In Yutong441/TBdev: Infer a Trophoblast Developmental Trajectory

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  warning = FALSE
)

library (TBdev)

Load data

The single cell RNA-seq dataset is located in data_dir. The output of this analysis will be stored in save_dir and sup_save_dir

root <- '.'
data_dir <- paste (root, 'data', sep='/')
all_data <- get (load (paste (data_dir, 'final_merged_tb.Robj', sep='/') ))

Correlation

For example, we will quantify the correlation between the transcriptome of in vitro cells i.e. human embryonic stem cells (hESC) and human trophoblast stem cells (hTSC) with the in vivo cell lines.

select_cells2 <- all_data$date == 'in_vitro'
select_cells1 <- !select_cells2

# you may choose to run 'partial_corr' or 'distance' to quantify partial
# correlation and Euclidean distance respectively
all_cor <- compute_all_cor (all_data, 
                            method= 'correlation', 
                            assay='RNA',
                            select_cells1=select_cells1,
                            select_cells2=select_cells2
)

Show the results in heatmap.

AP <- list (pointsize=3, legend_point_size=3, fontsize=11, point_fontsize=4,
            font_fam= 'sans')
cell_heat (all_cor, [e-mail], 
           features=c('broad_type', 'broad_type'), #group by features
           # first one for select_cells1 (rows), second one for
           # select_cells2 (columns)
           # below are the parameters to `seurat_heat`
           row_legend_labels='comparison',
           center_scale=T,
           column_scale=T, 
           row_scale=F,
           grid_height=4, AP = AP,
           column_rotation = 90,
           main_height = 25, main_width=7,
           automatic =F #make sure all legends lie in the same column
)

in violin plot

cell_violin (all_cor, [e-mail], c('broad_type', 'broad_type'), 
             column_scale=T, num_col=2, AP=AP)

Quadratic programming

Use quadratic programming to determine cell-cell similarity

library (DeconRNASeq)
select_types <- c('hTSC_OKAE', 'hTSC_TURCO', 'hESC', 'hESC_YAN')
compare_types <- c('ICM', 'TB', 'CTB', 'STB', 'EVT')
cell_sim <- get_cell_similarity (all_data, 
                                 group.by=c('broad_type', 'broad_type'), 
                                 # in the order of select_types, compare_types
                                 select_types= select_types, 
                                 compare_types=compare_types
)

plot_cell_similarity (all_data, cell_sim, group.by= 'broad_type', DR='pca',
                      AP=AP)

Yutong441/TBdev documentation built on Dec. 18, 2021, 8:22 p.m.