将传说按较高分类、菌丝和属进行分组？ggplot2 [英] Grouping legend by higher classification, filum and genus? ggplot2

问题描述

我想改进图例()

我希望根据更高的分类(门)对图例进行分组，但同时显示属(属)。

或相等，但仅选择每个菌丝中含量最丰富的20个属

我想要这样的东西： Im try to make something like this

我正在运行此代码：

ggplot(d4) + 
  geom_bar(aes(x=Pacientes, y=`Relative abundance`,fill=Genus), position="fill", stat="identity") +     scale_x_discrete("Patients") + 
  scale_y_continuous("Relative abundance",labels=scales::percent) +
  labs(title = "CAP",subtitle = "Relative abundance of phylum and genus") +
  theme_classic() + 
  scale_fill_manual(values=c ("#FC000D", "#30E500", "#E10072", "#730183",       "#B58E2C","#10A542","#6C1429",
  "#00B9B9", "#E36582","orange3","#800009",
  "#5E230B","#CC6187","#949285","#FF6A00",
  "#FF9D69","#B08A04","#005A3F","#120A5F","#E7BECD")) 

output

示例数据

d4<-structure(list(Pacientes = c("5-006", "5-005", "5-005", "5-001", 
"5-003", "5-002", "5-001", "5-001", "5-005", "5-001", "5-003", 
"5-003", "5-007", "5-006", "5-003", "5-001", "5-002", "5-003", 
"5-002", "5-002", "5-001", "5-002", "5-003", "5-005", "5-002", 
"5-001", "5-006", "5-005", "5-007", "5-005"), Filum = c("Firmicutes", 
"Firmicutes", "Firmicutes", "Firmicutes", "Firmicutes", "Firmicutes", 
"Firmicutes", "Firmicutes", "Proteobacteria", "Proteobacteria", 
"Proteobacteria", "Proteobacteria", "Proteobacteria", "Proteobacteria", 
"Proteobacteria", "Proteobacteria", "Proteobacteria", "Proteobacteria", 
"Spirochaetes", "Spirochaetes", "Spirochaetes", "Spirochaetes", 
"Firmicutes", "Firmicutes", "Bacteroidetes", "Bacteroidetes", 
"Bacteroidetes", "Bacteroidetes", "Bacteroidetes", "Firmicutes"
), Genus = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 12L, 13L, 
13L, 13L, 13L, 13L, 14L, 15L, 15L, 15L, 15L, 20L, 20L, 20L, 20L, 
21L, 21L, 25L, 25L, 25L, 25L, 25L, 26L), .Label = c("Streptococcus", 
"Veillonella", "Haemophilus", "Actinobacillus", "Serratia", "Fusobacterium", 
"Neisseria", "Moraxella", "Abiotrophia", "Granulicatella", "Actinomyces", 
"Oribacterium", "Aggregatibacter", "Escherichia-Shigella", "Lautropia", 
"Geobacillus", "Leptotrichia", "Johnsonella", "Campylobacter", 
"Treponema 2", "Gemella", "Megasphaera", "Atopobium", "Bifidobacterium", 
"Capnocytophaga", "Selenomonas", "Mycoplasma", "Porphyromonas", 
"Alloprevotella", "Lachnoanaerobaculum", "Eikenella", "[Eubacterium] brachy group", 
"Stomatobaculum", "Atopostipes", "Selenomonas 3", "Kingella", 
"Dialister", "F0058", "Parvimonas", "No identificado", "Solobacterium", 
"Otros finales", "Olsenella", "Filifactor", "Rodentibacter", 
"Alloscardovia", "Otros", "[Eubacterium] yurii group", "Anaeroglobus", 
"Staphylococcus", "Ruminococcaceae UCG-014", "Lactobacillus", 
"Rothia", "Selenomonas 4", "Scardovia", "Fluviicola", "Cardiobacterium", 
"Bilophila", "Simonsiella", "[Eubacterium] nodatum group", "Catonella", 
"Peptoniphilus", "uncultured", "Shuttleworthia", "Butyrivibrio 2", 
"Peptostreptococcus", "Mogibacterium", "Bergeyella", "Peptococcus", 
"Faucicola", "Blautia", "Rikenellaceae RC9 gut group", "Eggerthia", 
"Desulfobulbus", "Tannerella", "Lactococcus", "Prevotella", "Otro", 
"Prevotella 7", "Erysipelotrichaceae UCG-006", "Defluviitaleaceae UCG-011", 
"W5053", "Craurococcus", "Dolosigranulum", "Sneathia", "Anaerococcus", 
"Pseudoramibacter", "Family XIII UCG-001", "Ruminococcus 2", 
"Howardella", "Cryptobacterium", "Listeria", "Pantoea", "Akkermansia", 
"Prevotella 6", "Macrococcus", "Paracoccus", "Comamonas", "TM7 phylum sp. oral clone FR058", 
"Peptoanaerobacter", "Rubellimicrobium", "Fastidiosipila", "Brachymonas", 
"Candidatus Tammella", "Slackia", "DNF00809", "Truepera", "Finegoldia", 
"Erysipelotrichaceae UCG-004", "uncultured bacterium", "Bulleidia", 
"Flexilinea", "Methylobacterium", "Propionivibrio", "Ochrobactrum"
), class = "factor"), `Relative abundance` = c(1.797989737427, 
1.17051056033446, 0.967773967968912, 0.890190018788368, 0.875168325944855, 
0.203636768715721, 0.190978038791412, 0.0011814814596022, 0.0411830680204194, 
0.031674955321716, 0.018003527003462, 0.00559797167763897, 0.0032912697803204, 
0.0189880948864639, 0.0386794525465004, 0.0337566131314913, 0.0279054668553661, 
0.0242766309437308, 0.000168783065657456, 0.000112522043771638, 
8.43915328287282e-05, 8.43915328287282e-05, 0.0290306872930825, 
0.0263020277316203, 0.00458527328369423, 0.00393827153200732, 
0.00351631386786368, 0.00210978832071821, 0.000900176350173101, 
0.021744884958869)), row.names = c(NA, -30L), class = c("tbl_df", 
"tbl", "data.frame"))

