Main
After pre-processing, we create a SpaTalk object with st_data and st_meta for spot-based ST data. In practise, we generate simulated spot-based ST from single-cell STARmap data using generate_spot(). For Slide-seq (10um) close to single-cell resolution, we recommend spot_max_cell as 1. For 10X Visum (55um), we recommend spot_max_cell as 30 as stated in the recent review.
library(SpaTalk)# load starmap data
load(paste0(system.file(package = 'SpaTalk'), "/extdata/starmap_data.rda"))
load(paste0(system.file(package = 'SpaTalk'), "/extdata/starmap_meta.rda"))
# generate spot-based ST from single-cell STARmap data
st_data <- generate_spot(st_data = as.matrix(starmap_data),
st_meta = starmap_meta,
x_min = 0,
x_res = 200,
x_max = 3800,
y_min = 0,
y_res = 500,
y_max = 17500)## Using celltype as value column: use value.var to override.## Aggregation function missing: defaulting to lengthst_meta <- st_data$st_meta
st_data <- st_data$st_data
# create SpaTalk data
obj <- createSpaTalk(st_data = as.matrix(st_data),
st_meta = st_meta[, c(1,2,3)],
species = "Mouse",
if_st_is_sc = F,
spot_max_cell = max(st_meta$cell_real))Use plot_st_pie_generate() to view the cell-type composition
plot_st_pie_generate(st_meta = st_meta, pie_scale = 1.3, xy_ratio = 1.8)
Cell-type decomposition
First, we apply NNLM and spatial mapping to reconstruct the ST data at single-cell resolution using dec_celltype()
# decode the cell-type composition
obj <- dec_celltype(object = obj,
sc_data = as.matrix(starmap_data),
sc_celltype = starmap_meta$celltype)## Performing Non-negative regression for each spot
## Generating single-cell data for each spot
## ***Done***Use plot_st_pie() to view predicted cell-type composition
# Scatter pie plot for each spot
plot_st_pie(object = obj, pie_scale = 1.3, xy_ratio = 1.8)
Use plot_st_celltype_percent() to view cell-type percent
# plot cell-type percent across spatial spots
plot_st_celltype_percent(object = obj, celltype = 'Oligo',size = 4)
Use plot_st_gene() to view gene expression
# plot marker gene expression across spatial spots
plot_st_gene(object = obj, gene = 'Plp1',size = 4, if_use_newmeta = F)
Use plot_st_cor_heatmap() to view correlation heatmap
# correlation between marker gene expression and cell type percent across spatial spots
plot_st_cor_heatmap(object = obj,
marker_genes = c("Plp1","Vip","Sst","Lamp5","Pcp4","Mfge8","Pvalb"),
celltypes = c("Oligo","VIP","SST","eL2_3","eL6","Astro","PVALB"),
scale = "none",
color_low = 'blue',
color_high = 'yellow',
color_mid = 'yellow')
ST at single-cell resolution
Use plot_st_celltype() to view cell type in space
# plot cell type in reconstructed ST atlas
plot_st_celltype(object = obj, celltype = 'Oligo', size = 2)
Use plot_st_gene() to view gene expression in space
# plot marker gene expression in reconstructed ST atlas
plot_st_gene(object = obj,gene = 'Plp1', if_use_newmeta = T, size = 2)
Use plot_st_celltype_density() to view cell-type density in space
# plot cell-type density in reconstructed ST atlas
plot_st_celltype_density(object = obj,
celltype = 'Oligo',
type = 'raster',
color_low = 'purple',
color_high = 'yellow')
plot_st_celltype_density(object = obj,
celltype = 'Oligo',
type = 'contour',
color_low = 'purple',
color_high = 'yellow')
Use plot_st_celltype_all() to view all cell types in space
# plot all cell types in reconstructed ST atlas
plot_st_celltype_all(object = obj, size = 2)
Infer cell-cell communications
