Leiden Clustering

[sophmeye]

Describe the bug After preprocessing and reducing dimensions on my cds, I am unable to cluster using the default leiden clustering method. I am given the error: Error in leidenbase::leiden_find_partition(graph_result[["g"]], partition_type = partition_type, : REAL() can only be applied to a 'numeric', not a 'NULL'

I am, however, able to use the louvain clustering method without any errors.

To Reproduce cluster_cells(cds, reduction_method = "UMAP", resolution = 1e-5)

traceback() 3: leidenbase::leiden_find_partition(graph_result[["g"]], partition_type = partition_type, initial_membership = initial_membership, edge_weights = edge_weights, node_sizes = node_sizes, seed = random_seed, resolution_parameter = cur_resolution_parameter, num_iter = num_iter, verbose = verbose) 2: leiden_clustering(data = reduced_dim_res, pd = colData(cds), weight = weight, nn_index = nn_index, k = k, nn_control = nn_control, num_iter = num_iter, resolution_parameter = resolution, random_seed = random_seed, verbose = verbose, ...) 1: cluster_cells(cds_CZ_x_fake, reduction_method = "UMAP", resolution = 1e-05)

Expected behavior Leidenbase should have run community detection

sessionInfo(): R version 4.3.0 (2023-04-21 ucrt) Platform: x86_64-w64-mingw32/x64 (64-bit) Running under: Windows 10 x64 (build 19045)

Matrix products: default

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

time zone: America/New_York tzcode source: internal

attached base packages: [1] stats4 stats graphics grDevices utils datasets methods base

other attached packages: [1] leidenbase_0.1.18 monocle3_1.3.1 SingleCellExperiment_1.22.0 SummarizedExperiment_1.30.2 GenomicRanges_1.52.0 GenomeInfoDb_1.36.0 IRanges_2.34.0
[8] S4Vectors_0.38.1 MatrixGenerics_1.12.2 matrixStats_1.0.0 Biobase_2.60.0 BiocGenerics_0.46.0

loaded via a namespace (and not attached): [1] gtable_0.3.3 ggplot2_3.4.2 ggrepel_0.9.3 lattice_0.21-8 vctrs_0.6.2 tools_4.3.0 bitops_1.0-7 generics_0.1.3
[9] parallel_4.3.0 tibble_3.2.1 fansi_1.0.4 pkgconfig_2.0.3 Matrix_1.5-4 sparseMatrixStats_1.12.0 assertthat_0.2.1 lifecycle_1.0.3
[17] GenomeInfoDbData_1.2.10 stringr_1.5.0 farver_2.1.1 compiler_4.3.0 munsell_0.5.0 terra_1.7-37 codetools_0.2-19 RCurl_1.98-1.12
[25] pillar_1.9.0 nloptr_2.0.3 crayon_1.5.2 MASS_7.3-58.4 uwot_0.1.14 DelayedArray_0.26.3 viridis_0.6.3 boot_1.3-28.1
[33] nlme_3.1-162 parallelly_1.36.0 tidyselect_1.2.0 digest_0.6.31 stringi_1.7.12 future_1.32.0 reshape2_1.4.4 dplyr_1.1.2
[41] listenv_0.9.0 labeling_0.4.2 splines_4.3.0 grid_4.3.0 colorspace_2.1-0 cli_3.6.1 magrittr_2.0.3 S4Arrays_1.0.4
[49] utf8_1.2.3 withr_2.5.0 DelayedMatrixStats_1.22.1 scales_1.2.1 XVector_0.40.0 globals_0.16.2 igraph_1.5.0 lme4_1.1-33
[57] gridExtra_2.3 viridisLite_0.4.2 irlba_2.3.5.1 RcppAnnoy_0.0.20 rlang_1.1.1 Rcpp_1.0.10 glue_1.6.2 rstudioapi_0.14
[65] minqa_1.2.5 R6_2.5.1 plyr_1.8.8 zlibbioc_1.46.0

Additional context I have previously clustered this same dataset successfully with the exact parameters stated.

[pyjm]

I am having the same error! Any thoughts on this would be very helpful.

> cds <- cluster_cells(cds)
Error in leidenbase::leiden_find_partition(graph_result[["g"]], partition_type = partition_type,  : 
  REAL() can only be applied to a 'numeric', not a 'NULL'
> traceback()
3: leidenbase::leiden_find_partition(graph_result[["g"]], partition_type = partition_type, 
       initial_membership = initial_membership, edge_weights = edge_weights, 
       node_sizes = node_sizes, seed = random_seed, resolution_parameter = cur_resolution_parameter, 
       num_iter = num_iter, verbose = verbose)
2: leiden_clustering(data = reduced_dim_res, pd = colData(cds), 
       weight = weight, nn_index = nn_index, k = k, nn_control = nn_control, 
       num_iter = num_iter, resolution_parameter = resolution, random_seed = random_seed, 
       verbose = verbose, ...)
1: cluster_cells(cds)

> utils::sessionInfo()
R version 4.2.3 (2023-03-15)
Platform: x86_64-conda-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Thank you!!

[ghost]

Same error too. Monocle3 is the 1.3.1 version. It can work on the linux server system, but not work on my macos computer. The input file is the same.

[QiuyuLian]

Same error too.

[QiuyuLian]

I found if I specified "cluster_method = 'louvain'", it would work.... There might be something wrong with default parameter using leiden.

[pyjm]

Thanks @QiuyuLian!! Changing to "louvain" also works for me.

[iichelhadi]

same issue but what if I want to use leiden instead of louvain?

[cierra-walker]

same issue, following.

[eegerber]

same issue, following.

[sands58]

Downgrading the igraph package to version 1.4.3 can solve this problem.

[eegerber]

Downgrading the igraph package to version 1.4.3 can solve this problem.

Thank you sands58!! This worked for me.

[iichelhadi]

Downgrading the igraph package to version 1.4.3 can solve this problem.

worked for me as well

[chanucy97]

Thank you@Qiuyu Lian

[Daisyxiaozhou]

Thanks @QiuyuLian!! Changing to "louvain" also works for me. What version of R is yours, please?

[Daisyxiaozhou]

Downgrading the igraph package to version 1.4.3 can solve this problem.

worked for me as well

What version of R is yours, please?

[Krithika-Bhuvan]

For anyone interested this is the code packageurl <- "https://cran.r-project.org/src/contrib/Archive/igraph/igraph_1.4.3.tar.gz" install.packages(packageurl, repos=NULL, type="source")

Still getting the error even after downgrading

[mesposit]

For anyone interested this is the code packageurl <- "https://cran.r-project.org/src/contrib/Archive/igraph/igraph_1.4.3.tar.gz" install.packages(packageurl, repos=NULL, type="source")

Still getting the error even after downgrading

Have you tried restarting your R session and then loading all the necessary packages again?

[Krithika-Bhuvan]

Still tried that. It fails to load igraph Reason: image not found Error: loading failed Execution halted

Also tried to install monocle3 on google colab R , and that one says that it does not work with its R version

[brgew]

Hi,

The latest leidenbase version 0.1.25 should fix this problem. It can be installed from CRAN.

Best Wishes, Brent

[LiMo-bioxer]

This method is work well, many thanks.