notebook_merge.Rmd

---
title: "Module - Merging and Normalizing"
author: "BTC Pipeline"
date: "`r Sys.time()`"
output:
  html_document:
    toc: TRUE
    toc_float: TRUE
    toc_depth: 4
    code_folding: show
    theme: united
    df_print: paged
    self_contained: TRUE
params:
  project_name: 'Test'
  input_qc_approved: './dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_ASCITES_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-065_S1_CD45P_ASCITES_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-065_S1_CD45N_INFRACOLIC_OMENTUM_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-065_S1_CD45P_RIGHT_OVARY_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-065_S1_CD45P_INFRACOLIC_OMENTUM_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-022_S1_CD45P_BOWEL_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-022_S1_CD45N_RIGHT_ADNEXA_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45N_BOWEL_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_BOWEL_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-022_S1_CD45P_RIGHT_ADNEXA_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_INFRACOLIC_OMENTUM_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_PELVIC_PERITONEUM_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45N_RIGHT_OVARY_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_LEFT_OVARY_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45N_LEFT_OVARY_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_LEFT_UPPER_QUADRANT_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45N_RIGHT_UPPER_QUADRANT_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_RIGHT_UPPER_QUADRANT_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45P_RIGHT_OVARY_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-009_S1_CD45N_INFRACOLIC_OMENTUM_seurat_object.RDS;./dev/data/normalize/SPECTRUM-OV-065_S1_CD45N_RIGHT_FALLOPIAN_TUBE_seurat_object.RDS'
  thr_n_features: 2000
  n_threads: 8
  n_memory: 16
  workdir: !r here::here()
  timestamp: !r Sys.Date()
  auto_save: TRUE
---

------------------------------------------------------------------------

# Project Name: `r params$project_name`

```{r setup, include = FALSE, message = FALSE, warning = FALSE, echo = TRUE}

# Project parameters 
project_name <- params$project_name

# Inputs and thresholds
input_qc_approved <- strsplit(params$input_qc_approved, split = ';')[[1]]

thr_n_features <- params$thr_n_features

# Optional parameters

# Dataflow/Computational parameters
n_threads <- params$n_threads
n_memory <- params$n_memory

# Output parameters
work_directory <- params$workdir
timestamp <- params$timestamp
auto_save <- params$auto_save

# Rmarkdown options
knitr::opts_chunk$set(
  echo = TRUE,
  error = TRUE,
  message = FALSE, 
  warning = FALSE,
  fig.height = 8, 
  fig.width = 12,
  fig.align = 'center',
  dpi = 300
  )

knitr::opts_knit$set(
  root.dir = work_directory
  )


```

## Loading library

```{r library}

library(readr)
library(dplyr)
library(ggplot2)
library(Seurat)

```

## Folder structure

```{r project_directory}

if(!dir.exists(work_directory)) {
  dir.create(work_directory, recursive = T)
}

for(sub_directory in c('data', 'figures', 'figures/merge')) {
  dir.create(paste0(work_directory, '/', sub_directory))
}

# Cirro-related edition
file.create(c('figures/merge/.empty', '.dummy'))

```

## Merging and Normalizing

```{r merge, message = FALSE, warning = FALSE, echo = TRUE}

# Adjusting names
sample_label <- sub(
  '(\\S+)_seurat_object.RDS', '\\1', basename(input_qc_approved))

sample_label <- setNames(
  input_qc_approved,
  sample_label
)

# Reading approved samples
seurat_object_list <- list()
for (label in names(sample_label)) {
  seurat_object_list[[label]] <- readRDS(file = sample_label[[label]])
  seurat_object_list[[label]] <- RenameCells(
    seurat_object_list[[label]], add.cell.id = label)

}

# Merge all the objects in the list
seurat_object <- purrr::reduce(
  seurat_object_list, project = project_name, merge, do.normalize = FALSE)

```

```{r normalization, message = FALSE, warning = FALSE, echo = TRUE}

seurat_object <- NormalizeData(
  object = seurat_object, 
  normalization.method = "LogNormalize",
  scale.factor = 10000)

seurat_object <- FindVariableFeatures(
  seurat_object, selection.method = "vst", nfeatures = thr_n_features)

seurat_object <- ScaleData(seurat_object)

```

## Dimensionality Reduction

```{r dimensionality_reduction}

# Running PCA and printing out dimensions
seurat_object <- RunPCA(seurat_object, verbose = FALSE, npcs = 100)
print(seurat_object[["pca"]], dims = 1:5, nfeatures = 5)

```

### PCA Loadings

```{r pca_loading_plot, fig.height = 14}

VizDimLoadings(seurat_object, dims = 1:5, reduction = "pca")

```

### Features Plot

```{r variable_plot}

# Identify the 10 most highly variable genes
variable_top10 <- head(VariableFeatures(seurat_object), 10)

# plot variable features with and without labels
variable_plot <- VariableFeaturePlot(seurat_object)
variable_plot <- LabelPoints(
  plot = variable_plot, points = variable_top10, repel = TRUE)

variable_plot

```

### Elbow Plot

```{r elbow_plot}

elbow_plot <- ElbowPlot(object = seurat_object, ndims = 100)

# Saving Elbow Plot
if(auto_save) {
  
  ggsave(plot = elbow_plot, filename = 'figures/merge/Elbow_plots.pdf')
  
}

elbow_plot

```

## Saving Seurat object

```{r object_dump}

if(auto_save) {
  
  saveRDS(seurat_object, file = paste0('data/', project_name, '_merged_object.RDS'))
  
}

```

------------------------------------------------------------------------

## Parameters log

```{r params_log, message = FALSE, warning = FALSE, echo = FALSE}

print(params)

```

## Session info

```{r session_info, message = FALSE, warning = FALSE, echo = FALSE}

sessionInfo()

```