CODEWITHSUNDEEP
rggplot2shiny
user3234810
user3234810

Reputation: 482

R shiny - fix color/shape for dynamic plots

I have an app designed to plot 'pre-computed' data frames stored in a MYSQL database.

The data frames consist of four columns - id, genotype, A-value and B-value. The plot is a simple A/B geom_point, see example code below.

I want to manually define the shape and colour of the points based on the 'Genotype' variable.

The values for genotype will always be; "-1", "0" , "1" and "2". However, in some data frames, not all of these values are present (e.g. 1 only) however I would like to keep the shape/color the same between plots.

Does anybody have a solution for this? 

library( ggplot2 )

# Example values 
id       <- c( "s1" , "s2" , "s3" , "s4" )
genotype <- c( -1 , 0 , 1 , 2 )
A_value  <- c( 100 , 110 , 120 , 130 )
B_value  <- c( 130, 120 , 110 , 100 )

# data_frame 
df <- data.frame( id, genotype , A_value , B_value )

# Simple plot 
ggplot(df, aes( A_value , B_value )) +
  geom_point( aes(shape = factor( df$genotype ) , color = factor(df$genotype) ))

Upvotes: 0

Views: 759

Answers (1)

kath
kath

Reputation: 7724

You can specify the levels of genotype and then use drop = FALSE in the diescrete shape and color scales. Note that you should not use df$genotype in the specification of aes in the geom_point-call as you are referring to the same dataframe.

ggplot(df, aes(A_value, B_value)) +
  geom_point(aes(shape = factor(genotype, levels = c(-1 , 0 , 1 , 2)),
                 color = factor(genotype, levels = c(-1 , 0 , 1 , 2)))) +
  scale_color_discrete(name = "genotype", drop = F) +
  scale_shape_discrete(name = "genotype", drop = F)

enter image description here

ggplot(df[1:3, ], aes(A_value, B_value)) +
  geom_point(aes(shape = factor(genotype, levels = c(-1 , 0 , 1 , 2)),
                 color = factor(genotype, levels = c(-1 , 0 , 1 , 2)))) +
  scale_color_discrete(name = "genotype", drop = F) +
  scale_shape_discrete(name = "genotype", drop = F)

enter image description here

Upvotes: 1

Related Questions