数据透视表Excel或R

Question

My data looks like this on excel:

Genename ID1 ID2 ID3
Gene1   R   H   R
Gene1   R   H   R
Gene1   H   R   H
Gene2   H   R   H
Gene2   R   R   H
Gene2   H   R   R
Gene2   R   R   R

I would like to create a column with the total number of individuals with at least one H per gene. So, it should look like this;

Genename Het
Gene1 3
Gene2 2

I have hundreds of gene so I need an automated way to get these counts.Thanks in advance.

Answer 1

try this

library(data.table)
data <- data.table(data)
res <- data[,list("Genename"=Genename,"Het"=rowSums(data=="H")>0)]
res <- res[,list("Het"=sum(Het)),by=Genename]

> res
#   Genename Het
#1:    Gene1   3
#2:    Gene2   3

Answer 2

We can group by 'Genename', get a logical index for any values that are 'H' in each column, get the sum within summarise and then we use rowSums to get the expected output.

library(dplyr)
df1 %>% 
   group_by(Genename) %>% 
   summarise_each(funs(sum(any(.=='H')))) %>% 
   transmute(Genename= Genename, Het = rowSums(.[-1L]))
   Genename   Het
#    (chr) (dbl)
#1    Gene1     3
#2    Gene2     2

---

Or as @aosmith mentioned distinct would be an option after converting the 'wide' to 'long' format with gather .

library(tidyr)
gather(df1, Var1, Var2, -Genename) %>% 
            group_by(Genename, Var1) %>%
            distinct(Var2) %>%
            group_by(Genename) %>%
            summarise(Het= sum(Var2=='H'))

Update

If we need the count of IDs with no 'H' and at least one 'R' per 'Genename'

df1 %>% 
     group_by(Genename) %>%
     summarise_each(funs(all(.!='H') & any(.=='R'))) %>% 
     transmute(Genename=Genename, Het= rowSums(.[-1L]))
#   Genename   Het
#     (chr) (dbl)
#1    Gene1     0
#2    Gene2     1