Loading scRNA-seq Data

Loading scRNA-seq Data

Objective

• Load expression matrix data and make it compatible with Seurat.
  
• Add cell metadata to the Seurat object.

Data Loading

Load Required Libraries

# code:1-1
library(Seurat)
library(tidyverse)  # Includes dplyr, ggplot2, etc.

Load Data Information

Load the path to the expression matrix data and a text file containing metadata.

# code:1-2
# Path to the text file summarizing expression matrix information
data_sheet_path <- "./data_info/datasheet.tsv"

# code:1-3
data_sheet_df <- read_tsv(data_sheet_path)

Check the contents of the data file.
It contains metadata and paths to the datasets.

# code:1-4
data_sheet_df

Extract metadata associated with the expression data as vectors.

# Metadata vectors
# code:1-5
age_vec <- data_sheet_df %>% pull(age)
area_vec <- data_sheet_df %>% pull(area)
condition_vec <- data_sheet_df %>% pull(condition)
sample_vec <- paste0(age_vec, "_", area_vec, "_", condition_vec)

names(age_vec) <- sample_vec
names(area_vec) <- sample_vec
names(condition_vec) <- sample_vec
names(sample_vec) <- sample_vec

# Path vectors for expression matrix data
# code:1-6
path_vec <- data_sheet_df %>% pull(path)
names(path_vec) <- sample_vec

Load Expression Matrices

# Manage multiple expression matrices as a list.
# code:1-7
orig_mtx_list <- list()
for(smn in names(path_vec)){
  print(smn)
  orig_mtx_list[[smn]] <- Read10X_h5(path_vec[smn])
}

## [1] "P1_cortex_WT"
## [1] "P1_cortex_Fezf2KO"

Save the raw expression matrix data as an R object.

# code:1-8
saveRDS(orig_mtx_list, "./output/01_load/obj/orig_mtx_list.RDS")

Create Seurat Object

Convert the loaded expression matrix into a Seurat object.
min.cells = 3: Genes expressed in fewer than 3 cells are removed.

# code:1-9
# Prepare a list to store Seurat objects
created_seurat_obj_list <- list()

# Create Seurat objects
for(snm in names(orig_mtx_list)) {  # Iterate over multiple files
  print(snm)
  created_seurat_obj_list[[snm]] <- CreateSeuratObject(orig_mtx_list[[snm]], min.cells = 3, project = snm)
}

## [1] "P1_cortex_WT"
## [1] "P1_cortex_Fezf2KO"

Check the created Seurat objects.

# code:1-10
created_seurat_obj_list

## $P1_cortex_WT
## An object of class Seurat 
## 16705 features across 6552 samples within 1 assay 
## Active assay: RNA (16705 features, 0 variable features)
##  1 layer present: counts
## 
## $P1_cortex_Fezf2KO
## An object of class Seurat 
## 18611 features across 8386 samples within 1 assay 
## Active assay: RNA (18611 features, 0 variable features)
##  1 layer present: counts

Check the metadata of the Seurat objects.

# code:1-11
head(created_seurat_obj_list[[1]][[]])

head(created_seurat_obj_list[[2]][[]])

Add Metadata to Seurat Object

# code:1-12
for(smn in names(created_seurat_obj_list)){
  print(smn)
  created_seurat_obj_list[[smn]][[]]["age"]        <- age_vec[smn]
  created_seurat_obj_list[[smn]][[]]["area"]        <- area_vec[smn]
  created_seurat_obj_list[[smn]][[]]["condition"]   <- condition_vec[smn]
  created_seurat_obj_list[[smn]][[]]["sample_name"] <- sample_vec[smn]
}

## [1] "P1_cortex_WT"
## [1] "P1_cortex_Fezf2KO"

Verify that metadata has been added.

# code:1-13
head(created_seurat_obj_list[[1]][[]])

head(created_seurat_obj_list[[2]][[]])

Save Seurat Object Before Quality Control

# code:1-14
saveRDS(created_seurat_obj_list, "./output/01_load/obj/created_seurat_obj_list.RDS")