Broken guides with ggplot2 v3.5.0

[teng-gao]

With ggplot2 v3.4.1, the guides are collected normally:

library(numbat)
library(dplyr)
nb = readRDS(url('http://pklab.med.harvard.edu/teng/data/nb_TNBC1.rds'))

nb$bulk_clones %>% 
  filter(n_cells > 50) %>%
  plot_bulks(
    min_LLR = 10, # filtering CNVs by evidence
    legend = TRUE
  )

With ggplot2 v3.5.0, guides are not combined correctly and some are missing:

SessionInfo:

R version 4.2.2 Patched (2022-11-10 r83330)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 20.04.6 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.9.0
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.9.0

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] dplyr_1.1.3  numbat_1.4.0 Matrix_1.6-5

loaded via a namespace (and not attached):
 [1] viridis_0.6.2          tidyr_1.3.0            tidygraph_1.3.1       
 [4] jsonlite_1.8.8         viridisLite_0.4.1      ggraph_2.1.0          
 [7] RhpcBLASctl_0.21-247.1 RcppParallel_5.1.6     yulab.utils_0.0.6     
[10] ggrepel_0.9.2          pillar_1.9.0           lattice_0.20-45       
[13] glue_1.7.0             quadprog_1.5-8         phangorn_2.11.1       
[16] uuid_1.1-0             digest_0.6.27          polyclip_1.10-4       
[19] colorspace_2.1-0       ggfun_0.0.9            htmltools_0.5.4       
[22] pkgconfig_2.0.3        logger_0.2.2           purrr_1.0.1           
[25] patchwork_1.2.0        tidytree_0.4.6         scales_1.3.0          
[28] ggplotify_0.1.0        tweenr_2.0.2           ggforce_0.4.1         
[31] tibble_3.2.1           generics_0.1.0         farver_2.1.1          
[34] ggplot2_3.5.0          withr_3.0.0            parallelDist_0.2.6    
[37] repr_1.1.6             lazyeval_0.2.2         cli_3.6.2             
[40] magrittr_2.0.3         crayon_1.5.2           evaluate_0.20         
[43] fansi_1.0.4            nlme_3.1-161           MASS_7.3-58           
[46] Cairo_1.6-0            tools_4.2.2            data.table_1.14.10    
[49] lifecycle_1.0.4        stringr_1.5.1          aplot_0.1.9           
[52] hahmmr_1.0.0           ggtree_3.6.2           munsell_0.5.0         
[55] compiler_4.2.2         gridGraphics_0.5-1     rlang_1.1.3           
[58] grid_4.2.2             pbdZMQ_0.3-9           IRkernel_1.3.2        
[61] igraph_2.0.2           base64enc_0.1-3        labeling_0.4.2        
[64] scistreer_1.2.0        gtable_0.3.4           codetools_0.2-18      
[67] graphlayouts_0.8.4     R6_2.5.1               gridExtra_2.3         
[70] fastmap_1.1.0          utf8_1.2.2             fastmatch_1.1-3       
[73] treeio_1.22.0          ape_5.7-1              stringi_1.8.3         
[76] parallel_4.2.2         IRdisplay_1.1          Rcpp_1.0.12           
[79] vctrs_0.6.5            tidyselect_1.2.0

[thomasp85]

Can you please provide a minimal reprex?

Specifically I think this is due to changes in ggplot2 in how missing data points contribute to guides, rather than an issue in patchwork, but I may be mistaken

[trekonom]

The issue was cross-posted here and as the comments by @teunbrand make clear is due to the recent change in ggplot2 which now requires to set show.legend=TRUE in the geom to display a legend key for missing categories and hence requires a fix on side of numbat.