Cytonorm doesn't work when specifying different cellular and clustering parameters

[hrj21]

Hello gang!

I'm sure I've been able to do this in the past with no issue, but I now seem unable to specify different parameters for alignment and clustering with cytonorm. When I do so, the script errors at train.cytonorm(). Here's an example that works when the same parameters are used for cellular.cols and cluster.cols, but fails when the latter is a subset of the former. I can provide the data too if you can't reproduce this.

# load packages ----
library(Spectre)
package.load()

# read data ----
abs_path <- R.utils::getAbsolutePath("../data/pre-batch correction")

data_list <- read.files(
  file.loc = abs_path, 
  file.type = ".fcs", 
  do.embed.file.names = TRUE
)

cell_dat <- do.merge.files(dat = data_list)

# transform parameters ----
to_asinh <- names(cell_dat)[1:8]
cell_dat <- do.asinh(cell_dat, to_asinh, cofactor = 500) 
transformed_cols <- paste0(to_asinh, "_asinh")

# add metadata ----
# just adds columns "Batch" and "Sample"
meta_dat <- read.csv("../data/pre-batch correction/sample_details.csv")

cell_dat <- do.add.cols(
  cell_dat, 
  base.col = "FileName", 
  add.dat = meta_dat, 
  add.by = "FileName"
)

# prepare for alignment ----
reference_files <- c("Mock_01_A", "Mock_05_B")
reference_dat <- do.filter(cell_dat, "Sample", reference_files)

cytonorm <- prep.cytonorm(
  dat = reference_dat,
  cellular.cols = transformed_cols,
  cluster.cols = transformed_cols[1:4], # this causes train.cytonorm to fail
  batch.col = "Batch",
  sample.col = "Sample"
) 

# train cytonorm ----
cytonorm <- train.cytonorm(model = cytonorm, align.cols = transformed_cols) # fails

The final output is:

Training alignment - setup
Working directory is '[I removed this]'
Training alignment - file (batch) preparation
Training alignment - file (batch) and metacluster splitting
 -- running File '1'
 -- running File '2'
Training alignment - learning conversions
Processing cluster 1
Computing Quantiles
  A (FileID 1)

Error in flowCore::exprs(ff)[, channels, drop = FALSE] : 
  subscript out of bounds

Here is my sessionInfo(), with Spectre 1.2.0, CytoNorm 2.0.8, and FlowSOM 2.0.8:

R version 4.4.1 (2024-06-14 ucrt)
Platform: x86_64-w64-mingw32/x64
Running under: Windows 11 x64 (build 22631)

Matrix products: default


locale:
[1] LC_COLLATE=English_United Kingdom.utf8  LC_CTYPE=English_United Kingdom.utf8    LC_MONETARY=English_United Kingdom.utf8
[4] LC_NUMERIC=C                            LC_TIME=English_United Kingdom.utf8    

time zone: Europe/London
tzcode source: internal

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

other attached packages:
 [1] Biobase_2.64.0                 BiocGenerics_0.52.0            CytoNorm_2.0.8                 viridis_0.6.5                 
 [5] viridisLite_0.4.2              uwot_0.2.2                     umap_0.2.10.0                  scattermore_1.2               
 [9] scales_1.3.0                   Rtsne_0.17                     rsvd_1.0.5                     rstudioapi_0.17.1             
[13] RColorBrewer_1.1-3             pheatmap_1.0.12                patchwork_1.3.0                irlba_2.3.5.1                 
[17] Matrix_1.7-2                   gtools_3.9.5                   gridExtra_2.3                  ggthemes_5.1.0                
[21] ggpubr_0.6.0                   ggpointdensity_0.1.0           FlowSOM_2.14.0                 igraph_2.1.4                  
[25] flowCore_2.16.0                factoextra_1.0.7               ggplot2_3.5.1                  dendsort_0.3.4                
[29] data.table_1.16.4              colorRamps_2.3.4               Spectre_1.2.0.0.0.0.1728133410

loaded via a namespace (and not attached):
 [1] tidyselect_1.2.1            dplyr_1.1.4                 farver_2.1.2                R.utils_2.12.3             
 [5] tweenr_2.0.3                XML_3.99-0.18               lifecycle_1.0.4             cluster_2.1.6              
 [9] magrittr_2.0.3              compiler_4.4.1              rlang_1.1.4                 tools_4.4.1                
[13] knitr_1.49                  ggsignif_0.6.4              labeling_0.4.3              askpass_1.2.1              
[17] reticulate_1.40.0           ConsensusClusterPlus_1.68.0 abind_1.4-8                 withr_3.0.2                
[21] purrr_1.0.2                 RProtoBufLib_2.16.0         R.oo_1.27.0                 grid_4.4.1                 
[25] polyclip_1.10-7             stats4_4.4.1                colorspace_2.1-1            ggridges_0.5.6             
[29] MASS_7.3-60.2               cli_3.6.3                   ragg_1.3.3                  generics_0.1.3             
[33] RSpectra_0.16-2             ggforce_0.4.2               stringr_1.5.1               matrixStats_1.5.0          
[37] vctrs_0.6.5                 jsonlite_1.8.9              carData_3.0-5               cytolib_2.16.0             
[41] car_3.1-3                   S4Vectors_0.42.1            rstatix_0.7.2               ggrepel_0.9.6              
[45] Formula_1.2-5               systemfonts_1.2.1           ggnewscale_0.5.0            tidyr_1.3.1                
[49] glue_1.8.0                  stringi_1.8.4               gtable_0.3.6                munsell_0.5.1              
[53] tibble_3.2.1                pillar_1.10.1               openssl_2.3.2               R6_2.6.1                   
[57] textshaping_1.0.0           evaluate_1.0.3              lattice_0.22-6              png_0.1-8                  
[61] R.methodsS3_1.8.2           backports_1.5.0

Please let me know what you think.

Best wishes
Hefin

[tomashhurst]

@hrj21 hey mate, this is actually an old issue that I thought we fixed by v1.2, but looking at the code I think we applied some extra checks for prep.cytonorm but not where the error actually occurs in train.cytonorm. We are just prepping v1.3 so I'll add a fix now and it should be up shortly.

[hrj21]

Hi Tom, I'm using Spectre 1.3.0.0.0.0.1758242461 and CytoNorm 2.0.8, and I still get this behaviour. Are you able to reproduce this or am I doing something wrong?

Best wishes
Hefin