How to map row elements to column after doing intersection?

user_14010653 0
Viewed 38

This was a previous question I asked I nearly have the solution provided by stack community

My dummy dataframe

df1 = data.frame(genes = c('gene1', 'gene3', 'gene4', 'gene2','gene11'))
df2 = data.frame(genes = c('gene3', 'gene2', 'gene5', 'gene1', "genet",'gene11'))
df3 = data.frame(genes = c('gene6', 'gene3', 'gene4', 'gdene7','gene11', 'genex', "gene10"))
dfList <- list(df1, df2, df3)

To see if my set of genes are present in each dataframe or not I ran this

dfList %>% 
  bind_rows(.id="df") %>% 
  pivot_wider(names_from=df, names_prefix="df", values_from=df)

Which gives something like this

genes   df1 df2 df3
<chr>   <chr>   <chr>   <chr>
gene1   1   2   NA
gene3   1   2   3
gene4   1   NA  3
gene2   1   2   NA
gene11  1   2   3
gene5   NA  2   NA
genet   NA  2   NA
gene6   NA  NA  3
gdene7  NA  NA  3
genex   NA  NA  3
gene10  NA  NA  3

In the above dataframe I can see for each gene, whether or not that gene is in df1, df2, or df3. The problem I have is to find genes based on their presence and absence in different columns/condition

The below code as suggested to do what Im trying to do.

dfList %>% 
  bind_rows(.id="df") %>% 
  group_by(genes) %>% 
  summarise(minDF=min(df), maxDF=max(df)) %>% 
  filter(minDF == maxDF & maxDF == 3) %>% 
  pull(genes)

'gdene7''gene10''gene6''genex'

It identifies genes, in this example, that are only present in df3. How do I identify genes that are, for example:

  • present in all three frames?
  • present in two specified frames, say df1 and df2
  • present only in one frame, say df2?

Any help or suggestion would be really appreciated

r
1 Answers

There are a couple of options here

  1. You can modify your pivot_longer() call to include the value_fn param to make logical vectors, and then use rowwise(), and use any or all. For example, if you want to get the genes that are in df1, df2, and df3, you can do this:
dfList %>% 
  bind_rows(.id="df") %>% 
  pivot_wider(names_from=df, names_prefix="df", values_from=df,values_fn = \(x) !is.na(x)) %>% 
  rowwise() %>% 
  filter(all(df1,df2,df3)) %>% 
  pull(genes)

[1] "gene3"  "gene11"

If you change the filter(all()) line to filter(all(df1,df2)), this will give you the genes in both df1 and df2, without regard to df3

[1] "gene1"  "gene3"  "gene2"  "gene11"
  1. However, you don't need these intermediate steps. Instead, you can use this flexible function below, called find_genes(), which takes a function f, such as intersect, union, or setdiff, and any number of dfs, and returns the set of genes across those dfs according to the set-function requested
find_genes <- function(f,...) {
  Reduce(f,c(...))
}

Now, if you want to know the genes only in df3, and not in any others, do this:

find_genes(setdiff,df3,df2,df1)

Output:

[1] "gene6"  "gdene7" "genex"  "gene10"

If you want to know genes that are present in both df1 and df2, pass intersect to the find_genes() function, like this:

find_genes(intersect, df1,df2)

Output:

[1] "gene1"  "gene3"  "gene2"  "gene11"

If you want to know genes that are present in all three, pass all three frames

find_genes(intersect, df3, df2, df1)

Output:

[1] "gene3"  "gene11"

If you want the union of genes in two frames, pass union:

find_genes(union, df1, df3)

Output:

 "gene1"  "gene3"  "gene4"  "gene2"  "gene11" "gene6"  "gdene7" "genex"  "gene10"
user_44475400
Related