4 TKO chimera processing

setwd(here::here())

library("metacell")
library("tgstat")
library("tgconfig")
library("Matrix")
library("dplyr")

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

scdb_init("scrna_db/", force_reinit = T)

## initializing scdb to scrna_db/

mat <- scdb_mat("tko_chim")

4.0.0.1 Remove cells belonging to E8.5 chimera embryos. Only use chimera embryos from E7.5

e85_embryos <- paste0("TKO29_", c(13:19))
ig_cells <- colnames(mat@mat)[mat@cell_metadata[colnames(mat@mat), "embryo"] %in% e85_embryos]
ig_cells <- c(mat@ignore_cells, ig_cells)

mat <- scm_ignore_cells(scmat = mat, ig_cells = ig_cells)

4.0.0.2 Remove extraembryonic ectoderm and parietal endoderm cells

source("scripts/pipeline/remove_exe_ectoderm_and_parietal_endo_cls.R")
mat <- remove_exe_ectoderm_and_parietal_endoderm(mat_query = mat)
scdb_add_mat(id = "tko_chim", mat = mat)

4.0.0.3 Merge cells with the wildtype atlas

source("scripts/pipeline/merge_umi_mat_with_wt10_umi_mat.R")
mat_new <- merge_umi_mat_with_wt10(scmat = mat)
scdb_add_mat(id = "tko_chim_wt10", mat = mat_new)

4.0.0.4 Generate single-cell balanced kNN-graph

This step needs the file
data/tet_tko.bad_genes.txt

source("scripts/pipeline/gen_cgraph.R")

## initializing scdb to scrna_db/

gen_cgraph(mat_nm = "tko_chim_wt10", force_recompute_cgraph = F)

## [1] TRUE

4.0.0.5 Transfer cell type annotation to chimera cells

source("scripts/pipeline/transfer_cell_type_annotation.R")
transfer_color_chimera_tetraploid(mat_nm = "tko_chim_wt10")

4.0.0.6 Time chimera embryos based on their host and KO cells

source("scripts/pipeline/tko_chim_timing.R")
source("scripts/pipeline/transfer_time_annotation.R")

## Loading required package: slam

## Loading required package: sparsesvd

tko_chim_timing()
# tko_chimera_plot_cor_time_dist_per_emb(mat_nm = "tko_chim_wt10")

4.0.0.7 Export infered time and cell type into table

source("scripts/pipeline/create_metadata.R")
create_tko_chimera_metadata()

4.0.0.8 Generate a separate UMI matrix and metacell object for the TKO cells

source("scripts/pipeline/generat_ko_mat.R")
generate_ko_mat_chimera()

source("scripts/pipeline/generic_mc.R")
tgconfig::override_params(config_file = "config/tet_chim.yaml", package = "metacell")

bad_genes <- read.table("data/tet_tko.bad_genes.txt", sep = "\t", stringsAsFactors = F)
bad_genes <- bad_genes[, 1]

mat_nm <- "tko_chim_ko"

generate_mc(name = mat_nm, add_bad_genes = bad_genes)

# Run the line below if you want to transfer cell type annotation from the WT atlas to the tko_chim_ko metacell object

# wt_atlas <- mcell_gen_atlas(
#     mat_id = "sing_emb_wt10",
#     mc_id = "sing_emb_wt10_recolored",
#     gset_id = "sing_emb_wt10",
#     mc2d_id = "sing_emb_wt10_recolored"
# )

# cmp <- mcell_proj_on_atlas(
#     mat_id = "tko_chim_ko",
#     mc_id = "tko_chim_ko",
#     atlas = wt_atlas,
#     fig_cmp_dir = paste0("figs/atlas_projection_", "tko_chim_ko"), recolor_mc_id = "tko_chim_ko", ten2mars = F
# )

tgconfig::set_param(param = "mcell_mc2d_T_edge", value = 0.03, package = "metacell")
mcell_mc2d_force_knn(mc2d_id = "tko_chim_ko", mc_id = "tko_chim_ko", graph_id = "tko_chim_ko", graph_parametric = T, feats_gset = "tko_chim_ko")

## got 74 feat genes for mc graph construction

mcell_mc2d_plot("tko_chim_ko")

## png 
##   2