13c_cluster_downstream.Rmd

---
title: "What predicts the number of downstream genes of a target?"
author: "Britta Velten"
date:  "`r format(Sys.Date(),'%e %B, %Y')`"
output: BiocStyle::html_document
params:
  utils_path: "../../io/Magpie_Utils.R"
  io_path: "../../io/Magpie_io.R"
  date: "2022-08-15"
  section_name: "10a_number_of_DEGs"
---

# Preparations
```{r prep, warning=FALSE, message=FALSE, echo = F}
# to render pngs in html
library(Cairo)
knitr::opts_chunk$set(fig.path = plotsdir, dev=c("CairoPNG", "pdf"))
```

# Generate Heatmap

## Downstream Genes with Signal

```{r, fig.width = 20, fig.height = 4, echo =F}
#targets2consider <- unlist(pluripotency_tfs)
downstreams2consider <- pluripotency_lfcs %>%
  dplyr::filter(pval_adj < 0.1 & 
                  abs(lfc) > 0.2) %>%
  .$downstream_gene_name %>% as.character() %>%
  table() %>% as.data.frame() %>%
  dplyr::filter(Freq >=1) %>%
  .$`.` %>% as.character()
#downstreams2consider <- sort(unique(downstreams2consider, unlist(pluripotency_tfs)))

df4heatmap <- pluripotency_lfcs %>%
  dplyr::filter(downstream_gene_name %in% downstreams2consider) %>%
  dplyr::select(c('genes_in_complex', 'downstream_gene_name', 'lfc', 'pval_adj')) %>%
  #mutate(lfc = ifelse(pval_adj < 0.1, lfc, 0)) %>%
  reshape2::dcast(genes_in_complex ~ downstream_gene_name, value.var = 'lfc') %>%
  column_to_rownames('genes_in_complex') %>%
  as.matrix()

my_heatmap(df4heatmap,
           #treeheight_row=0, treeheight_col = 0,
           min_c = -0.5, max_c = 0.5,
           border_col = NA)
```

## Cluster Downstream Genes

```{r, echo = F}
fviz_nbclust(t(df4heatmap), kmeans, method = "silhouette")+
  labs(subtitle = "Silhouette method")
cluster_res <- hclust(dist(t(df4heatmap)))
downstream_clust <- data.frame(
  gene = colnames(df4heatmap),
  cluster = factor(as.numeric(cutree(cluster_res, k = 2)))
)
```


## Functional Enrichment of Clusters
### Cluster 1

These are the genes that are down-regulated upon OCT4 knockdown and up-regulated due to knockdown of MAX.

Genes 



```{r}
cluster_1 <- downstream_clust %>%
  dplyr::filter(cluster == 1) %>% .$gene
#cluster_1_enrichment <- gprofiler2::gost(cluster_1)$result
```

Genes found to be uniquely expressed in iPSCs: https://www.nature.com/articles/s41598-020-66845-6. 

```{r}
ipsc_marker_genes <- c('MT1G', 'USP44', 'ESRG', 'SPP1', 'PTPRZ1', 'CRABP1', 'CNMD', 'THY1', 'VSNL1', 'HHLA1', 'SFRP2')
length(which(ipsc_marker_genes %in% cluster_1))

other_genes <- cluster_1[which(!(cluster_1 %in% ipsc_marker_genes))]
represented_pathways <- lapply(msigdb_hallmark_pathways, FUN = function(l){
  length(which(l %in% cluster_1))
}) %>% unlist()
represented_pathways <- represented_pathways[represented_pathways > 0]
```

### Cluster 2
# SessionInfo
```{r}
sessionInfo()
```