Scatter plot of variables in a column of a tibble using dplyr & ggplot2

Question

I have the following tibble that would like to use to make a scatter plot (using ggplot2) of logcpm values of AA_Colon vs. BA_Colon matched by gene.

             gene   sample     logcpm
             <chr>    <chr>      <dbl>
 1 ENSG00000169903 AA_Colon 0.31536340
 2 ENSG00000145321 AA_Colon 0.19735593
 3 ENSG00000171560 AA_Colon 0.00000000
 4 ENSG00000171557 AA_Colon 0.19735593
 5 ENSG00000106327 AA_Colon 0.06882901
 6 ENSG00000228278 AA_Colon 0.13452328
 7 ENSG00000138115 AA_Colon 0.31536340
 8 ENSG00000148702 AA_Colon 0.00000000
 9 ENSG00000140107 AA_Colon 0.00000000
10 ENSG00000197723 AA_Colon 0.00000000
11 ENSG00000169903 BA_Colon 1.14724849
12 ENSG00000145321 BA_Colon 0.08113901
13 ENSG00000171560 BA_Colon 0.36654820
14 ENSG00000171557 BA_Colon 0.23088996
15 ENSG00000106327 BA_Colon 0.08113901
16 ENSG00000228278 BA_Colon 0.08113901
17 ENSG00000138115 BA_Colon 0.42987550
18 ENSG00000148702 BA_Colon 0.00000000
19 ENSG00000140107 BA_Colon 0.00000000
20 ENSG00000197723 BA_Colon 0.08113901

Currently, I'm doing (which works):

tibble %>%
    spread(key = sample, value = logcpm) %>%
    ggplot(aes(x = AA_Colon, y = BA_Colon)) +
    geom_point()

But, I'm wondering if there's a more elegant way to directly work with the tidy format and extract the two vectors to plot instead of spreading the data into two columns.

Answer 1

When the data are in tidy format, ggplot gives scatterplot with just two columns of dots, one for AA_Colon and one for BA_Colon.

ggplot(tibble, aes(x = sample, y = logcpm)) + 
        geom_point()

Maybe boxplots with geom_jitter would be more useful.

ggplot(tibble, aes(x = sample, y = logcpm)) +
geom_boxplot() + 
geom_jitter(width = 0.3)