重命名大型图形网络中的图形对象

Question

我有一个包含50个网络信息的large igraph，如下所示：

Source                              Target
STAR-delta                          c-Lgt-1
SUN-2 receptor                      c-Jun
H2R reg class IA (p7)               STAR-delta
PtdIns(4,5)P2 intracellular         PtdIns(3,4,5)P3 intracellular
ATP + PtdIns(4,5)P2                 ATP + PtdIns(4,5)P2 

另一方面，我有一个数据集(x)，其中包含我的网络的ID和符号。

Object.Name                         ID                Symbol
STAR-delta                         5580               PRKCD
SUN-2 receptor                     1439               CSF2RB
H2R reg class IA (p7)              5295               PIK3R1
PtdIns(4,5)P2 intracellular        No ID              PtdIns(4,5)P2 intracellular
ATP + PtdIns(4,5)P2                No ID              ATP + PtdIns(4,5)P2 
c-Lgt-1                            5894               RAF1
c-Jun                              3725               JUN
STAR-delta                         5580               PRKCD
PtdIns(3,4,5)P3 intracellular      No ID              PtdIns(3,4,5)P3 intracellular
ATP + PtdIns(4,5)P2                No ID              ATP + PtdIns(4,5)P2 

因此，我想将顶点名称更改为符号(以ID作为属性)。我试着用

V(g)[V(g)$name == "x$Object.Name"]$Symbol <- x$Symbol

但它不起作用。我希望得到这样的输出：

Source                                Target
PRKCD                                 RAF1
CSF2RB                                JUN
PIK3R1                                PRKCD
PtdIns(4,5)P2 intracellular           PtdIns(3,4,5)P3 intracellular
ATP + PtdIns(4,5)P2                   ATP + PtdIns(4,5)P2 

注意: dataset x和g的顶点名称顺序不同。

Answer 1

一种选择是基于V(g)$name和x$Object.Name的x$Symbol子集。若要创建顶点属性，可以使用set_vertex_attr()。尝尝这个

links <- read.table(text = "Source                              Target
                    STAR-delta                          c-Lgt-1
                    'SUN-2 receptor'                      c-Jun
                    'H2R reg class IA (p7)'               STAR-delta
                    'PtdIns(4,5)P2 intracellular'         'PtdIns(3,4,5)P3 intracellular'
                    'ATP + PtdIns(4,5)P2'                 'ATP + PtdIns(4,5)P2'",
                    header = TRUE)

x <- read.table(text = "Object.Name                         ID                Symbol
                'PtdIns(4,5)P2 intracellular'        'No ID'              'PtdIns(4,5)P2 intracellular'
                'SUN-2 receptor'                     1439               CSF2RB
                'H2R reg class IA (p7)'              5295               PIK3R1
                c-Jun                              3725               JUN
                'ATP + PtdIns(4,5)P2'                'No ID'              'ATP + PtdIns(4,5)P2' 
                c-Lgt-1                            5894               RAF1
                STAR-delta                         5580               PRKCD
                'PtdIns(3,4,5)P3 intracellular'      'No ID'              'PtdIns(3,4,5)P3 intracellular'
                'ATP + PtdIns(4,5)P2'                'No ID'              'ATP + PtdIns(4,5)P2'
                STAR-delta                         5580               PRKCD",
                header = TRUE)

## create network
library(igraph)
g <- graph_from_data_frame(links, directed = TRUE, vertices = unique(x$Object.Name))
V(g)$name # vertex names
E(g) # edge list

## change vertex names
library(dplyr)
x <- distinct(x) # get unique rows of x
unique(x$Symbol[V(g)$name %in% x$Object.Name]) # this will subset x$Symbol
# based on vertex names. This vector has then the same length with V(g)$name 
V(g)$name <- as.character(unique(x$Symbol[V(g)$name %in% x$Object.Name]))
V(g)$name # vertex names
E(g) # edge list. Expected output

## set vertex attributes
vertex_attr(g) # only name as vertex attribute
# x[match(V(g)$name, x$Symbol), ] not useful here. Can be used to reoder rows
g <- set_vertex_attr(g, "ID", index = V(g), as.vector(x$ID))
vertex_attr(g) # with name and ID as vertex attributes

Answer 2

这是您的数据

edge_igraph <- igraph::graph_from_literal(
  "STAR-delta" - "c-Lgt-1",
  "SUN-2 receptor" - "c-Jun",
  "H2R reg class IA (p7)" -  "STAR-delta",
  "PtdIns(4,5)P2 intracellular"- "PtdIns(3,4,5)P3 intracellular",
  "ATP + PtdIns(4,5)P2" - "ATP + PtdIns(4,5)P2")

x_df <- tibble::tribble(
                    ~Object.Name,      ~ID,                          ~Symbol,
                    "STAR-delta",   "5580",                          "PRKCD",
                "SUN-2 receptor",   "1439",                         "CSF2RB",
         "H2R reg class IA (p7)",   "5295",                         "PIK3R1",
   "PtdIns(4,5)P2 intracellular",  "No ID",    "PtdIns(4,5)P2 intracellular",
           "ATP + PtdIns(4,5)P2",  "No ID",            "ATP + PtdIns(4,5)P2",
                       "c-Lgt-1",   "5894",                           "RAF1",
                         "c-Jun",   "3725",                            "JUN",
                    "STAR-delta",   "5580",                          "PRKCD",
 "PtdIns(3,4,5)P3 intracellular",  "No ID",  "PtdIns(3,4,5)P3 intracellular",
           "ATP + PtdIns(4,5)P2",  "No ID",            "ATP + PtdIns(4,5)P2"
            )

这是一个使用tidygraph的tidyverse解决方案

library(tidygraph)

mygraph <- as_tbl_graph(edge_igraph) %>% 
  graph_join(tbl_graph(nodes=distinct(x_df)), 
                 by=c("name"="Object.Name")) %>% 
  select(name=Symbol, ID)

您可以使用ggraph进一步可视化它

library(ggraph)

ggraph(mygraph, layout='mds')+
  geom_node_label(aes(label=name))+
  geom_edge_link(aes(start_cap = circle(8, 'mm'),
                     end_cap = circle(8, 'mm')),
                 arrow = arrow(length = unit(3, 'mm')), 
                 width=1.5, colour="dodgerblue")+
  theme_void()

​

​