How to to assign logarithmic scale to “Intersection size” using ComplexUpset library?

Question

When I assigning a logarithmic scale to “Intersection size” using ComplexUpset library the bars disappear. Is there a way around?

Answer 1

This is because in ggplot2's geom_bar does not play well with log10 scale/transform when using stat other than "identity" (see Bar plot with log scales). CompelxUpset plots each data point individually to allow for colouring each observation separately. You can revert to plotting pre-summarised data by passing the following function as data to geom_bar:

presence = ComplexUpset:::get_mode_presence('exclusive_intersection')

summarise_values = function(df) {
    aggregate(
        as.formula(paste0(presence, '~ intersection')),
        df,
        FUN=sum
    )
}

and then plot it, starting with the example:

library(ComplexUpset)
library(ggplot2)
movies = as.data.frame(ggplot2movies::movies)
movies[movies$mpaa == '', 'mpaa'] = NA
movies = na.omit(movies)
genres = colnames(movies)[18:24]

to get log10 transform only:

upset(
    movies,
    genres,
    base_annotations=list(
        'log10(intersection size)'=(
            ggplot()
            + geom_bar(data=summarise_values, stat='identity', aes(y=!!presence))
            + ylab('Intersection size')
            + scale_y_continuous(trans='log10')
        )
    ),
    min_size=5,
    width_ratio=0.1
)

to get log10 on axis too:

upset(
    movies,
    genres,
    base_annotations=list(
        'log10(intersection size)'=(
            ggplot()
            + geom_bar(
                data=summarise_values,
                stat='identity',
                aes(y=log10(!!presence))
            )
            + ylab('log10(intersection size)')
        )
    ),
    width_ratio=0.1
)