bseqsc, cdseq, cpm and momf fail with latest Docker image on provided test data

[thegrebe]

Hi,

I am running the Nextflow pipeline as explained here, with the latest Docker image (pulled on 2025-04-01).

The only modifications I made to the nextflow.config file are input/output directories and general benchmarking parameters:

  data_dir_bulk = "/home/sciliciumtheo/omnideconv_bench/data/bulks/"                                                     // absolute path to directory that contains RNA-seq datasets
  data_dir_sc = "/home/sciliciumtheo/omnideconv_bench/data/single_cell/"                                                 // absolute path to directory that contains scRNA-seq datasets


  /*** output directories ***/  
  results_dir_general = "/home/sciliciumtheo/omnideconv_bench/test_results_fail"                                                     // absolute path to directory in which final results of the main and subsampling workflow are stored
  preProcess_dir = "/home/sciliciumtheo/omnideconv_bench/tmp"                                              // absolute path to directory in which temporary files are stored
  

  /*** Benchmarking Parameters [general] ***/  
  single_cell_list = ["sc1"]                                                                    // list of scRNA-seq dataset names that will be used for deconvolution. 
  bulk_list = ["bulk1","bulk2"]                                                                 // list of RNA-seq dataset names that will be used for deconvolution. 
  method_list = ["bseqsc", "cdseq", "cpm", "momf"]

Any method of bseqsc, cdseq, cpm and momf (added to nextflow.config) fail with similar error (here for cpm):

ERROR ~ Error executing process > 'CREATE_SIGNATURE (2)'

Caused by:
  Process `CREATE_SIGNATURE (2)` terminated with an error exit status (1)


Command executed:

  computeSignaturesNF.R 'sc1' '/home/sciliciumtheo/omnideconv_bench/data/single_cell/' 'bulk2' '/home/sciliciumtheo/omnideconv_bench/data/bulks/' 'cpm' '/home/sciliciumtheo/omnideconv_bench/test_results_fail_cpm' 'false' '0' '0' '4' '/home/sciliciumtheo/omnideconv_bench/pipeline'

Command exit status:
  1

Command output:
  [1] "Starting signature building script ..."
  R version 4.4.0 (2024-04-24)
  Platform: x86_64-pc-linux-gnu
  Running under: Ubuntu 20.04.6 LTS

  Matrix products: default
  BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
  LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3;  LAPACK version 3.9.0

  locale:
  [1] C

  time zone: Etc/UTC
  tzcode source: system (glibc)

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

  other attached packages:
  [1] omnideconv_0.0.0.9000 snowfall_1.84-6.3     snow_0.4-4

  loaded via a namespace (and not attached):
   [1] Matrix_1.7-0        gtable_0.3.6        jsonlite_1.8.9
   [4] dplyr_1.1.4         compiler_4.4.0      Rcpp_1.0.14
   [7] tidyselect_1.2.1    Biobase_2.62.0      parallel_4.4.0
  [10] tidyr_1.3.1         png_0.1-8           scales_1.3.0
  [13] fastmap_1.2.0       lattice_0.22-6      reticulate_1.40.0
  [16] ggplot2_3.5.1       R6_2.5.1            generics_0.1.3
  [19] BiocGenerics_0.48.1 htmlwidgets_1.6.4   tibble_3.2.1
  [22] munsell_0.5.1       pillar_1.10.1       RColorBrewer_1.1-3
  [25] rlang_1.1.5         lazyeval_0.2.2      viridisLite_0.4.2
  [28] plotly_4.10.4       cli_3.6.3           withr_3.0.2
  [31] magrittr_2.0.3      digest_0.6.37       grid_4.4.0
  [34] rappdirs_0.3.3      lifecycle_1.0.4     vctrs_0.6.5
  [37] glue_1.8.0          data.table_1.16.4   colorspace_2.1-1
  [40] purrr_1.0.4         httr_1.4.7          tools_4.4.0
  [43] pkgconfig_2.0.3     htmltools_0.5.8.1
  python:         /root/.local/share/r-miniconda/envs/r-omnideconv/bin/python
  libpython:      /root/.local/share/r-miniconda/envs/r-omnideconv/lib/libpython3.10.so
  pythonhome:     /root/.local/share/r-miniconda/envs/r-omnideconv:/root/.local/share/r-miniconda/envs/r-omnideconv
  version:        3.10.14 (main, May  6 2024, 19:42:50) [GCC 11.2.0]
  numpy:          /root/.local/share/r-miniconda/envs/r-omnideconv/lib/python3.10/site-packages/numpy
  numpy_version:  1.26.4