Based on the constructed ST atlas at single-cell resolution, we use find_lr_path() and dec_cci() to infer cell-cell communications mediated by ligand-receptor interactions (LRIs) in space using LRIs from CellTalkDB, pathways from KEGG and Reactome, and transcriptional factors (TFs) from AnimalTFDB
# Filter LRIs with downstream targets
obj <- find_lr_path(object = obj, lrpairs = lrpairs, pathways = pathways)## Checking input data
## Begin to filter lrpairs and pathways
## ***Done***# Infer cell-cell communications from SST to PVALB neurons
obj <- dec_cci(object = obj,
celltype_sender = 'SST',
celltype_receiver = 'PVALB')## Begin to find LR pairs# Get LR and downstream pathways
obj_lr_path <- get_lr_path(object = obj,
celltype_sender = 'SST',
celltype_receiver = 'PVALB',
ligand = 'Sst',
receptor = 'Sstr2')
obj_lr_path$tf_path## src dest src_tf dest_tf hop co_ratio tf
## 909510 Sstr2 Gnai1 NO NO 1 0.3076923 Smad3
## 268551 Gnai1 Mapk3 NO NO 2 0.1538462 Smad3
## 300531 Mapk3 Smad3 NO YES 3 0.1538462 Smad3
## 327207 Smad3 Serpine1 YES NO 4 0.2307692 Smad3obj_lr_path$path_pvalue## celltype_sender celltype_receiver ligand receptor
## 1 SST PVALB Sst Sstr2
## 2 SST PVALB Sst Sstr2
## 3 SST PVALB Sst Sstr2
## receptor_pathways pvalue gene_count
## 1 cAMP signaling pathway 1.009418e-05 3
## 2 Neuroactive ligand-receptor interaction 6.039598e-03 2
## 3 Growth hormone synthesis, secretion and action 1.281881e-08 4
## gene
## 1 Sst,Sstr2,Gnai1
## 2 Sst,Sstr2
## 3 Sst,Sstr2,Gnai1,Mapk3Use plot_ccdist() to view distribution of senders and receivers
# Point plot with spatial distribution of celltype_sender and celltype_receiver
plot_ccdist(object = obj,
celltype_sender = 'SST',
celltype_receiver = 'PVALB',
size = 2,
arrow_length = 0.1)
Use plot_cci_lrpairs() to view all LRIs of senders and receivers
Given the limited LR pairs, we do not show the result here
# Heatmap with LR pairs of celltype_sender and celltype_receiver
# plot_cci_lrpairs(object = obj, celltype_sender = 'SST', celltype_receiver = 'PVALB')Use plot_lrpair() to view the specific LRI
# Point plot with LR pair from celltype_sender to celltype_receiver
plot_lrpair(object = obj,
celltype_sender = 'SST',
ligand = 'Sst',
celltype_receiver = 'PVALB',
receptor = 'Sstr2',
if_plot_density = F,
size = 2,
arrow_length = 0.1)
Use plot_lrpair_vln() to view violin plot with spatial distance of LRI
# Violin plot with spatial distance of LR pair between senders and receivers and between all cell-cell pairs
plot_lrpair_vln(object = obj,
celltype_sender = 'SST',
ligand = 'Sst',
celltype_receiver = 'PVALB',
receptor = 'Sstr2')##
## Wilcoxon rank sum test with continuity correction
##
## data: celltype_dist1 and celltype_dist2
## W = 5685812, p-value = 5.395e-13
## alternative hypothesis: true location shift is greater than 0
Use plot_lr_path() to view network with LR and downstream pathways
# Plot network with LR and downstream pathways
plot_lr_path(object = obj,
celltype_sender = 'SST',
ligand = 'Sst',
celltype_receiver = 'PVALB',
receptor = 'Sstr2')
Use plot_path2gene() to view river plot of significantly activated pathways
# River plot of significantly activated pathways and related downstream genes of receptors
plot_path2gene(object = obj,
celltype_sender = 'SST',
ligand = 'Sst',
celltype_receiver = 'PVALB',
receptor = 'Sstr2')## Warning: The `.dots` argument of `group_by()` is deprecated as of dplyr 1.0.0.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was generated.
