How do you create a table with each segment of the Venn diagram

[jane-murphy]

Hello

Thank you for your code - it is very nice. I am using the code to look at 3 gene sets with the overlapping and unique genes represented by the Venn diagram. I have used show_elements to look at these genes directly on the Venn diagram but I am wondering whether there is a convenient way to have them in a table?

I am especially interested in a way to view the unique genes in each set (i.e. not overlapping with either of the other 2 sets).

Best wishes,
Jane

[mj163163]

Hope he can add that function to export the data.

[lucaskbobadilla]

You can just use dplyr::setdiff in the list of elements you used to create the plot

[Dragonmasterx87]

I dont understand why the authors didnt outline how to do this as the whole purpose of the venn dagram is to find associations across gene sets. Anyway here's a more appropriate solution than arbitrarily looking at intersections such as dplyr::setdiff

# You have made a list of associations called x
x <- list(
 a = data_a,
 b = data_b,
 c = data_c
)

venn <- Venn(x)
data <- process_data(venn) # data is a ggVennDiagram object

# Look at all sets of genes forming overlaps
mylist <- data@region[["item"]]

# These will result in numerically named lists. This is stupid, because its difficult to match a list to the 
# Venn diagram unless you count the values which is also stupid. So the more appropriate way of doing this is:
names(mylist) # this will return a null because no names exist
names(mylist) <- data@region[["name"]]
names(mylist)
mylist  #named list of association type and genes

Enjoy!
🐉

[yanlinlin82]

Hi, @Dragonmasterx87

Thank you for the workaround code. I haven't fixed this issue because I didn't quite understand the problem (or had no idea of a better design) and had insufficient time to code and test. If you don't mind, please send me a pull request, and I will merge it into the package.

[xiasijian]

I have a try, maybe it is not perfect. This is my R code below.

## ggvenn data
export_ggvenn_data=function(venn_list,save_name){
  max_length=c()
  exclude_list=list()
  for(i in names(venn_list)){
    set1=venn_list[[i]]
    other_name=names(venn_list[-which(names(venn_list)==i)])
    other_value=c()
    for(j in other_name){
      #print(j)
      set2=venn_list[[j]]
      other_value=c(other_value,set2)
        
      ## complementary set
      set1_comple=setdiff(set1,set2)
      exclude_name=paste0(i,"_setdiff_",j)
      exclude_list[[exclude_name]]=set1_comple
      
      ## export length in order to cbind
      max_length=c(max_length,length(set1_comple))
      
    } 
      set1_exclude=setdiff(set1,other_value)
      unique_name=paste0(i,"_unique")
      exclude_list[[unique_name]]=set1_exclude
  }
  
  ## intersect
  overlap_set=Reduce(intersect,venn_list)

  ## if the intersected result is zero,give a element.
  if(length(overlap_set)==0){
    exclude_list[["overlap"]]="zero"
  }else{
    exclude_list[["overlap"]]=overlap_set
  }
  
  
  for(i in 1:length(names(exclude_list))){
    if(i ==1 ){
      
      vec1=exclude_list[[i]]
      length(vec1)=max(max_length)
    }else{
      vec2=exclude_list[[i]]
      length(vec2)=max(max_length)
      vec1=cbind(vec1,vec2)
    }
    
  }
  colnames(vec1)=names(exclude_list)
  vec1=as.data.frame(vec1)
  write.csv(vec1,file=paste0(save_name,"_ggvenn_overlap_data.csv"),row.names=F)
  
}

the venn_list is the input of ggvenn function.

[xiasijian]

Recently I wrote a simple function for this:

venny_table = function(set_list,list_name){
  names(setlist) = list_name
  len_list = length(list_name)
  
  ## all overlap items----------------------------------
  
  overlap_items = Reduce(intersect,
                         list(set1,set2,set3))
  
  intersect_setdiff_for_two_sets = function(set1, set2){
    common = intersect(set1,set2)
    set12 = setdiff(set1,set2)
    set21 = setdiff(set2,set1)
    
    common_save = paste(common,collapse = ";")
    set12_save = paste(set12,collapse = ";")
    set21_save = paste(set21, collapse = ";")
    out = c(set12_save,set21_save,common_save)
    return(out)
  }
  
  
  get_pair_type = function(j,len_list){
    pairs = c()
    for(i in (j+1):len_list){
      
      pair = c(list_name[j],list_name[i])
      pair_name = paste(pair,collapse = "_and_")
      pairs = c(pairs,pair_name)
    }
    return(pairs)
    
  }
  
  ## create comparison-----------------------------
  all_pairs = c()
  for(j in 1:(len_list-1)){
    print(j)
    test = get_pair_type(j = j,len_list=len_list)
    print(test)
    all_pairs = c(all_pairs,test)
  }
  
  ## make comparison table------------------------
  out_list = list()
  for(i in all_pairs){
    name_split = strsplit(i,"_and_")
    set1_name = name_split[[1]][1]
    set2_name =  name_split[[1]][2]
    out = intersect_setdiff_for_two_sets(set1 =setlist[[set1_name]] ,
                                         set2 = setlist[[set2_name]])
    print(out)
    
    out_list[[i]] = out
  }
  
  
  out_df = do.call(cbind,out_list)
  rownames(out_df) = c("left","right","left_and_right")
  
  overlap_save = paste(overlap_items,collapse = ";")
  export_df = rbind(out_df,rep(overlap_save,length(all_pairs)))
  rownames(export_df)[nrow(export_df)] = "overlap"
  
  write.csv(export_df,file = paste0("groups_intersect_setdiff_details.csv"),
            row.names = T)
  
} 



df1 = data.frame(gene_name = c(1,2,3,4,5),
                 biotype = rep("protein_coding",5))

df2 = data.frame(gene_name = c(10,11,12,5,7),
                 biotype = rep("protein_coding",5))

df3 = data.frame(gene_name = c(5,8,9,2,1),
                 biotype = rep("protein_coding",5))

df4 = data.frame(gene_name = c(2,1,5,7,9),
                 biotype = rep("protein_coding",5))

check_col = "gene_name"
set1 = df1[,check_col]
set2 = df2[,check_col]

set3 = df3[,check_col]
set4 = df4[,check_col]
list_name = c("set1","set2","set3","set4")
setlist = list(set1,set2,set3,set4)

venny_table(set_list = set_list ,list_name = list_name )

[xiasijian]