  NOTE: Python version was forced by use_python() function
  [1] "Method: cpm; sc-norm: ; bulk-norm: "

Command error:
  Loading required package: snowfall
  Loading required package: snow
  Skipping python installation checks ...
  R version 4.4.0 (2024-04-24)
  Platform: x86_64-pc-linux-gnu
  Running under: Ubuntu 20.04.6 LTS

  Matrix products: default
  BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
  LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3;  LAPACK version 3.9.0

  locale:
  [1] C

  time zone: Etc/UTC
  tzcode source: system (glibc)

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

  other attached packages:
  [1] omnideconv_0.0.0.9000 snowfall_1.84-6.3     snow_0.4-4

  loaded via a namespace (and not attached):
   [1] Matrix_1.7-0        gtable_0.3.6        jsonlite_1.8.9
   [4] dplyr_1.1.4         compiler_4.4.0      Rcpp_1.0.14
   [7] tidyselect_1.2.1    Biobase_2.62.0      parallel_4.4.0
  [10] tidyr_1.3.1         png_0.1-8           scales_1.3.0
  [13] fastmap_1.2.0       lattice_0.22-6      reticulate_1.40.0
  [16] ggplot2_3.5.1       R6_2.5.1            generics_0.1.3
  [19] BiocGenerics_0.48.1 htmlwidgets_1.6.4   tibble_3.2.1
  [22] munsell_0.5.1       pillar_1.10.1       RColorBrewer_1.1-3
  [25] rlang_1.1.5         lazyeval_0.2.2      viridisLite_0.4.2
  [28] plotly_4.10.4       cli_3.6.3           withr_3.0.2
  [31] magrittr_2.0.3      digest_0.6.37       grid_4.4.0
  [34] rappdirs_0.3.3      lifecycle_1.0.4     vctrs_0.6.5
  [37] glue_1.8.0          data.table_1.16.4   colorspace_2.1-1
  [40] purrr_1.0.4         httr_1.4.7          tools_4.4.0
  [43] pkgconfig_2.0.3     htmltools_0.5.8.1
  python:         /root/.local/share/r-miniconda/envs/r-omnideconv/bin/python
  libpython:      /root/.local/share/r-miniconda/envs/r-omnideconv/lib/libpython3.10.so
  pythonhome:     /root/.local/share/r-miniconda/envs/r-omnideconv:/root/.local/share/r-miniconda/envs/r-omnideconv
  version:        3.10.14 (main, May  6 2024, 19:42:50) [GCC 11.2.0]
  numpy:          /root/.local/share/r-miniconda/envs/r-omnideconv/lib/python3.10/site-packages/numpy
  numpy_version:  1.26.4

  NOTE: Python version was forced by use_python() function
  [1] "Method: cpm; sc-norm: ; bulk-norm: "
  Error in if (sc_norm == "counts") { : argument is of length zero
  Execution halted

Work dir:
  /home/sciliciumtheo/omnideconv_bench/work/86/d44564698550225faaf45fd95d86eb

Container:
  alexd13/omnideconv_benchmark:latest

Tip: you can try to figure out what's wrong by changing to the process work dir and showing the script file named `.command.sh`

 -- Check '.nextflow.log' file for details

For bseqsc, the only difference is:

  NOTE: Python version was forced by use_python() function
  [1] "Method: bseqsc; sc-norm: ; bulk-norm: "
  Error in if (sc_norm == "counts") { : argument is of length zero
  Execution halted

Best,
Theo

[alex-d13]

Hi,

Thanks for using deconvBench! I am currently reworking the Docker image, I already pushed a version 1.4, that you might try? Happy for Feedback!

