Modeling continuous cell-level covariates
Introduction
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).
Standard processing
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()# only keep singlet cells with sufficient reads
sce <- sce[rowSums(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$StimStatus <- sce$stimIn 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:
Pseudobulk
Now compute the pseudobulk using standard code:
sce$id <- paste0(sce$StimStatus, sce$ind)
# Create pseudobulk
pb <- aggregateToPseudoBulk(sce,
assay = "counts",
cluster_id = "cell",
sample_id = "id",
verbose = FALSE
)The means per variable, cell type, and sample are stored in the
pseudobulk SingleCellExperiment object:
## # A tibble: 128 × 5
## # Groups: cell [8]
## cell id cluster value1 value2
## <fct> <fct> <dbl> <dbl> <dbl>
## 1 B cells ctrl101 3.96 0.113 -0.0545
## 2 B cells ctrl1015 4.00 -0.00104 0.0940
## 3 B cells ctrl1016 4 0.0600 -0.00432
## 4 B cells ctrl1039 4.04 0.172 -0.249
## 5 B cells ctrl107 4 0.0882 -0.229
## 6 B cells ctrl1244 4 0.0938 0.0530
## 7 B cells ctrl1256 4.01 -0.0171 -0.155
## 8 B cells ctrl1488 4.02 -0.0600 0.0122
## 9 B cells stim101 4.09 -0.0889 0.109
## 10 B cells stim1015 4.06 -0.0504 -0.00298
## # ℹ 118 more rowsAnalysis
Including these variables in a regression formula uses the summarized
values from the corresponding cell type. This happens behind the scenes,
so the user doesn’t need to distinguish bewteen sample-level variables
stored in colData(pb) and cell-level variables stored in
metadata(pb)$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 value2Details
A variable in colData(sce) is handled according to if
the variable is
- continuous: the mean per donor/cell type is stored in
metadata(pb)$aggr_means - discrete
- [constant within each donor/cell type] it is stored in
colData(pb) - [varies within each donor/cell type] there is no good way to summarize it. The variable is dropped.
- [constant within each donor/cell type] it is stored in
Session Info
## R version 4.5.2 (2025-10-31)
## Platform: x86_64-pc-linux-gnu
## Running under: Ubuntu 24.04.3 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/libopenblasp-r0.3.26.so; LAPACK version 3.12.0
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_US.UTF-8 LC_COLLATE=C
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## time zone: Etc/UTC
## tzcode source: system (glibc)
##
## attached base packages:
## [1] stats4 stats graphics grDevices utils datasets methods
## [8] base
##
## other attached packages:
## [1] muscData_1.24.0 scater_1.38.0
## [3] scuttle_1.20.0 ExperimentHub_3.0.0
## [5] AnnotationHub_4.0.0 BiocFileCache_3.0.0
## [7] dbplyr_2.5.1 muscat_1.24.0
## [9] dreamlet_1.8.0 SingleCellExperiment_1.32.0
## [11] SummarizedExperiment_1.40.0 Biobase_2.70.0
## [13] GenomicRanges_1.62.0 Seqinfo_1.0.0
## [15] IRanges_2.44.0 S4Vectors_0.48.0
## [17] BiocGenerics_0.56.0 generics_0.1.4
## [19] MatrixGenerics_1.22.0 matrixStats_1.5.0
## [21] variancePartition_1.40.0 BiocParallel_1.44.0
## [23] limma_3.66.0 ggplot2_4.0.0
## [25] BiocStyle_2.38.0
##
## loaded via a namespace (and not attached):
## [1] bitops_1.0-9 httr_1.4.7
## [3] RColorBrewer_1.1-3 doParallel_1.0.17
## [5] Rgraphviz_2.54.0 