Chapter 6 Muraro human pancreas (CEL-seq)
6.1 Introduction
This performs an analysis of the Muraro et al. (2016) CEL-seq dataset, consisting of human pancreas cells from various donors.
6.2 Data loading
Converting back to Ensembl identifiers.
library(AnnotationHub)
edb <- AnnotationHub()[["AH73881"]]
gene.symb <- sub("__chr.*$", "", rownames(sce.muraro))
gene.ids <- mapIds(edb, keys=gene.symb,
keytype="SYMBOL", column="GENEID")
# Removing duplicated genes or genes without Ensembl IDs.
keep <- !is.na(gene.ids) & !duplicated(gene.ids)
sce.muraro <- sce.muraro[keep,]
rownames(sce.muraro) <- gene.ids[keep]
6.3 Quality control
This dataset lacks mitochondrial genes so we will do without. For the one batch that seems to have a high proportion of low-quality cells, we compute an appropriate filter threshold using a shared median and MAD from the other batches (Figure 6.1).
library(scater)
stats <- perCellQCMetrics(sce.muraro)
qc <- quickPerCellQC(stats, percent_subsets="altexps_ERCC_percent",
batch=sce.muraro$donor, subset=sce.muraro$donor!="D28")
sce.muraro <- sce.muraro[,!qc$discard]
colData(unfiltered) <- cbind(colData(unfiltered), stats)
unfiltered$discard <- qc$discard
gridExtra::grid.arrange(
plotColData(unfiltered, x="donor", y="sum", colour_by="discard") +
scale_y_log10() + ggtitle("Total count"),
plotColData(unfiltered, x="donor", y="detected", colour_by="discard") +
scale_y_log10() + ggtitle("Detected features"),
plotColData(unfiltered, x="donor", y="altexps_ERCC_percent",
colour_by="discard") + ggtitle("ERCC percent"),
ncol=2
)

Figure 6.1: Distribution of each QC metric across cells from each donor in the Muraro pancreas dataset. Each point represents a cell and is colored according to whether that cell was discarded.
We have a look at the causes of removal:
## low_lib_size low_n_features high_altexps_ERCC_percent
## 663 700 738
## discard
## 773
6.4 Normalization
library(scran)
set.seed(1000)
clusters <- quickCluster(sce.muraro)
sce.muraro <- computeSumFactors(sce.muraro, clusters=clusters)
sce.muraro <- logNormCounts(sce.muraro)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0878 0.5411 0.8208 1.0000 1.2108 13.9869
plot(librarySizeFactors(sce.muraro), sizeFactors(sce.muraro), pch=16,
xlab="Library size factors", ylab="Deconvolution factors", log="xy")

Figure 6.2: Relationship between the library size factors and the deconvolution size factors in the Muraro pancreas dataset.
6.5 Variance modelling
We block on a combined plate and donor factor.
block <- paste0(sce.muraro$plate, "_", sce.muraro$donor)
dec.muraro <- modelGeneVarWithSpikes(sce.muraro, "ERCC", block=block)
top.muraro <- getTopHVGs(dec.muraro, prop=0.1)
par(mfrow=c(8,4))
blocked.stats <- dec.muraro$per.block
for (i in colnames(blocked.stats)) {
current <- blocked.stats[[i]]
plot(current$mean, current$total, main=i, pch=16, cex=0.5,
xlab="Mean of log-expression", ylab="Variance of log-expression")
curfit <- metadata(current)
points(curfit$mean, curfit$var, col="red", pch=16)
curve(curfit$trend(x), col='dodgerblue', add=TRUE, lwd=2)
}

Figure 6.3: Per-gene variance as a function of the mean for the log-expression values in the Muraro pancreas dataset. Each point represents a gene (black) with the mean-variance trend (blue) fitted to the spike-in transcripts (red) separately for each donor.
6.6 Data integration
library(batchelor)
set.seed(1001010)
merged.muraro <- fastMNN(sce.muraro, subset.row=top.muraro,
batch=sce.muraro$donor)
We use the proportion of variance lost as a diagnostic measure:
## D28 D29 D30 D31
## [1,] 0.060847 0.024121 0.000000 0.00000
## [2,] 0.002646 0.003018 0.062421 0.00000
## [3,] 0.003449 0.002641 0.002598 0.08162
6.8 Clustering
snn.gr <- buildSNNGraph(merged.muraro, use.dimred="corrected")
colLabels(merged.muraro) <- factor(igraph::cluster_walktrap(snn.gr)$membership)
tab <- table(Cluster=colLabels(merged.muraro), CellType=sce.muraro$label)
library(pheatmap)
pheatmap(log10(tab+10), color=viridis::viridis(100))

Figure 6.4: Heatmap of the frequency of cells from each cell type label in each cluster.
## Donor
## Cluster D28 D29 D30 D31
## 1 104 6 57 112
## 2 59 21 77 97
## 3 12 75 64 43
## 4 28 149 126 120
## 5 87 261 277 214
## 6 21 7 54 26
## 7 1 6 6 37
## 8 6 6 5 2
## 9 11 68 5 30
## 10 4 2 5 8
gridExtra::grid.arrange(
plotTSNE(merged.muraro, colour_by="label"),
plotTSNE(merged.muraro, colour_by="batch"),
ncol=2
)