Best,
Alex

[thegrebe]

Hi,

Thanks for the quick feedback!

I just pulled the 1.4 version, changed the docker profile in nextflow.config accordingly, but the problem is still present:

ERROR ~ Error executing process > 'CREATE_SIGNATURE (1)'

Caused by:
  Process `CREATE_SIGNATURE (1)` terminated with an error exit status (1)


Command executed:

  computeSignaturesNF.R 'sc1' '/home/sciliciumtheo/omnideconv_bench/data/single_cell/' 'bulk1' '/home/sciliciumtheo/omnideconv_bench/data/bulks/' 'cdseq' '/home/sciliciumtheo/omnideconv_bench/test_results_1.4' 'false' '0' '0' '4' '/home/sciliciumtheo/omnideconv_bench/pipeline'

Command exit status:
  1

Command output:
  [1] "Starting signature building script ..."
  R version 4.4.0 (2024-04-24)
  Platform: x86_64-pc-linux-gnu
  Running under: Ubuntu 20.04.6 LTS

  Matrix products: default
  BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
  LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3;  LAPACK version 3.9.0

  locale:
  [1] C

  time zone: Etc/UTC
  tzcode source: system (glibc)

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

  other attached packages:
  [1] omnideconv_0.1.1  snowfall_1.84-6.3 snow_0.4-4

  loaded via a namespace (and not attached):
   [1] Matrix_1.7-0        gtable_0.3.6        jsonlite_1.9.1
   [4] dplyr_1.1.4         compiler_4.4.0      Rcpp_1.0.14
   [7] tidyselect_1.2.1    Biobase_2.62.0      parallel_4.4.0
  [10] tidyr_1.3.1         png_0.1-8           scales_1.3.0
  [13] fastmap_1.2.0       lattice_0.22-6      reticulate_1.42.0
  [16] ggplot2_3.5.1       R6_2.6.1            generics_0.1.3
  [19] BiocGenerics_0.48.1 htmlwidgets_1.6.4   tibble_3.2.1
  [22] munsell_0.5.1       pillar_1.10.1       RColorBrewer_1.1-3
  [25] rlang_1.1.5         lazyeval_0.2.2      viridisLite_0.4.2
  [28] plotly_4.10.4       cli_3.6.4           withr_3.0.2
  [31] magrittr_2.0.3      digest_0.6.37       grid_4.4.0
  [34] rappdirs_0.3.3      lifecycle_1.0.4     vctrs_0.6.5
  [37] glue_1.8.0          data.table_1.17.0   colorspace_2.1-1
  [40] purrr_1.0.4         httr_1.4.7          tools_4.4.0
  [43] pkgconfig_2.0.3     htmltools_0.5.8.1
  python:         /root/.local/share/r-miniconda/envs/r-omnideconv/bin/python
  libpython:      /root/.local/share/r-miniconda/envs/r-omnideconv/lib/libpython3.8.so
  pythonhome:     /root/.local/share/r-miniconda/envs/r-omnideconv:/root/.local/share/r-miniconda/envs/r-omnideconv
  version:        3.8.16 | packaged by conda-forge | (default, Feb  1 2023, 16:01:55)  [GCC 11.3.0]
  numpy:          /root/.local/share/r-miniconda/envs/r-omnideconv/lib/python3.8/site-packages/numpy
  numpy_version:  1.23.5

  NOTE: Python version was forced by use_python() function
  [1] "Method: cdseq; sc-norm: ; bulk-norm: "

Command error:
  Loading required package: snowfall
  Loading required package: snow
  Skipping python installation checks ...
  R version 4.4.0 (2024-04-24)
  Platform: x86_64-pc-linux-gnu
  Running under: Ubuntu 20.04.6 LTS

  Matrix products: default
  BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
  LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3;  LAPACK version 3.9.0

  locale:
  [1] C

  time zone: Etc/UTC
  tzcode source: system (glibc)

