Chapter 12 Bach mouse mammary gland (10X Genomics)

12.1 Introduction

This performs an analysis of the Bach et al. (2017) 10X Genomics dataset, from which we will consider a single sample of epithelial cells from the mouse mammary gland during gestation.

12.2 Data loading

library(scRNAseq)
sce.mam <- BachMammaryData(samples="G_1")

library(scater)
rownames(sce.mam) <- uniquifyFeatureNames(
    rowData(sce.mam)$Ensembl, rowData(sce.mam)$Symbol)

library(AnnotationHub)
ens.mm.v97 <- AnnotationHub()[["AH73905"]]
rowData(sce.mam)$SEQNAME <- mapIds(ens.mm.v97, keys=rowData(sce.mam)$Ensembl,
    keytype="GENEID", column="SEQNAME")

12.3 Quality control

unfiltered <- sce.mam

is.mito <- rowData(sce.mam)$SEQNAME == "MT"
stats <- perCellQCMetrics(sce.mam, subsets=list(Mito=which(is.mito)))
qc <- quickPerCellQC(stats, percent_subsets="subsets_Mito_percent")
sce.mam <- sce.mam[,!qc$discard]

colData(unfiltered) <- cbind(colData(unfiltered), stats)
unfiltered$discard <- qc$discard

gridExtra::grid.arrange(
    plotColData(unfiltered, y="sum", colour_by="discard") + 
        scale_y_log10() + ggtitle("Total count"),
    plotColData(unfiltered, y="detected", colour_by="discard") + 
        scale_y_log10() + ggtitle("Detected features"),
    plotColData(unfiltered, y="subsets_Mito_percent", 
        colour_by="discard") + ggtitle("Mito percent"),
    ncol=2
)

Distribution of each QC metric across cells in the Bach mammary gland dataset. Each point represents a cell and is colored according to whether that cell was discarded.

Figure 12.1: Distribution of each QC metric across cells in the Bach mammary gland dataset. Each point represents a cell and is colored according to whether that cell was discarded.

plotColData(unfiltered, x="sum", y="subsets_Mito_percent", 
    colour_by="discard") + scale_x_log10()

Percentage of mitochondrial reads in each cell in the Bach mammary gland dataset compared to its total count. Each point represents a cell and is colored according to whether that cell was discarded.

Figure 12.2: Percentage of mitochondrial reads in each cell in the Bach mammary gland dataset compared to its total count. Each point represents a cell and is colored according to whether that cell was discarded.

colSums(as.matrix(qc))

##              low_lib_size            low_n_features high_subsets_Mito_percent 
##                         0                         0                       143 
##                   discard 
##                       143

12.4 Normalization

library(scran)
set.seed(101000110)
clusters <- quickCluster(sce.mam)
sce.mam <- computeSumFactors(sce.mam, clusters=clusters)
sce.mam <- logNormCounts(sce.mam)

summary(sizeFactors(sce.mam))

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.264   0.520   0.752   1.000   1.207  10.790

plot(librarySizeFactors(sce.mam), sizeFactors(sce.mam), pch=16,
    xlab="Library size factors", ylab="Deconvolution factors", log="xy")

Figure 12.3: Relationship between the library size factors and the deconvolution size factors in the Bach mammary gland dataset.

12.5 Variance modelling

We use a Poisson-based technical trend to capture more genuine biological variation in the biological component.

set.seed(00010101)
dec.mam <- modelGeneVarByPoisson(sce.mam)
top.mam <- getTopHVGs(dec.mam, prop=0.1)

plot(dec.mam$mean, dec.mam$total, pch=16, cex=0.5,
    xlab="Mean of log-expression", ylab="Variance of log-expression")
curfit <- metadata(dec.mam)
curve(curfit$trend(x), col='dodgerblue', add=TRUE, lwd=2)

Per-gene variance as a function of the mean for the log-expression values in the Bach mammary gland dataset. Each point represents a gene (black) with the mean-variance trend (blue) fitted to simulated Poisson counts.

Figure 12.4: Per-gene variance as a function of the mean for the log-expression values in the Bach mammary gland dataset. Each point represents a gene (black) with the mean-variance trend (blue) fitted to simulated Poisson counts.

12.6 Dimensionality reduction

library(BiocSingular)
set.seed(101010011)
sce.mam <- denoisePCA(sce.mam, technical=dec.mam, subset.row=top.mam)
sce.mam <- runTSNE(sce.mam, dimred="PCA")

ncol(reducedDim(sce.mam, "PCA"))

## [1] 15

12.7 Clustering

We use a higher k to obtain coarser clusters (for use in doubletCluster() later).

snn.gr <- buildSNNGraph(sce.mam, use.dimred="PCA", k=25)
colLabels(sce.mam) <- factor(igraph::cluster_walktrap(snn.gr)$membership)

table(colLabels(sce.mam))

## 
##   1   2   3   4   5   6   7   8   9  10 
## 550 847 639 477  54  88  39  22  32  24

plotTSNE(sce.mam, colour_by="label")

Obligatory $t$-SNE plot of the Bach mammary gland dataset, where each point represents a cell and is colored according to the assigned cluster.

Figure 12.5: Obligatory $t$-SNE plot of the Bach mammary gland dataset, where each point represents a cell and is colored according to the assigned cluster.

