3_epithelium_monocle.rmd

---
title: "scSEQ analysis of the developing epithelium"
author: Nick Negretti
date: 08/09/21
output: rmarkdown::github_document
---

# Analysis of the lung epithelium

## Load libraries and helper functions
```{r, results="hide", message = FALSE, load-libraries}
setwd("~/postdoc/code/devo_scseq_github")
renv::activate()
source("./helper_functions/globals.R")
#source("./helper_functions/libraries.R")
library(monocle3)
library(tidyverse)
library(knitr) # for kable
options(knitr.table.format = "html")
library(kableExtra) # for pretty tables kable_styling()

opts_knit$set(root.dir = getwd())

source("./helper_functions/trajectory.R")
source("./helper_functions/cluster.R")
source("./helper_functions/colors.R")
source("./helper_functions/brackets.R")
source("./helper_functions/heatmaps.R")

#N_WORKERS <- 12
#plan("multiprocess", workers = N_WORKERS)
epi_cds <- readRDS("./data/20210802_celldataset_epi_relabel.rds")
epi_cds



```
## Load data from merge and clean pipeline
```{r}
epi_cds <- readRDS("./data/20210802_celldataset_epi_relabel.rds")
epi_cds



at1_genes <- c("Aqp5",
                    "Sftpc",
                    "Hopx",
                    "Ager",
                    "Col4a3")
cds_subset <- epi_cds[rownames(epi_cds) %in% at1_genes,]
cds_subset@colData$timepoint_numeric <- as.numeric(cds_subset@colData$timepoint_numeric)


gene_fits_test <- fit_models(cds_subset, model_formula_str = "~timepoint_numeric", cores = 20)
fit_coefs_test <- coefficient_table(gene_fits_test)
filt_fits_test <- fit_coefs_test %>% filter(term == "timepoint_numeric") %>% select(gene_id, term, q_value, estimate)


##
gene_fits_all <- fit_models(epi_cds, model_formula_str = "~timepoint_numeric", cores = 20)
fit_coefs_all <- coefficient_table(gene_fits_all)
filt_fits_all <- fit_coefs_all %>% filter(term == "timepoint_numeric") %>% select(gene_id, term, q_value, estimate)
saveRDS(filt_fits_all, "./data/20210802_epi_filt_fits_all.rds")
gc()


##
gene_fits_at1 <- fit_models(epi_cds[,epi_cds@colData$celltype == "AT1"], model_formula_str = "~timepoint_numeric", cores = 20)
fit_coefs_at1 <- coefficient_table(gene_fits_at1)
filt_fits_at1 <- fit_coefs_at1 %>% filter(term == "timepoint_numeric") %>% select(gene_id, term, q_value, estimate)
saveRDS(filt_fits_at1, "./data/20210802_epi_filt_fits_at1.rds")
gc()

##
gene_fits_at2 <- fit_models(epi_cds[,epi_cds@colData$celltype == "AT2"], model_formula_str = "~timepoint_numeric", cores = 20)
fit_coefs_at2 <- coefficient_table(gene_fits_at2)
filt_fits_at2 <- fit_coefs_at2 %>% filter(term == "timepoint_numeric") %>% select(gene_id, term, q_value, estimate)
saveRDS(filt_fits_at2, "./data/20210802_epi_filt_fits_at2.rds")
gc()

##
gene_fits_ciliated <- fit_models(epi_cds[,epi_cds@colData$celltype == "Ciliated"], model_formula_str = "~timepoint_numeric", cores = 20)
fit_coefs_ciliated <- coefficient_table(gene_fits_ciliated)
filt_fits_ciliated <- fit_coefs_ciliated %>% filter(term == "timepoint_numeric") %>% select(gene_id, term, q_value, estimate)
saveRDS(filt_fits_ciliated, "./data/20210802_epi_filt_fits_ciliated.rds")
gc()


gene_fits_secretory <- fit_models(epi_cds[,epi_cds@colData$celltype == "Secretory"], model_formula_str = "~timepoint_numeric", cores = 20)
fit_coefs_secretory <- coefficient_table(gene_fits_secretory)
filt_fits_secretory <- fit_coefs_secretory %>% filter(term == "timepoint_numeric") %>% select(gene_id, term, q_value, estimate)
saveRDS(filt_fits_secretory, "./data/20210802_epi_filt_fits_secretory.rds")
gc()


```