分组列为filum，正常图例为数据中的属。

谢谢

更新

我正在尝试完整数据，但图例太大，请添加guides(fill=guide_legend(ncol=3, byrow=TRUE))，但功能不佳。

*更新2

跑得很好。但为了做得更好，我需要为每个组添加比例颜色，如下所示：如果您看到Firmicuts有很多单词，就是为每个组添加比例颜色。所以，可能是红色鳞片的菲米米特，蓝色鳞片的变形杆菌

更新3 想法是在栏中只显示一种颜色(这很好)。我需要按图例中的每个细丝标明颜色比例。

最终更新

所有数据的最终绘图如下：

我将每个名称放入每个名称中，仅修改COLS对象。

cols <- c(Streptococcus="#EE0600",Veillonella= "#FFE200", Haemophilus="#5EE200", Actinobacillus= "#3A0B88", Serratia= "#9E8D00", Fusobacterium= "#8EAC55", Neisseria=  "#F5B800",Moralexa= "#900027", Abiotrophia ="#EB2B5F", Granulicatella= "#990015", Actinomyces= "#009876",
Oriobacterium= "#F9342F", Agreggatibacter= "#CBF377",Escherichia_Shigenella ="#DE5FA4",Lautropia= "#617F29",Geobacillus ="#FF7061",Leptotrichia= "#B94700", Johnsonella= "#FF2811", Campylobacter= "#054474",Treponema_2=  "#FFAF43")

非常感谢

推荐答案

实现所需效果的一个选项是通过ggnewscale程序包，该程序包允许使用多个比例和图例来实现相同的美学效果。

1. 将您的颜色放入命名矢量中，该矢量将为您的Genus
分配一种颜色
2. 列出Filum与关联的Genus%s。为此，我使用dplyr::distinct和split。

library(ggplot2)
library(ggnewscale)
library(dplyr)

cols <- c("#FC000D", "#30E500", "#E10072", "#730183",
"#B58E2C", "#10A542", "#6C1429",
"#00B9B9", "#E36582", "orange3", "#800009",
"#5E230B", "#CC6187", "#949285", "#FF6A00",
"#FF9D69", "#B08A04", "#005A3F", "#120A5F", "#E7BECD")
cols <- rep_len(cols, length.out = length(levels(d4$Genus)))
names(cols) <- levels(d4$Genus)

groups <- d4 %>% 
  distinct(Filum, Genus) %>% 
  # Add order of Filum and legends
  mutate(order = as.numeric(forcats::fct_inorder(Filum))) %>% 
  split(.$Filum)

3. 对于每个Filum，添加一个显示整个数据的geom_col和一个scale_fill_manual，其中使用limits参数，我们仅显示与此Filum相关联的Genus。执行此操作后，所有其他Genus将分配给我们为其选择transparent颜色的NA。

要对我们在这里所做的事情有一个基本概念，只需两个组的代码：

ggplot(d4) +
  geom_col(aes(x = Pacientes, y=`Relative abundance`, fill = Genus), position = "fill") +
  scale_fill_manual(values = cols, limits = groups$Firmicutes$Genus, na.value = "transparent") +
  new_scale_fill() +
  geom_col(aes(x = Pacientes, y=`Relative abundance`, fill = Genus), position = "fill") +
  scale_fill_manual(values = cols, limits = groups$Bacteroidetes$Genus, na.value = "transparent")

我们可以使用lapply遍历这些组并动态添加层，而不是复制和粘贴(如果我们有很多组，这会变得很麻烦)：

  
ggplot(d4) +
  lapply(groups, function(x) {
    list(
      geom_col(aes(x = Pacientes, y=`Relative abundance`, fill = Genus), position = "fill"),
      scale_fill_manual(name = unique(x$Filum),
                        values = cols, limits = x$Genus, na.value = "transparent",
                        guide = guide_legend(order = unique(x$order))),
      new_scale_fill()  
    )
  }) +
  scale_x_discrete("Patients") +
  scale_y_continuous("Relative abundance", labels = scales::percent) +
  labs(title = "CAP", subtitle = "Relative abundance of phylum and genus") +
  theme_classic()

编辑如果要创建包含多列的图例，可以通过guide_legend中的guide_legend进行编辑。因为我们现在有多个图例，所以使用guides(fill = ....)将不起作用。

ggplot(d4) +
  lapply(groups, function(x) {
    list(
      geom_col(aes(x = Pacientes, y=`Relative abundance`, fill = Genus), position = "fill"),
      scale_fill_manual(name = unique(x$Filum),
                        values = cols, limits = x$Genus, na.value = "transparent",
                        guide = guide_legend(order = unique(x$order), ncol = 3, byrow = TRUE)),
      new_scale_fill()  
    )
  }) +
  scale_x_discrete("Patients") +
  scale_y_continuous("Relative abundance", labels = scales::percent) +
  labs(title = "CAP", subtitle = "Relative abundance of phylum and genus") +
  theme_classic()

这篇关于将传说按较高分类、菌丝和属进行分组？ggplot2的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！