Session Info

R version 4.5.1 (2025-06-13)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 24.04.3 LTS

Matrix products: default
BLAS/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_GB              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: America/New_York
tzcode source: system (glibc)

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

other attached packages:
 [1] BiocSingular_1.25.0         scran_1.37.0               
 [3] AnnotationHub_3.99.6        BiocFileCache_2.99.5       
 [5] dbplyr_2.5.0                scater_1.37.0              
 [7] ggplot2_3.5.2               scuttle_1.19.0             
 [9] ensembldb_2.33.1            AnnotationFilter_1.33.0    
[11] GenomicFeatures_1.61.6      AnnotationDbi_1.71.1       
[13] scRNAseq_2.23.0             SingleCellExperiment_1.31.1
[15] SummarizedExperiment_1.39.1 Biobase_2.69.0             
[17] GenomicRanges_1.61.1        Seqinfo_0.99.2             
[19] IRanges_2.43.0              S4Vectors_0.47.0           
[21] BiocGenerics_0.55.1         generics_0.1.4             
[23] MatrixGenerics_1.21.0       matrixStats_1.5.0          
[25] BiocStyle_2.37.1            rebook_1.19.0              

loaded via a namespace (and not attached):
  [1] RColorBrewer_1.1-3       jsonlite_2.0.0           CodeDepends_0.6.6       
  [4] magrittr_2.0.3           ggbeeswarm_0.7.2         gypsum_1.5.0            
  [7] farver_2.1.2             rmarkdown_2.29           BiocIO_1.19.0           
 [10] vctrs_0.6.5              memoise_2.0.1            Rsamtools_2.25.2        
 [13] RCurl_1.98-1.17          htmltools_0.5.8.1        S4Arrays_1.9.1          
 [16] curl_6.4.0               BiocNeighbors_2.3.1      Rhdf5lib_1.31.0         
 [19] SparseArray_1.9.1        rhdf5_2.53.4             sass_0.4.10             
 [22] alabaster.base_1.9.5     bslib_0.9.0              alabaster.sce_1.9.0     
 [25] httr2_1.2.1              cachem_1.1.0             GenomicAlignments_1.45.2
 [28] igraph_2.1.4             lifecycle_1.0.4          pkgconfig_2.0.3         
 [31] rsvd_1.0.5               Matrix_1.7-3             R6_2.6.1                
 [34] fastmap_1.2.0            digest_0.6.37            dqrng_0.4.1             
 [37] irlba_2.3.5.1            ExperimentHub_2.99.5     RSQLite_2.4.2           
 [40] beachmat_2.25.4          labeling_0.4.3           filelock_1.0.3          
 [43] httr_1.4.7               abind_1.4-8              compiler_4.5.1          
 [46] bit64_4.6.0-1            withr_3.0.2              BiocParallel_1.43.4     
 [49] viridis_0.6.5            DBI_1.2.3                HDF5Array_1.37.0        
 [52] alabaster.ranges_1.9.1   alabaster.schemas_1.9.0  rappdirs_0.3.3          
 [55] DelayedArray_0.35.2      bluster_1.19.0           rjson_0.2.23            
 [58] tools_4.5.1              vipor_0.4.7              beeswarm_0.4.0          
 [61] glue_1.8.0               h5mread_1.1.1            restfulr_0.0.16         
 [64] rhdf5filters_1.21.0      grid_4.5.1               Rtsne_0.17              
 [67] cluster_2.1.8.1          gtable_0.3.6             metapod_1.17.0          
 [70] ScaledMatrix_1.17.0      XVector_0.49.0           ggrepel_0.9.6           
 [73] BiocVersion_3.22.0       pillar_1.11.0            limma_3.65.3            
 [76] dplyr_1.1.4              lattice_0.22-7           rtracklayer_1.69.1      
 [79] bit_4.6.0                tidyselect_1.2.1         locfit_1.5-9.12         
 [82] Biostrings_2.77.2        knitr_1.50               gridExtra_2.3           
 [85] bookdown_0.43            ProtGenerics_1.41.0      edgeR_4.7.3             
 [88] xfun_0.52                statmod_1.5.0            UCSC.utils_1.5.0        
 [91] lazyeval_0.2.2           yaml_2.3.10              evaluate_1.0.4          
 [94] codetools_0.2-20         tibble_3.3.0             alabaster.matrix_1.9.0  
 [97] BiocManager_1.30.26      graph_1.87.0             cli_3.6.5               
[100] jquerylib_0.1.4          dichromat_2.0-0.1        Rcpp_1.1.0              
[103] GenomeInfoDb_1.45.9      dir.expiry_1.17.0        png_0.1-8               
[106] XML_3.99-0.18            parallel_4.5.1           blob_1.2.4              
[109] bitops_1.0-9             viridisLite_0.4.2        alabaster.se_1.9.0      
[112] scales_1.4.0             purrr_1.1.0              crayon_1.5.3            
[115] rlang_1.1.6              cowplot_1.2.0            KEGGREST_1.49.1

References

Bach, K., S. Pensa, M. Grzelak, J. Hadfield, D. J. Adams, J. C. Marioni, and W. T. Khaled. 2017. “Differentiation dynamics of mammary epithelial cells revealed by single-cell RNA sequencing.” Nat Commun 8 (1): 2128.