part2_RNAvelocity_Seurat.Rmd

---
title: "RNA Velocity analysis using velocyto, Seurat, and scVelo (Part 2: Seurat data)""
subtitle:  "Extract data from Seurat to merge with Loom file for scVelo RNA velocity"
author: "Katherine Beigel (beigelk@chop.edu)"
---

# Saving Seurat object as h5ad

This script does not include any Seurat processing steps. The "WT_Seurat.rds" input file is expected to be a fully processed Seurat object (with UMAP, clustering, etc.).

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
# main
library(tidyverse)
library(Seurat)
library(SeuratDisk)
library(viridis)
library(ggplot2)
library(scales)
library(mgsub)
library(cowplot)

```

## Set the directory for where to read/write data files

```{r Directories}

datadir <- "/dir/where/to/load/SeuratObj/from/"
outdir <- "/dir/where/to/write/outputfiles/to/"

```

## Load the RDS file

```{r Load rds Seurat file}

seu <- readRDS(file = paste0(datadir, "WT_Seurat.rds"))

# Optional: if you need to subset the object to only look at some cells
# Idents(object = seu) <- "orig.ident"
# seu <- subset(seu, idents = "WT_sub")

# Check to see how the UMAP plot looks
DimPlot(seu, group.by = "cell_type", label = TRUE)

```

## Save the Seurat object as a h5Seurat, convert to .h5ad
Vignette: https://mojaveazure.github.io/seurat-disk/articles/convert-anndata.html

```{r Convert Seurat to .h5ad, eval = FALSE}

# Uses pkg SeuratDisk
SaveH5Seurat(seu, filename = paste0(outdir, "WT_seurat.h5Seurat"))
Convert(paste0(outdir, "WT_seurat.h5Seurat"), dest = "h5ad") # this file will be used in Part 3: scVelo RNA velocity

```