figure_2.Rmd

---
title: 'Figure 3'
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
knitr::opts_knit$set(root.dir='/local1/USERS/jfleck/data/PUBLIC_ORGANOIDS')
```

This notebook reproduces the main analyses from figure 2 of the manuscript. First we import the necessary packages.

```{r message=FALSE, warning=FALSE, results='hide'}
library(tidyverse)
library(patchwork)
library(Seurat)
library(voxhunt)
```

Now we load the data. The loaded seurat object contains the neuronal popultions of the datasets shown in the manuscript. We further subset the ones shown in figure 2.

```{r}
load_aba_data('voxhunt_data/')
```

We can plot annotations of the mouse brain at different timepoints with the `plot_annotation()` function:

```{r, fig.height=3, fig.width=5}
e11 <- voxhunt::plot_annotation('E11')
e13 <- voxhunt::plot_annotation('E13')
e15 <- voxhunt::plot_annotation('E15')
e18 <- voxhunt::plot_annotation('E18')
p4 <- voxhunt::plot_annotation('P4')
p14 <- voxhunt::plot_annotation('P14')
p28 <- voxhunt::plot_annotation('P28')
p56 <- voxhunt::plot_annotation('P56')

e11 + e13 + e15 + e18 + p4 + p14 + p28 + p56 
```


Since for each voxel, the expression of almost 2000 genes is registered, we can plot some common markers:

```{r, fig.height=4, fig.width=6}
marker_genes <- c('NEUROD6', 'EOMES', 'DCN', 'DLX1', 'DLX2', 
                  'GBX2', 'OTX2', 'GATA3', 'PAX8')
voxhunt::plot_expression('E15', marker_genes, slices=8:12) & no_legend()
```

We can also look for markers of specific brain structures:

```{r, fig.height=4, fig.width=6}
c4_markers <- structure_markers('E15', 'custom_4')
hipp_markers <- c4_markers %>% 
    filter(group=='hippocampus') %>% 
    top_n(8, auc) %>% pull(gene)

cb_markers <- c4_markers %>% 
    filter(group=='cerebellar hemisphere') %>% 
    top_n(8, auc) %>% pull(gene)

p1 <- voxhunt::plot_expression(
    'E15', 
    hipp_markers,
    nrow=2
) & no_legend() 

p2 <- voxhunt::plot_expression(
    'E15', 
    cb_markers,
    nrow=2
) & no_legend()

p1 / p2 
```

As shown in figure 2e), we can do this DE analysis on different levels of annotation at all available developmental stages:

```{r}
c2_markers <- structure_markers('P4', 'custom_2')
top_markers <- c2_markers %>% 
    group_by(group) %>% 
    top_n(1, auc) %>% 
    pull(gene)

voxhunt::plot_expression('P4', top_markers) & no_legend()
```