Number of shared components in R dataframe - r

Question

I have a dataframe as follows:
deput(data)
structure(list(genome = c("A", "A", "A", "A", "B", "B", "B",
"B"), gene = c("esaA", "esaB", "esaC", "esaC", "essA", "essB",
"essC", "esaA")), class = c("spec_tbl_df", "tbl_df", "tbl", "data.frame"
), row.names = c(NA, -8L), spec = structure(list(cols = list(
genome = structure(list(), class = c("collector_character",
"collector")), gene = structure(list(), class = c("collector_character",
"collector"))), default = structure(list(), class = c("collector_guess",
"collector")), skip = 1), class = "col_spec"))
I am interested in the number of items which are shared between each genome, e.g here between A and B it is 1. I can do this (very very slowly) in a loop:
out=NULL
for(i in guuids){
print(i)
m<-filter(data,genome==i)
g<-unique(m$gene)
for(z in guuids){
print(z)
p<-filter(data,genome==z)
p<-filter(p,gene %in% g)
p<-nrow(p)
out=rbind(out,data.frame(i,z,p))
}
}
but this obviously does not scale and there is surely a better way?

Answer 1

table can count them for you:
table(df)
gene
genome esaA esaB esaC essA essB essC
A 1 1 2 0 0 0
B 1 0 0 1 1 1
table(df[,c('gene','genome')])
genome
gene A B
esaA 1 1
esaB 1 0
esaC 2 0
essA 0 1
essB 0 1
essC 0 1
The returned object is a matrix, so you can simply test for which rows/matrices have more than one non-zero entry.

Answer 2

You can do this based on regex:
First you define all unique genes in group A as an alternation pattern; you do the same for B. Then you subset df on those rows for which grepl finds the same gene matches:
gene_A <- paste0("(", paste0(unique(df$gene[df$genome=="A"]), collapse = "|"), ")")
gene_B <- paste0("(", paste0(unique(df$gene[df$genome=="B"]), collapse = "|"), ")")
df[grepl(gene_B, df$gene) == grepl(gene_A, df$gene),]
# A tibble: 2 x 2
genome gene
<chr> <chr>
1 A esaA
2 B esaA
Data:
df <- structure(list(genome = c("A", "A", "A", "A", "B", "B", "B",
"B"), gene = c("esaA", "esaB", "esaC", "esaC", "essA", "essB",
"essC", "esaA")), class = c("spec_tbl_df", "tbl_df", "tbl", "data.frame"
), row.names = c(NA, -8L), spec = structure(list(cols = list(
genome = structure(list(), class = c("collector_character",
"collector")), gene = structure(list(), class = c("collector_character",
"collector"))), default = structure(list(), class = c("collector_guess",
"collector")), skip = 1), class = "col_spec"))

Answer 3

Perhaps you can try intersect + unstack
> do.call(intersect, unname(unstack(rev(data))))
[1] "esaA"