Since read counts are summed across cells in a pseudobulk approach, modeling continuous cell-level covariates also requires a collapsing step. Here we summarize the values of a variable from a set of cells using the mean, and store the value for each cell type. Including these variables in a regression formula uses the summarized values from the corresponding cell type.
We demonstrate this feature on a lightly modified analysis of PBMCs from 8 individuals stimulated with interferon-β (Kang, et al, 2018, Nature Biotech).
Here is the code from the main vignette:
library(dreamlet)
library(muscat)
library(ExperimentHub)
library(scater)
# Download data, specifying EH2259 for the Kang, et al study
eh <- ExperimentHub()
sce <- eh[["EH2259"]]
# only keep singlet cells with sufficient reads
sce <- sce[rowSums2(counts(sce) > 0) > 0, ]
sce <- sce[, colData(sce)$multiplets == "singlet"]
# compute QC metrics
qc <- perCellQCMetrics(sce)
# remove cells with few or many detected genes
ol <- isOutlier(metric = qc$detected, nmads = 2, log = TRUE)
sce <- sce[, !ol]
# set variable indicating stimulated (stim) or control (ctrl)
sce$StimStatus <- sce$stim
In many datasets, continuous cell-level variables could be mapped reads, gene count, mitochondrial rate, etc. There are no continuous cell-level variables in this dataset, so we can simulate two from a normal distribution:
sce$value1 <- rnorm(ncol(sce))
sce$value2 <- rnorm(ncol(sce))
Now compute the pseudobulk using standard code:
sce$id <- paste0(sce$StimStatus, sce$ind)
# Create pseudobulk
pb <- aggregateToPseudoBulk(sce,
assay = "counts",
cluster_id = "cell",
sample_id = "id",
verbose = FALSE
)
The means per variable, cell type, and sample are stored in the pseudobulk SingleCellExperiment
object:
metadata(pb)$aggr_means
## # A tibble: 128 × 5
## # Groups: cell [8]
## cell id cluster value1 value2
## <fct> <fct> <dbl> <dbl> <dbl>
## 1 B cells ctrl101 3.96 -0.102 -0.0989
## 2 B cells ctrl1015 4.00 0.0305 -0.00961
## 3 B cells ctrl1016 4 0.146 0.115
## 4 B cells ctrl1039 4.04 -0.0196 0.149
## 5 B cells ctrl107 4 0.173 0.0838
## 6 B cells ctrl1244 4 0.0573 -0.0564
## 7 B cells ctrl1256 4.01 0.125 0.0516
## 8 B cells ctrl1488 4.02 0.0351 0.0560
## 9 B cells stim101 4.09 0.00692 0.109
## 10 B cells stim1015 4.06 0.00291 -0.00965
## # ℹ 118 more rows
Including these variables in a regression formula uses the summarized values from the corresponding cell type. This happens behind the scenes, so the user doesn’t need to distinguish bewteen sample-level variables stored in colData(pb)
and cell-level variables stored in metadata(pb)$aggr_means
.
Variance partition and hypothesis testing proceeds as ususal:
form <- ~ StimStatus + value1 + value2
# Normalize and apply voom/voomWithDreamWeights
res.proc <- processAssays(pb, form, min.count = 5)
# run variance partitioning analysis
vp.lst <- fitVarPart(res.proc, form)
# Summarize variance fractions genome-wide for each cell type
plotVarPart(vp.lst, label.angle = 60)
# Differential expression analysis within each assay
res.dl <- dreamlet(res.proc, form)
# dreamlet results include coefficients for value1 and value2
res.dl
## class: dreamletResult
## assays(8): B cells CD14+ Monocytes ... Megakaryocytes NK cells
## Genes:
## min: 164
## max: 5262
## details(7): assay n_retain ... n_errors error_initial
## coefNames(4): (Intercept) StimStatusstim value1 value2
## R version 4.3.2 (2023-10-31)
## Platform: aarch64-apple-darwin20 (64-bit)
## Running under: macOS Ventura 13.6.1
##
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRlapack.dylib; LAPACK version 3.11.0
