Error in `translate_shape_string()`: ! Shape aesthetic contains invalid values: 19.

[ltzyl]

Hello, I have noticed that when using DimPlot, it consistently reports the following error:
> DimPlot(rna_integrated_harmony,reduction = 'umap.harmony')

Rasterizing points since number of points exceeds 100,000.
To disable this behavior set raster=FALSE
Error in translate_shape_string():
! Shape aesthetic contains invalid values: 19.
Run rlang::last_trace() to see where the error occurred.

I am unaware of the reason, and the number of groups is far below 19.If I set raster to TRUE, it will work normally.

version.string R version 4.3.0 (2023-04-21)

packageVersion("Seurat")
[1] '5.1.0'
packageVersion("ggplot2")
[1] '3.5.2'
packageVersion("patchwork")
[1] '1.3.2'
packageVersion("rlang")
[1] '1.1.6'

[teunbrand]

Thanks for the report! Can you please reduce this to a minimal reproducible example, preferably using the regex package and just using ggplot2 (i.e. not Seurat's DimPlot())? Also have you tried wether the error persists with the latest ggplot2 version (4.0.0)?

[ltzyl]

Right, to resolve this issue, I downloaded several versions of Seurat, including 5.0.3 and v4. The problem persisted when using DimPlot. My dataset comprises 170,000 cells, so I'm unable to provide an instance. Following your advice, I successfully plotted using ggplot2 with the following code:

ggplot(plot_data, aes(x = UMAP1, y = UMAP2, color = Cluster)) + geom_point(alpha = 0.7) + scale_color_viridis_d() + theme_minimal() + labs(title = "UMAP Plot colored by wnn_cluster_res_0.5", x = "UMAP1", y = "UMAP2") + theme(legend.title = element_text(size = 12), legend.text = element_text(size = 10))

I was using version 3.5.2 of ggplot2 precisely because of the previous VlnPlot error, but after upgrading to 4.0, I find the issue persists:

library(Signac)
library(Seurat)
library(ggplot2)
DimPlot(combined_data,reduction='wnn.umap',group.by='wnn_cluster_res_0.5')
Rasterizing points since number of points exceeds 100,000.
To disable this behavior set raster=FALSE
Error in translate_shape_string():
! Shape aesthetic contains invalid values: 19.
Run rlang::last_trace() to see where the error occurred.
rlang::last_trace()
<error/rlang_error>
Error in translate_shape_string():
! Shape aesthetic contains invalid values: 19.
Backtrace:
▆

1. ├─base (local) <fn>(x)
2. └─patchwork:::print.patchwork(x)
3. └─patchwork:::build_patchwork(plot, plot$layout$guides %||% "auto")

4. └─base::lapply(x$plots, plot_table, guides = guides)
   

5.   ├─patchwork (local) FUN(X[[i]], ...)
   

6.   └─patchwork:::plot_table.ggplot(X[[i]], ...)
   

7.     └─ggplot2::ggplotGrob(x)
   

8.       ├─ggplot2::ggplot_gtable(ggplot_build(x))
   

9.       └─ggplot2:::ggplot_gtable.ggplot_built(ggplot_build(x))
   

10.         └─plot$guides$assemble(theme)
    

11.           └─ggplot2 (local) assemble(..., self = self)
    

12.             └─self$draw(theme, default_position, theme$legend.direction)
    

13.               └─ggplot2 (local) draw(..., self = self)
    

14.                 └─guides[[i]]$draw(...)
    

15.                   └─ggplot2 (local) draw(..., self = self)
    

16.                     └─self$build_decor(params$decor, grobs, elems, params)
    

17.                       └─ggplot2 (local) build_decor(...)
    

18.                         ├─base::unlist(lapply(seq_len(params$n_breaks), draw), FALSE)
    

19.                         └─base::lapply(seq_len(params$n_breaks), draw)
    

20.                           └─ggplot2 (local) FUN(X[[i]], ...)
    

21.                             └─base::lapply(...)
    

22.                               └─ggplot2 (local) FUN(X[[i]], ...)
    

23.                                 └─g$draw_key(data, g$params, key_size)
    

24.                                   └─ggplot2 (local) draw_key(...)
    

25.                                     └─ggplot2::translate_shape_string(data$shape %||% 19)
    

Run rlang::last_trace(drop = FALSE) to see 2 hidden frames. > packageVersion("Signac")
[1] ‘1.16.0’
packageVersion("Seurat")
[1] ‘5.2.1’
packageVersion("ggplot2")
[1] ‘4.0.0’`

[teunbrand]

I apologise if I've been unclear, but the reason I'm asking for a reproducible example is so that we can debug and test solutions on our end, or deduce that the problem lies elsewhere. This ask wasn't intended as advice. We don't require your 170k cell dataset, we merely need to know how to trigger the error. If there is built-in data in Seurat that'll trigger the error, that'll be just fine. If we can reproduce the error without Seurat, that'll be even better.

[ltzyl]

I understand your point, but I regret that I cannot provide specific examples, as I have only worked with non-public datasets and have not tested this with publicly available datasets. Based on previous testing, this likely relates to Seurat's built-in functions. I suspect it may be connected to the scale of the dataset. Should you happen to possess a dataset exceeding 10,000 cells, you might wish to verify this. Thank you.

[smouksassi]

when cross posting please linkg to it this probably does not belon to ggplot2 issue tracker as it was submitted here:
satijalab/seurat#10201