilaginea 0.1291
Southeast_Asia Southeast_Asia Amyda_cartilaginea 0.0283
Indonesia West_Asia Amyda_cartilaginea 0.7614
South_Asia Europe Amyda_cartilaginea 2.8484
Australasia Europe Apodora_papuana 0.0368
Indonesia North_America Apodora_papuana 0.324
Indonesia Europe Apodora_papuana 0.0691
Europe Europe Apodora_papuana 0.0106
Indonesia East_Asia Apodora_papuana 0.0129
Europe North_America Apodora_papuana 0.0034
East_Asia East_Asia Apodora_papuana 2.00E-04
Indonesia Southeast_Asia Apodora_papuana 0.0045
East_Asia North_America Apodora_papuans 0.0042"),
header = TRUE, stringsAsFactors = FALSE)
另外，我删除了一些值非常小的行。
df = df[df[[4]] > 0.01, ]
为物种和地区分配颜色。
library(circlize)
library(RColorBrewer)
all_species = unique(df[[3]])
color_species = structure(brewer.pal(length(all_species), "Set1"), names = all_species)
all_regions = unique(c(df[[1]], df[[2]]))
color_regions = structure(brewer.pal(length(all_regions), "Set2"), names = all_regions)
按物种分组
首先，我将演示如何按物种对和弦图进行分组。
如前所述，我们species|region用作唯一标识符。
df2 = data.frame(from = paste(df[[3]], df[[1]], sep = "|"),
to = paste(df[[3]], df[[2]], sep = "|"),
value = df[[4]], stringsAsFactors = FALSE)
接下来我们将所有扇区的顺序调整为先按物种排序，然后按地区排序。
combined = unique(data.frame(regions = c(df[[1]], df[[2]]),
species = c(df[[3]], df[[3]]), stringsAsFactors = FALSE))
combined = combined[order(combined$species, combined$regions), ]
order = paste(combined$species, combined$regions, sep = "|")
我们希望链接的颜色与 regoins 的颜色相同
grid.col = structure(color_regions[combined$regions], names = order)
由于弦图是按物种分组的，因此物种之间的差距应该大于每个物种内部的差距。
gap = rep(1, length(order))
gap[which(!duplicated(combined$species, fromLast = TRUE))] = 5
准备好所有设置后，我们现在可以制作和弦图：
在下面的代码中，我们设置preAllocateTracks以便之后添加代表物种的圆形线。
circos.par(gap.degree = gap)
chordDiagram(df2, order = order, annotationTrack = c("grid", "axis"),
grid.col = grid.col, directional = TRUE,
preAllocateTracks = list(
track.height = 0.04,
track.margin = c(0.05, 0)
)
)
添加圆线以表示物种：
for(species in unique(combined$species)) {
l = combined$species == species
sn = paste(combined$species[l], combined$regions[l], sep = "|")
highlight.sector(sn, track.index = 1, col = color_species[species],
text = species, niceFacing = TRUE)
}
circos.clear()
以及地区和物种的传说：
legend("bottomleft", pch = 15, col = color_regions,
legend = names(color_regions), cex = 0.6)
legend("bottomright", pch = 15, col = color_species,
legend = names(color_species), cex = 0.6)
情节如下所示：
按地区分组
代码类似，我就不解释了，只是把代码附在帖子里。情节如下所示：
## group by regions
df2 = data.frame(from = paste(df[[1]], df[[3]], sep = "|"),
to = paste(df[[2]], df[[3]], sep = "|"),
value = df[[4]], stringsAsFactors = FALSE)
combined = unique(data.frame(regions = c(df[[1]], df[[2]]),
species = c(df[[3]], df[[3]]), stringsAsFactors = FALSE))
combined = combined[order(combined$regions, combined$species), ]
order = paste(combined$regions, combined$species, sep = "|")
grid.col = structure(color_species[combined$species], names = order)
gap = rep(1, length(order))
gap[which(!duplicated(combined$species, fromLast = TRUE))] = 5
circos.par(gap.degree = gap)
chordDiagram(df2, order = order, annotationTrack = c("grid", "axis"),
grid.col = grid.col, directional = TRUE,
preAllocateTracks = list(
track.height = 0.04,
track.margin = c(0.05, 0)
)
)
for(region in unique(combined$regions)) {
l = combined$regions == region
sn = paste(combined$regions[l], combined$species[l], sep = "|")
highlight.sector(sn, track.index = 1, col = color_regions[region],
text = region, niceFacing = TRUE)
}
circos.clear()
legend("bottomleft", pch = 15, col = color_regions,
legend = names(color_regions), cex = 0.6)
legend("bottomright", pch = 15, col = color_species, l
egend = names(color_species), cex = 0.6)
