seurat-wrappers
seurat-wrappers copied to clipboard
Published
20 hours ago •
satijalab
satijalab
Unexpected behaviour
Hello thanks for the tool.
When I run liger through SeuratWrappers regressing for (i) total reads per cell and (ii) mitochondrial percent I get a stringy UMAP plot. While I get a more normal looking UMAP plot if I don't regress for any variable.
Could you please advise?
WithOUT regressing variables. Looging "normal"
cancer_lymphoid %>%
NormalizeData() %>%
FindVariableFeatures() %>%
ScaleData(split.by = "sample", do.center = FALSE) %>%
SeuratWrappers::RunOptimizeALS(k = 20, lambda = 5, split.by = "sample") %>%
SeuratWrappers::RunQuantileNorm(split.by = "sample") %>%
RunUMAP(dims = 1:20, reduction = "iNMF", n.components = 3L) %>%
ggplot(aes(UMAP_1, UMAP_2)) +
geom_point(size = 0.2, shape=".")
Regressing for variables. Looking stringy
cancer_lymphoid %>%
NormalizeData() %>%
FindVariableFeatures() %>%
ScaleData(split.by = "sample", do.center = FALSE, vars.to.regress = c("sum", "subsets_Mito_percent")) %>%
SeuratWrappers::RunOptimizeALS(k = 20, lambda = 5, split.by = "sample") %>%
SeuratWrappers::RunQuantileNorm(split.by = "sample") %>%
RunUMAP(dims = 1:20, reduction = "iNMF", n.components = 3L) %>%
ggplot(aes(UMAP_1, UMAP_2)) +
geom_point(size = 0.2, shape=".")
Sessioninfo
> sessionInfo()
R version 4.0.3 (2020-10-10)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)
Matrix products: default
BLAS: /stornext/System/data/apps/R/R-4.0.3/lib64/R/lib/libRblas.so
LAPACK: /stornext/System/data/apps/R/R-4.0.3/lib64/R/lib/libRlapack.so
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=en_US.UTF-8
[4] LC_COLLATE=en_US.UTF-8 LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C LC_ADDRESS=C
[10] LC_TELEPHONE=C LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] parallel stats4 stats graphics grDevices utils datasets methods base
other attached packages:
[1] scales_1.1.1 functional_0.6 SingleR_1.4.0
[4] SeuratWrappers_0.3.0 liger_0.5.0 patchwork_1.1.0
[7] Matrix_1.2-18 cowplot_1.1.0 furrr_0.2.1
[10] future_1.21.0 googlesheets4_0.2.0 nanny_0.1.8
[13] tidybulk_1.2.1 DropletUtils_1.10.0 scran_1.18.1
[16] scater_1.18.3 SingleCellExperiment_1.12.0 ggExtra_0.9
[19] RColorBrewer_1.1-2 tidyHeatmap_1.1.5 tidysc_0.1.0
[22] forcats_0.5.0 stringr_1.4.0 dplyr_1.0.2
[25] purrr_0.3.4 readr_1.4.0 tidyr_1.1.2
[28] tibble_3.0.4 ggplot2_3.3.2 tidyverse_1.3.0
[31] tidyseurat_0.1.18 Seurat_3.2.2 SummarizedExperiment_1.20.0
[34] Biobase_2.50.0 GenomicRanges_1.42.0 GenomeInfoDb_1.26.0
[37] IRanges_2.24.0 S4Vectors_0.28.0 BiocGenerics_0.36.0
[40] MatrixGenerics_1.2.0 matrixStats_0.57.0
loaded via a namespace (and not attached):
[1] rappdirs_0.3.1 widyr_0.1.3 R.methodsS3_1.8.1
[4] bit64_4.0.5 irlba_2.3.3 DelayedArray_0.16.0
[7] R.utils_2.10.1 data.table_1.13.2 rpart_4.1-15