numDeriv_2016.8-1.1
## [7] tools_4.5.2 sctransform_0.4.2
## [9] backports_1.5.0 utf8_1.2.6
## [11] R6_2.6.1 metafor_4.8-0
## [13] mgcv_1.9-4 GetoptLong_1.0.5
## [15] withr_3.0.2 prettyunits_1.2.0
## [17] gridExtra_2.3 fdrtool_1.2.18
## [19] cli_3.6.5 sandwich_3.1-1
## [21] labeling_0.4.3 slam_0.1-55
## [23] sass_0.4.10 KEGGgraph_1.70.0
## [25] SQUAREM_2021.1 mvtnorm_1.3-3
## [27] S7_0.2.0 blme_1.0-6
## [29] mixsqp_0.3-54 zenith_1.12.0
## [31] parallelly_1.45.1 invgamma_1.2
## [33] RSQLite_2.4.4 shape_1.4.6.1
## [35] gtools_3.9.5 dplyr_1.1.4
## [37] Matrix_1.7-4 metadat_1.4-0
## [39] ggbeeswarm_0.7.2 abind_1.4-8
## [41] lifecycle_1.0.4 yaml_2.3.10
## [43] edgeR_4.8.0 mathjaxr_1.8-0
## [45] gplots_3.2.0 SparseArray_1.10.1
## [47] grid_4.5.2 blob_1.2.4
## [49] crayon_1.5.3 lattice_0.22-7
## [51] beachmat_2.26.0 msigdbr_25.1.1
## [53] annotate_1.88.0 KEGGREST_1.50.0
## [55] sys_3.4.3 maketools_1.3.2
## [57] pillar_1.11.1 knitr_1.50
## [59] ComplexHeatmap_2.26.0 rjson_0.2.23
## [61] boot_1.3-32 corpcor_1.6.10
## [63] future.apply_1.20.0 codetools_0.2-20
## [65] glue_1.8.0 data.table_1.17.8
## [67] vctrs_0.6.5 png_0.1-8
## [69] Rdpack_2.6.4 gtable_0.3.6
## [71] assertthat_0.2.1 cachem_1.1.0
## [73] zigg_0.0.2 xfun_0.54
## [75] rbibutils_2.4 S4Arrays_1.10.0
## [77] Rfast_2.1.5.2 reformulas_0.4.2
## [79] iterators_1.0.14 statmod_1.5.1
## [81] nlme_3.1-168 pbkrtest_0.5.5
## [83] bit64_4.6.0-1 filelock_1.0.3
## [85] progress_1.2.3 EnvStats_3.1.0
## [87] bslib_0.9.0 TMB_1.9.18
## [89] irlba_2.3.5.1 vipor_0.4.7
## [91] KernSmooth_2.23-26 colorspace_2.1-2
## [93] rmeta_3.0 DBI_1.2.3
## [95] DESeq2_1.50.1 tidyselect_1.2.1
## [97] bit_4.6.0 compiler_4.5.2
## [99] curl_7.0.0 httr2_1.2.1
## [101] graph_1.88.0 BiocNeighbors_2.4.0
## [103] DelayedArray_0.36.0 scales_1.4.0
## [105] caTools_1.18.3 remaCor_0.0.20
## [107] rappdirs_0.3.3 stringr_1.6.0
## [109] digest_0.6.37 minqa_1.2.8
## [111] rmarkdown_2.30 aod_1.3.3
## [113] XVector_0.50.0 RhpcBLASctl_0.23-42
## [115] htmltools_0.5.8.1 pkgconfig_2.0.3
## [117] lme4_1.1-37 sparseMatrixStats_1.22.0
## [119] lpsymphony_1.38.0 mashr_0.2.79
## [121] fastmap_1.2.0 rlang_1.1.6
## [123] GlobalOptions_0.1.2 DelayedMatrixStats_1.32.0
## [125] farver_2.1.2 jquerylib_0.1.4
## [127] IHW_1.38.0 zoo_1.8-14
## [129] jsonlite_2.0.0 BiocSingular_1.26.0
## [131] RCurl_1.98-1.17 magrittr_2.0.4
## [133] Rcpp_1.1.0 babelgene_22.9
## [135] viridis_0.6.5 EnrichmentBrowser_2.40.0
## [137] stringi_1.8.7 MASS_7.3-65
## [139] plyr_1.8.9 listenv_0.10.0
## [141] parallel_4.5.2 ggrepel_0.9.6
## [143] Biostrings_2.78.0 splines_4.5.2
## [145] hms_1.1.4 circlize_0.4.16
## [147] locfit_1.5-9.12 buildtools_1.0.0
## [149] reshape2_1.4.4 ScaledMatrix_1.18.0
## [151] BiocVersion_3.23.1 XML_3.99-0.20
## [153] evaluate_1.0.5 RcppParallel_5.1.11-1
## [155] BiocManager_1.30.26 nloptr_2.2.1
## [157] foreach_1.5.2 tidyr_1.3.1
## [159] purrr_1.2.0 future_1.67.0
## [161] clue_0.3-66 scattermore_1.2
## [163] ashr_2.2-63 rsvd_1.0.5
## [165] broom_1.0.10 xtable_1.8-4
## [167] fANCOVA_0.6-1 viridisLite_0.4.2
## [169] truncnorm_1.0-9 tibble_3.3.0
## [171] lmerTest_3.1-3 glmmTMB_1.1.13
## [173] memoise_2.0.1 beeswarm_0.4.0
## [175] AnnotationDbi_1.72.0 cluster_2.1.8.1
## [177] globals_0.18.0 GSEABase_1.72.0