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

  other attached packages:
  [1] omnideconv_0.1.1  snowfall_1.84-6.3 snow_0.4-4

  loaded via a namespace (and not attached):
   [1] Matrix_1.7-0        gtable_0.3.6        jsonlite_1.9.1
   [4] dplyr_1.1.4         compiler_4.4.0      Rcpp_1.0.14
   [7] tidyselect_1.2.1    Biobase_2.62.0      parallel_4.4.0
  [10] tidyr_1.3.1         png_0.1-8           scales_1.3.0
  [13] fastmap_1.2.0       lattice_0.22-6      reticulate_1.42.0
  [16] ggplot2_3.5.1       R6_2.6.1            generics_0.1.3
  [19] BiocGenerics_0.48.1 htmlwidgets_1.6.4   tibble_3.2.1
  [22] munsell_0.5.1       pillar_1.10.1       RColorBrewer_1.1-3
  [25] rlang_1.1.5         lazyeval_0.2.2      viridisLite_0.4.2
  [28] plotly_4.10.4       cli_3.6.4           withr_3.0.2
  [31] magrittr_2.0.3      digest_0.6.37       grid_4.4.0
  [34] rappdirs_0.3.3      lifecycle_1.0.4     vctrs_0.6.5
  [37] glue_1.8.0          data.table_1.17.0   colorspace_2.1-1
  [40] purrr_1.0.4         httr_1.4.7          tools_4.4.0
  [43] pkgconfig_2.0.3     htmltools_0.5.8.1
  python:         /root/.local/share/r-miniconda/envs/r-omnideconv/bin/python
  libpython:      /root/.local/share/r-miniconda/envs/r-omnideconv/lib/libpython3.8.so
  pythonhome:     /root/.local/share/r-miniconda/envs/r-omnideconv:/root/.local/share/r-miniconda/envs/r-omnideconv
  version:        3.8.16 | packaged by conda-forge | (default, Feb  1 2023, 16:01:55)  [GCC 11.3.0]
  numpy:          /root/.local/share/r-miniconda/envs/r-omnideconv/lib/python3.8/site-packages/numpy
  numpy_version:  1.23.5

  NOTE: Python version was forced by use_python() function
  [1] "Method: cdseq; sc-norm: ; bulk-norm: "
  Error in if (sc_norm == "counts") { : argument is of length zero
  Execution halted

Work dir:
  /home/sciliciumtheo/omnideconv_bench/work/1d/451d3a32c6358a4a0a55571454801b

Container:
  alexd13/omnideconv_benchmark:1.4

Tip: when you have fixed the problem you can continue the execution adding the option `-resume` to the run command line

 -- Check '.nextflow.log' file for details

[alex-d13]

Hi,

Sorry, I should have read your issue more carefully before.
You are trying to run methods that are part of the omnideconv R package. However, we did not include them in our benchmark of methods and, therefore, also not in the pipeline.
Currently, we are also not planning to include them here, but if you are interested in their performance, we would be highly appreciative if you want to start a pull request with the additional methods. In theory you only need to adapt the https://github.com/omnideconv/deconvBench/blob/main/pipeline/bin/computeSignaturesNF.R and https://github.com/omnideconv/deconvBench/blob/main/pipeline/bin/runDeconvolutionNF.R files and add the respective function call to omnideconv for a specific method parameter.

I hope you understand that currently, adding these additional methods is not our highest priority.

Best,
Alex

[thegrebe]

Just to inform you, I am in the process of adding these methods to deconvBench.

cpm and CDseq already implemented, currently testing momf.
I will probably finish at the beginning of next week.

[thegrebe]

momf just finished running smoothly.

Retrospectively, I don't think adding BSeq-SC is really worth it, for the following reasons:

1. this method needs a pre-defined set of marker genes, while other do not, and IMO this limits the scope of this tool;
2. it internally builds on CIBERSORT, which is a tool already supported;
3. the GitHub repo seems abandoned, no update in 8 years, broken links, issues mentioning broken links open without any response for years.

[alex-d13]

I totally agree, especially point 1) is a reason why we did not include it in our benchmark :)