[10] RCurl_1.98-1.2 generics_0.1.0 snow_0.4-3
[13] preprocessCore_1.52.0 RSQLite_2.2.1 RANN_2.6.1
[16] bit_4.0.4 spatstat.data_1.5-2 xml2_1.3.2
[19] lubridate_1.7.9.2 httpuv_1.5.4 assertthat_0.2.1
[22] viridis_0.5.1 celldex_1.0.0 hms_0.5.3
[25] promises_1.1.1 fansi_0.4.1 dbplyr_2.0.0
[28] readxl_1.3.1 igraph_1.2.6 DBI_1.1.0
[31] htmlwidgets_1.5.2 googledrive_1.0.1 ellipsis_0.3.1
[34] riverplot_0.6 RSpectra_0.16-0 backports_1.2.0
[37] deldir_0.2-3 sparseMatrixStats_1.2.0 vctrs_0.3.5
[40] remotes_2.2.0 Cairo_1.5-12.2 ROCR_1.0-11
[43] abind_1.4-5 withr_2.3.0 sctransform_0.3.1
[46] dittoSeq_1.2.0 mclust_5.4.7 goftest_1.2-2
[49] cluster_2.1.0 ExperimentHub_1.16.0 lazyeval_0.2.2
[52] crayon_1.3.4 labeling_0.4.2 edgeR_3.32.0
[55] pkgconfig_2.0.3 nlme_3.1-150 vipor_0.4.5
[58] rlang_0.4.8 globals_0.14.0 lifecycle_0.2.0
[61] miniUI_0.1.1.1 BiocFileCache_1.14.0 doSNOW_1.0.19
[64] modelr_0.1.8 rsvd_1.0.3 AnnotationHub_2.22.0
[67] cellranger_1.1.0 polyclip_1.10-0 lmtest_0.9-38
[70] Rhdf5lib_1.12.0 zoo_1.8-8 reprex_0.3.0
[73] beeswarm_0.2.3 ggridges_0.5.2 GlobalOptions_0.1.2
[76] pheatmap_1.0.12 png_0.1-7 viridisLite_0.3.0
[79] rjson_0.2.20 bitops_1.0-6 R.oo_1.24.0
[82] KernSmooth_2.23-18 rhdf5filters_1.2.0 blob_1.2.1
[85] DelayedMatrixStats_1.12.0 shape_1.4.5 parallelly_1.21.0
[88] beachmat_2.6.1 memoise_1.1.0 magrittr_2.0.1
[91] plyr_1.8.6 ica_1.0-2 zlibbioc_1.36.0
[94] compiler_4.0.3 dqrng_0.2.1 clue_0.3-57
[97] fitdistrplus_1.1-1 cli_2.2.0 XVector_0.30.0
[100] listenv_0.8.0 pbapply_1.4-3 MASS_7.3-53
[103] mgcv_1.8-33 tidyselect_1.1.0 stringi_1.5.3
[106] yaml_2.2.1 BiocSingular_1.6.0 locfit_1.5-9.4
[109] ggrepel_0.8.2 grid_4.0.3 tools_4.0.3
[112] future.apply_1.6.0 circlize_0.4.11 rstudioapi_0.13
[115] bluster_1.0.0 foreach_1.5.1 gridExtra_2.3
[118] farver_2.0.3 Rtsne_0.15 digest_0.6.27
[121] BiocManager_1.30.10 FNN_1.1.3 shiny_1.5.0
[124] Rcpp_1.0.5 broom_0.7.2 scuttle_1.0.0
[127] BiocVersion_3.12.0 later_1.1.0.1 RcppAnnoy_0.0.17
[130] httr_1.4.2 AnnotationDbi_1.52.0 ComplexHeatmap_2.6.1
[133] colorspace_2.0-0 rvest_0.3.6 fs_1.5.0
[136] tensor_1.5 reticulate_1.18 splines_4.0.3
[139] uwot_0.1.9 statmod_1.4.35 spatstat.utils_1.17-0
[142] plotly_4.9.2.1 xtable_1.8-4 jsonlite_1.7.1
[145] spatstat_1.64-1 R6_2.5.0 pillar_1.4.7
[148] htmltools_0.5.0 mime_0.9 glue_1.4.2
[151] fastmap_1.0.1 BiocParallel_1.24.1 BiocNeighbors_1.8.0
[154] interactiveDisplayBase_1.28.0 codetools_0.2-18 utf8_1.1.4
[157] lattice_0.20-41 curl_4.3 ggbeeswarm_0.6.0
[160] leiden_0.3.5 gtools_3.8.2 survival_3.2-7
[163] limma_3.46.0 munsell_0.5.0 GetoptLong_1.0.4
[166] rhdf5_2.34.0 GenomeInfoDbData_1.2.4 iterators_1.0.13
[169] HDF5Array_1.18.0 haven_2.3.1 reshape2_1.4.4
[172] gtable_0.3.0
I found that possibly regressing here is not the answer, but discarting some dimension might be.
https://github.com/welch-lab/liger/issues/113