##
## locale:
## [1] C/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## time zone: America/New_York
## tzcode source: internal
##
## attached base packages:
## [1] stats4 stats graphics grDevices utils datasets methods
## [8] base
##
## other attached packages:
## [1] muscData_1.16.0 scater_1.30.1
## [3] scuttle_1.12.0 ExperimentHub_2.10.0
## [5] AnnotationHub_3.10.0 BiocFileCache_2.10.1
## [7] dbplyr_2.4.0 muscat_1.16.0
## [9] dreamlet_1.0.3 SingleCellExperiment_1.24.0
## [11] SummarizedExperiment_1.32.0 Biobase_2.62.0
## [13] GenomicRanges_1.54.1 GenomeInfoDb_1.38.6
## [15] IRanges_2.36.0 S4Vectors_0.40.2
## [17] BiocGenerics_0.48.1 MatrixGenerics_1.14.0
## [19] matrixStats_1.0.0 variancePartition_1.32.5
## [21] BiocParallel_1.36.0 limma_3.58.1
## [23] ggplot2_3.4.4 BiocStyle_2.30.0
##
## loaded via a namespace (and not attached):
## [1] bitops_1.0-7 httr_1.4.7
## [3] RColorBrewer_1.1-3 doParallel_1.0.17
## [5] Rgraphviz_2.46.0 numDeriv_2016.8-1.1
## [7] tools_4.3.2 sctransform_0.4.1
## [9] backports_1.4.1 utf8_1.2.4
## [11] R6_2.5.1 metafor_4.4-0
## [13] mgcv_1.9-0 GetoptLong_1.0.5
## [15] withr_2.5.2 prettyunits_1.2.0
## [17] gridExtra_2.3 cli_3.6.1
## [19] sandwich_3.0-2 labeling_0.4.3
## [21] sass_0.4.7 KEGGgraph_1.62.0
## [23] SQUAREM_2021.1 mvtnorm_1.2-3
## [25] blme_1.0-5 mixsqp_0.3-48
## [27] zenith_1.4.2 parallelly_1.36.0
## [29] invgamma_1.1 RSQLite_2.3.2
## [31] generics_0.1.3 shape_1.4.6
## [33] gtools_3.9.4 dplyr_1.1.3
## [35] Matrix_1.6-1.1 metadat_1.2-0
## [37] ggbeeswarm_0.7.2 fansi_1.0.5
## [39] abind_1.4-5 lifecycle_1.0.3
## [41] multcomp_1.4-25 yaml_2.3.7
## [43] edgeR_4.0.16 mathjaxr_1.6-0
## [45] gplots_3.1.3 SparseArray_1.2.4
## [47] grid_4.3.2 blob_1.2.4
## [49] promises_1.2.1 crayon_1.5.2
## [51] lattice_0.22-5 beachmat_2.18.1
## [53] msigdbr_7.5.1 annotate_1.80.0
## [55] KEGGREST_1.42.0 magick_2.8.1
## [57] pillar_1.9.0 knitr_1.45
## [59] ComplexHeatmap_2.18.0 rjson_0.2.21
## [61] boot_1.3-28.1 estimability_1.4.1
## [63] corpcor_1.6.10 future.apply_1.11.0
## [65] codetools_0.2-19 glue_1.6.2
## [67] data.table_1.14.8 vctrs_0.6.4
## [69] png_0.1-8 Rdpack_2.5
## [71] gtable_0.3.4 assertthat_0.2.1
## [73] cachem_1.0.8 xfun_0.41
## [75] mime_0.12 rbibutils_2.2.16
## [77] S4Arrays_1.2.0 Rfast_2.0.9
## [79] coda_0.19-4 survival_3.5-7
## [81] iterators_1.0.14 statmod_1.5.0
## [83] ellipsis_0.3.2 interactiveDisplayBase_1.40.0
## [85] TH.data_1.1-2 nlme_3.1-163
## [87] pbkrtest_0.5.2 bit64_4.0.5
## [89] filelock_1.0.2 progress_1.2.2
## [91] EnvStats_2.8.1 bslib_0.5.1
## [93] TMB_1.9.6 irlba_2.3.5.1
## [95] vipor_0.4.5 KernSmooth_2.23-22
## [97] colorspace_2.1-0 rmeta_3.0
## [99] DBI_1.1.3 DESeq2_1.42.0
## [101] tidyselect_1.2.0 emmeans_1.8.9
## [103] curl_5.1.0 bit_4.0.5
## [105] compiler_4.3.2 graph_1.80.0
## [107] BiocNeighbors_1.20.2 DelayedArray_0.28.0
## [109] bookdown_0.36 scales_1.2.1
## [111] caTools_1.18.2 remaCor_0.0.16
## [113] rappdirs_0.3.3 stringr_1.5.0
## [115] digest_0.6.33 minqa_1.2.6
## [117] rmarkdown_2.25 aod_1.3.2
## [119] XVector_0.42.0 RhpcBLASctl_0.23-42
## [121] htmltools_0.5.7 pkgconfig_2.0.3
## [123] lme4_1.1-35 sparseMatrixStats_1.14.0
## [125] highr_0.10 mashr_0.2.79
## [127] fastmap_1.1.1 rlang_1.1.1
## [129] GlobalOptions_0.1.2 shiny_1.7.5.1
## [131] DelayedMatrixStats_1.24.0 farver_2.1.1
## [133] jquerylib_0.1.4 zoo_1.8-12
## [135] jsonlite_1.8.7 BiocSingular_1.18.0
## [137] RCurl_1.98-1.13 magrittr_2.0.3
## [139] GenomeInfoDbData_1.2.11 munsell_0.5.0
## [141] Rcpp_1.0.11 babelgene_22.9
## [143] viridis_0.6.4 EnrichmentBrowser_2.32.0
## [145] RcppZiggurat_0.1.6 stringi_1.7.12
## [147] zlibbioc_1.48.0 MASS_7.3-60
## [149] plyr_1.8.9 listenv_0.9.0
## [151] parallel_4.3.2 ggrepel_0.9.4
## [153] Biostrings_2.70.2 splines_4.3.2
## [155] hms_1.1.3 circlize_0.4.15
## [157] locfit_1.5-9.8 reshape2_1.4.4
## [159] ScaledMatrix_1.10.0 BiocVersion_3.18.1
## [161] XML_3.99-0.15 evaluate_0.23
## [163] RcppParallel_5.1.7 BiocManager_1.30.22
## [165] httpuv_1.6.12 nloptr_2.0.3
## [167] foreach_1.5.2 tidyr_1.3.0
## [169] purrr_1.0.2 future_1.33.0
## [171] clue_0.3-65 scattermore_1.2
## [173] ashr_2.2-63 rsvd_1.0.5
## [175] broom_1.0.5 xtable_1.8-4
## [177] fANCOVA_0.6-1 later_1.3.1
## [179] viridisLite_0.4.2 truncnorm_1.0-9
## [181] tibble_3.2.1 lmerTest_3.1-3
## [183] glmmTMB_1.1.8 memoise_2.0.1
## [185] beeswarm_0.4.0 AnnotationDbi_1.64.1
## [187] cluster_2.1.4 globals_0.16.2
## [189] GSEABase_1.64.0