Note
To infer all paired cell-cell communications, use dec_cci_all() instead of dec_cci()
# Infer all cell-cell communications
# obj <- dec_cci_all(object = obj)sessionInfo()## R version 4.1.1 (2021-08-10)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 10 x64 (build 19044)
##
## Matrix products: default
##
## locale:
## [1] LC_COLLATE=Chinese (Simplified)_China.936
## [2] LC_CTYPE=Chinese (Simplified)_China.936
## [3] LC_MONETARY=Chinese (Simplified)_China.936
## [4] LC_NUMERIC=C
## [5] LC_TIME=Chinese (Simplified)_China.936
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] SpaTalk_1.0 ggalluvial_0.12.3 ggplot2_3.3.5
##
## loaded via a namespace (and not attached):
## [1] Seurat_4.0.5 Rtsne_0.15 colorspace_2.0-2
## [4] ggsignif_0.6.3 deldir_1.0-6 ellipsis_0.3.2
## [7] ggridges_0.5.3 spatstat.data_2.1-0 ggpubr_0.4.0
## [10] leiden_0.3.9 listenv_0.8.0 farver_2.1.0
## [13] ggrepel_0.9.1 fansi_0.5.0 scatterpie_0.1.7
## [16] codetools_0.2-18 splines_4.1.1 knitr_1.36
## [19] polyclip_1.10-0 jsonlite_1.7.2 broom_0.7.10
## [22] ica_1.0-2 cluster_2.1.2 png_0.1-7
## [25] uwot_0.1.10 pheatmap_1.0.12 spatstat.sparse_2.0-0
## [28] ggforce_0.3.3 shiny_1.7.1 sctransform_0.3.2
## [31] compiler_4.1.1 httr_1.4.2 backports_1.3.0
## [34] assertthat_0.2.1 SeuratObject_4.0.2 Matrix_1.3-4
## [37] fastmap_1.1.0 lazyeval_0.2.2 later_1.3.0
## [40] tweenr_1.0.2 htmltools_0.5.2 prettyunits_1.1.1
## [43] tools_4.1.1 igraph_1.2.7 gtable_0.3.0
## [46] glue_1.4.2 RANN_2.6.1 reshape2_1.4.4
## [49] dplyr_1.0.7 Rcpp_1.0.7 carData_3.0-4
## [52] scattermore_0.7 jquerylib_0.1.4 vctrs_0.3.8
## [55] nlme_3.1-152 lmtest_0.9-38 xfun_0.30
## [58] stringr_1.4.0 globals_0.14.0 mime_0.12
## [61] miniUI_0.1.1.1 lifecycle_1.0.1 irlba_2.3.3
## [64] rstatix_0.7.0 goftest_1.2-3 NNLM_0.4.4
## [67] future_1.23.0 MASS_7.3-54 zoo_1.8-9
## [70] scales_1.1.1 spatstat.core_2.3-0 spatstat.utils_2.2-0
## [73] hms_1.1.1 promises_1.2.0.1 parallel_4.1.1
## [76] RColorBrewer_1.1-2 yaml_2.2.1 gridExtra_2.3
## [79] reticulate_1.22 pbapply_1.5-0 ggfun_0.0.4
## [82] sass_0.4.0 rpart_4.1-15 ggExtra_0.9
## [85] stringi_1.7.5 highr_0.9 rlang_0.4.12
## [88] pkgconfig_2.0.3 matrixStats_0.61.0 evaluate_0.14
## [91] lattice_0.20-44 tensor_1.5 ROCR_1.0-11
## [94] purrr_0.3.4 labeling_0.4.2 patchwork_1.1.1
## [97] htmlwidgets_1.5.4 cowplot_1.1.1 tidyselect_1.1.1
## [100] ggsci_2.9 parallelly_1.28.1 RcppAnnoy_0.0.19
## [103] plyr_1.8.6 magrittr_2.0.1 R6_2.5.1
## [106] generics_0.1.1 DBI_1.1.1 mgcv_1.8-36
## [109] pillar_1.6.4 withr_2.4.2 fitdistrplus_1.1-6
## [112] prettydoc_0.4.1 abind_1.4-5 survival_3.2-11
## [115] tibble_3.1.5 future.apply_1.8.1 car_3.0-12
## [118] crayon_1.4.2 KernSmooth_2.23-20 utf8_1.2.2
## [121] spatstat.geom_2.3-0 plotly_4.10.0 rmarkdown_2.13
## [124] progress_1.2.2 isoband_0.2.5 grid_4.1.1
## [127] data.table_1.14.2 digest_0.6.28 xtable_1.8-4
## [130] tidyr_1.1.4 httpuv_1.6.3 munsell_0.5.0
## [133] viridisLite_0.4.0 bslib_0.3.1