My ggplot code below is working on my Mac, but isn't working on my Linux instance.

As per the title, I get the error Error: Aesthetics must be either length 1 or the same as the data (1): x, y and colour with this code.

library(tidyverse)
library(plotly)
library(ggplot2)

plot_total <-
          demo %>%
          as.data.frame() %>%
          ggplot(aes(log2(baseMean), log2FoldChange, color=padj < 0.01,
                     text = paste(" log2(baseMean):", formatC(log2(baseMean), format = "e", digits = 2), "\n",
                                  "log2FoldChange:", formatC(log2FoldChange, format = "e", digits = 2), "\n",
                                  "Gene Name:", Gene, "\n",
                                  "EnsGeneID:", ENSG)))  
          geom_point(data = . %>% filter(!padj<0.01), cex = 0.1, size=0.1)  
          geom_point(data = . %>% filter(is.na(padj)), cex = 0.1, size=0.1)  
          geom_point(data = . %>% filter(padj<0.01), cex = 1, size=1)  
          scale_color_manual(values=c("FALSE" = "#30B7BC", "TRUE" = "#DE653A", "NA" = "#B3B3B3"))  
          labs(title = "Treated vs. Control", x = "log2Expression")  
          theme_bw()  
          NULL
    
    ggplotly(plot_total,
             tooltip = "text",
             type = 'scattergl') %>% 
      config(displaylogo = FALSE,
             modeBarButtonsToRemove = c('sendDataToCloud', 'select2d', 'lasso2d')) %>% 
      toWebGL()

The culprit seems to be the text variable. I have tried placing this variable inside each geom_point(aes()) where I'm filtering the data, but I get the same error. This is a small subset of the data:

CodePudding user response：

I don't really understand your question, as your code works fine. The only thing that was not working was "WebGL is not supported by your browser". I solved it forcing RStudio to use 'Desktop OpenGL'. In RStudio, I went to [Tools/Global Options.../Advanced] and choosing 'Desktop OpenGL' in Rendering engine.

Sample code:

library(tidyverse)
library(plotly)
library(ggplot2)



plot_total <-demo %>%
  as.data.frame() %>%
  ggplot(aes(log2(baseMean), log2FoldChange, color=padj < 0.01,
             text = paste(" log2(baseMean):", formatC(log2(baseMean), format = "e", digits = 2), "\n",
                          "log2FoldChange:", formatC(log2FoldChange, format = "e", digits = 2), "\n",
                          "Gene Name:", Gene, "\n",
                          "EnsGeneID:", ENSG)))  
  geom_point(data = . %>% filter(!padj<0.01), cex = 0.1, size=0.1)  
  geom_point(data = . %>% filter(is.na(padj)), cex = 0.1, size=0.1)  
  geom_point(data = . %>% filter(padj<0.01), cex = 1, size=1)  
  scale_color_manual(values=c("FALSE" = "#30B7BC", "TRUE" = "#DE653A", "NA" = "#B3B3B3"))  
  labs(title = "Treated vs. Control", x = "log2Expression")  
  theme_bw()  
  NULL

ggplotly(plot_total,
         tooltip = "text",
         type = 'scattergl') %>% 
  config(displaylogo = FALSE,
         modeBarButtonsToRemove = c('sendDataToCloud', 'select2d', 'lasso2d')) %>% 
  toWebGL()

Plot:

CodePudding user response：

The reason I was getting this error was because the df didn't contain NA under padj column for my particular dataset, so it was trying to filter what wasn't there. Adding an if/else statement fixed.