Figure 6.5: Obligatory \(t\)-SNE plots of the Muraro pancreas dataset. Each point represents a cell that is colored by cluster (left) or batch (right).
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] pheatmap_1.0.13 batchelor_1.25.0
[3] scran_1.37.0 scater_1.37.0
[5] ggplot2_3.5.2 scuttle_1.19.0
[7] ensembldb_2.33.1 AnnotationFilter_1.33.0
[9] GenomicFeatures_1.61.6 AnnotationDbi_1.71.1
[11] AnnotationHub_3.99.6 BiocFileCache_2.99.5
[13] dbplyr_2.5.0 scRNAseq_2.23.0
[15] SingleCellExperiment_1.31.1 SummarizedExperiment_1.39.1
[17] Biobase_2.69.0 GenomicRanges_1.61.1
[19] Seqinfo_0.99.2 IRanges_2.43.0
[21] S4Vectors_0.47.0 BiocGenerics_0.55.1
[23] generics_0.1.4 MatrixGenerics_1.21.0
[25] matrixStats_1.5.0 BiocStyle_2.37.1
[27] rebook_1.19.0
loaded via a namespace (and not attached):
[1] RColorBrewer_1.1-3 jsonlite_2.0.0
[3] CodeDepends_0.6.6 magrittr_2.0.3
[5] ggbeeswarm_0.7.2 gypsum_1.5.0
[7] farver_2.1.2 rmarkdown_2.29
[9] BiocIO_1.19.0 vctrs_0.6.5
[11] DelayedMatrixStats_1.31.0 memoise_2.0.1
[13] Rsamtools_2.25.2 RCurl_1.98-1.17
[15] htmltools_0.5.8.1 S4Arrays_1.9.1
[17] curl_6.4.0 BiocNeighbors_2.3.1
[19] Rhdf5lib_1.31.0 SparseArray_1.9.1
[21] rhdf5_2.53.4 sass_0.4.10
[23] alabaster.base_1.9.5 bslib_0.9.0
[25] alabaster.sce_1.9.0 httr2_1.2.1
[27] cachem_1.1.0 ResidualMatrix_1.19.0
[29] GenomicAlignments_1.45.2 igraph_2.1.4
[31] lifecycle_1.0.4 pkgconfig_2.0.3
[33] rsvd_1.0.5 Matrix_1.7-3
[35] R6_2.6.1 fastmap_1.2.0
[37] digest_0.6.37 dqrng_0.4.1
[39] irlba_2.3.5.1 ExperimentHub_2.99.5
[41] RSQLite_2.4.2 beachmat_2.25.4
[43] labeling_0.4.3 filelock_1.0.3
[45] httr_1.4.7 abind_1.4-8
[47] compiler_4.5.1 bit64_4.6.0-1
[49] withr_3.0.2 BiocParallel_1.43.4
[51] viridis_0.6.5 DBI_1.2.3
[53] HDF5Array_1.37.0 alabaster.ranges_1.9.1
[55] alabaster.schemas_1.9.0 rappdirs_0.3.3
[57] DelayedArray_0.35.2 bluster_1.19.0
[59] rjson_0.2.23 tools_4.5.1
[61] vipor_0.4.7 beeswarm_0.4.0
[63] glue_1.8.0 h5mread_1.1.1
[65] restfulr_0.0.16 rhdf5filters_1.21.0
[67] grid_4.5.1 Rtsne_0.17
[69] cluster_2.1.8.1 gtable_0.3.6
[71] metapod_1.17.0 BiocSingular_1.25.0
[73] ScaledMatrix_1.17.0 XVector_0.49.0
[75] ggrepel_0.9.6 BiocVersion_3.22.0
[77] pillar_1.11.0 limma_3.65.3
[79] dplyr_1.1.4 lattice_0.22-7
[81] rtracklayer_1.69.1 bit_4.6.0
[83] tidyselect_1.2.1 locfit_1.5-9.12
[85] Biostrings_2.77.2 knitr_1.50
[87] gridExtra_2.3 bookdown_0.43
[89] ProtGenerics_1.41.0 edgeR_4.7.3
[91] xfun_0.52 statmod_1.5.0
[93] UCSC.utils_1.5.0 lazyeval_0.2.2
[95] yaml_2.3.10 evaluate_1.0.4
[97] codetools_0.2-20 tibble_3.3.0
[99] alabaster.matrix_1.9.0 BiocManager_1.30.26
[101] graph_1.87.0 cli_3.6.5
[103] jquerylib_0.1.4 dichromat_2.0-0.1
[105] Rcpp_1.1.0 GenomeInfoDb_1.45.9
[107] dir.expiry_1.17.0 png_0.1-8
[109] XML_3.99-0.18 parallel_4.5.1
[111] blob_1.2.4 sparseMatrixStats_1.21.0
[113] bitops_1.0-9 viridisLite_0.4.2
[115] alabaster.se_1.9.0 scales_1.4.0
[117] purrr_1.1.0 crayon_1.5.3
[119] rlang_1.1.6 cowplot_1.2.0
[121] KEGGREST_1.49.1
References
Muraro, M. J., G. Dharmadhikari, D. Grun, N. Groen, T. Dielen, E. Jansen, L. van Gurp, et al. 2016. “A Single-Cell Transcriptome Atlas of the Human Pancreas.” Cell Syst 3 (4): 385–94.