Nsryjdtyk

> head(Gene)

  Key Func.ensGene Func.genericGene Func.refGene

1   1   intergenic       intergenic   intergenic

2   2   intergenic       intergenic   intergenic

3   3   intergenic       intergenic     intronic

4   4       exonic           exonic       exonic

5   5   intergenic       intergenic     intronic

6   6   intergenic       intergenic     intronic

Required Output:

Type          Func.ensGene Func.genericGene Func.refGene

exonic             1              1                1

intergenic         5              5                2

intronic           0              0                3

The solution I tried is working on only one column:

unique(Gene["Func.ensGene"])

Could I get the output table as shown above and get a barplot
where X-axis has the 'Type' and the bar represents counts from each column?

Simply use ?xtabs along with ?stack:

xtabs( ~ values + ind , stack(df1[,-1]))

or even shorter as @nicola suggests:

table(stack(df1[,-1]))

for both you get:

#            ind

#values       Func.ensGene Func.genericGene Func.refGene

#  exonic                1                1            1

#  intergenic            5                5            2

#  intronic              0                0            3

do you prefer further working on a data.frame?

as.data.frame.matrix(

    xtabs( ~ values + ind , stack(df1[,-1]))  # or again only table(stack(df1[,-1]))

)

We can get all unique levels from the dataframe and for every column calculate the count of each level by first converting it to factor with unique levels calculated before.

unique_names <- unique(unlist(df[-1]))

sapply(df[-1], function(x) table(factor(x, levels = unique_names)))



#           Func.ensGene Func.genericGene Func.refGene

#intergenic            5                5            2

#exonic                1                1            1

#intronic              0                0            3

Love the base R solutions but using data.table and some magrittr for readability you could get a data.frame directly (instead of a table):

library(magrittr)

library(data.table)

setDT(df)

df %>%

  melt(id.vars = "Key") %>%

  .[, .N, .(variable, value)] %>% 

  dcast(value ~ variable, value.var = "N", fill = 0)



        value Func.ensGene Func.genericGene Func.refGene

1:     exonic            1                1            1

2: intergenic            5                5            2

3:   intronic            0                0            3

Or much more concisely(as suggested by Henrik):

dcast(melt(df, "Key"), value ~ variable)

If you prefer tidyverse functions:

library(tidyr)

df %>%

  gather(key = Key) %>%

  group_by(Key, value) %>%

  count() %>%

  spread(Key, n, fill = 0)



# A tibble: 3 x 4

# Groups:   value [3]

  value      Func.ensGene Func.genericGene Func.refGene

  <chr>             <dbl>            <dbl>        <dbl>

1 exonic                1                1            1

2 intergenic            5                5            2

3 intronic              0                0            3

Data:

df <- data.frame(

  Key              = 1:6, 

  Func.ensGene     = c("intergenic", "intergenic", "intergenic", "exonic", "intergenic", "intergenic"), 

  Func.genericGene = c("intergenic", "intergenic", "intergenic", "exonic", "intergenic", "intergenic"), 

  Func.refGene     = c("intergenic", "intergenic", "intronic", "exonic", "intronic", "intronic"),

  stringsAsFactors = FALSE

)