---
title: "Pathway activity inference in bulk RNA-seq"
author:
  - name: Pau Badia-i-Mompel
    affiliation:
    - Heidelberg Universiy
output: 
  BiocStyle::html_document:
    self_contained: true
    toc: true
    toc_float: true
    toc_depth: 3
    code_folding: show
package: "`r pkg_ver('decoupleR')`"
vignette: >
  %\VignetteIndexEntry{Pathway activity inference in bulk RNA-seq}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(
    collapse = TRUE,
    comment = "#>"
)
```
 
Bulk RNA-seq yield many molecular readouts that are hard to interpret by
themselves. One way of summarizing this information is by inferring pathway
activities from prior knowledge.

In this notebook we showcase how to use `decoupleR` for pathway activity
inference with a bulk RNA-seq data-set where the transcription factor FOXA2 was
knocked out in pancreatic cancer cell lines. 

The data consists of 3 Wild Type (WT) samples and 3 Knock Outs (KO). They are
freely available in 
[GEO](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE119931).

# Loading packages
First, we need to load the relevant packages:

```{r "load packages", message = FALSE}
## We load the required packages
library(decoupleR)
library(dplyr)
library(tibble)
library(tidyr)
library(ggplot2)
library(pheatmap)
library(ggrepel)
```

# Loading the data-set
Here we used an already processed bulk RNA-seq data-set. We provide the
normalized log-transformed counts, the experimental design meta-data and the
Differential Expressed Genes (DEGs) obtained using `limma`.
We can open the data like this:
```{r "load data"}
inputs_dir <- system.file("extdata", package = "decoupleR")
data <- readRDS(file.path(inputs_dir, "bk_data.rds"))
```

From `data` we can extract the mentioned information. Here we see the normalized
log-transformed counts:
```{r "counts"}
# Remove NAs and set row names
counts <- data$counts %>%
  dplyr::mutate_if(~ any(is.na(.x)), ~ if_else(is.na(.x),0,.x)) %>% 
  column_to_rownames(var = "gene") %>% 
  as.matrix()
head(counts)
```

The design meta-data:
```{r "design"}
design <- data$design
design
```

And the results of `limma`, of which we are interested in extracting the
obtained t-value from the contrast:
```{r "deg"}
# Extract t-values per gene
deg <- data$limma_ttop %>%
    select(ID, t) %>% 
    filter(!is.na(t)) %>% 
    column_to_rownames(var = "ID") %>%
    as.matrix()
head(deg)
```

# PROGENy model
[PROGENy](https://saezlab.github.io/progeny/) is a comprehensive resource
containing a curated collection of pathways and their target genes, with weights
for each interaction. For this example we will use the human weights
(mouse is also available) and we will use the top 100 responsive genes ranked
by p-value. We can use `decoupleR` to retrieve it from `OmniPath`:

```{r "progeny", message=FALSE}
net <- get_progeny(organism = 'human', top = 100)
net
```

# Activity inference with Weighted Mean
To infer activities we will run the Weighted Mean method (`wmean`). It infers
regulator activities by first multiplying each target feature by its associated
weight which then are summed to an enrichment score `wmean`. Furthermore,
permutations of random target features can be performed to obtain a null
distribution that can be used to compute a z-score `norm_wmean`, or a corrected
estimate `corr_wmean` by multiplying `wmean` by the minus log10 of the obtained
empirical p-value.

In this example we use `wmean` but we could have used any other.
To see what methods are available use `show_methods()`.

To run `decoupleR` methods, we need an input matrix (`mat`), an input prior
knowledge network/resource (`net`), and the name of the columns of net that we
want to use.

```{r "sample_wmean", message=FALSE}
# Run wmean
sample_acts <- run_wmean(mat=counts, net=net, .source='source', .target='target',
                  .mor='weight', times = 100, minsize = 5)
sample_acts
```

# Visualization

From the obtained results, we will select the `norm_wmean` activities and we 
will observe the obtained activities per sample in a heat-map:
```{r "heatmap"}
# Transform to wide matrix
sample_acts_mat <- sample_acts %>%
  filter(statistic == 'norm_wmean') %>%
  pivot_wider(id_cols = 'condition', names_from = 'source',
              values_from = 'score') %>%
  column_to_rownames('condition') %>%
  as.matrix()

# Scale per sample
sample_acts_mat <- scale(sample_acts_mat)

# Choose color palette
palette_length = 100
my_color = colorRampPalette(c("Darkblue", "white","red"))(palette_length)

my_breaks <- c(seq(-3, 0, length.out=ceiling(palette_length/2) + 1),
               seq(0.05, 3, length.out=floor(palette_length/2)))

# Plot
pheatmap(sample_acts_mat, border_color = NA, color=my_color, breaks = my_breaks) 
```

We can observe that WT samples have higher activities for p53 and TGFb than KO.
On the other hand, KO show higher activities for MAPK and PI3K.

We can also infer pathway activities from the t-values of the DEGs between KO
and WT:
```{r "contrast_wmean", message=FALSE}
# Run wmean
contrast_acts <- run_wmean(mat=deg, net=net, .source='source', .target='target',
                  .mor='weight', times = 100, minsize = 5)
contrast_acts
```

We select the `norm_wmean` activities and then we show the changes
in activity between KO and WT:

```{r "barplot"}
# Filter norm_wmean
f_contrast_acts <- contrast_acts %>%
  filter(statistic == 'norm_wmean')

# Plot
ggplot(f_contrast_acts, aes(x = reorder(source, score), y = score)) + 
    geom_bar(aes(fill = score), stat = "identity") +
    scale_fill_gradient2(low = "darkblue", high = "indianred", 
        mid = "whitesmoke", midpoint = 0) + 
    theme_minimal() +
    theme(axis.title = element_text(face = "bold", size = 12),
        axis.text.x = 
            element_text(angle = 45, hjust = 1, size =10, face= "bold"),
        axis.text.y = element_text(size =10, face= "bold"),
        panel.grid.major = element_blank(), 
        panel.grid.minor = element_blank()) +
    xlab("Pathways")
```

As observed before, the pathways p53 and TGFb are deactivated in KO when
compared to WT, while JAK-STAT and others seem to be activated.

We can further visualize the most responsive genes in each pathway along their
t-values to interpret the results. For example, let's see the genes that are 
belong to the MAPK pathway:
```{r "targets"}
pathway <- 'MAPK'

df <- net %>%
  filter(source == pathway) %>%
  arrange(target) %>%
  mutate(ID = target, color = "3") %>%
  column_to_rownames('target')
inter <- sort(intersect(rownames(deg),rownames(df)))
df <- df[inter, ]
df['t_value'] <- deg[inter, ]
df <- df %>%
  mutate(color = if_else(weight > 0 & t_value > 0, '1', color)) %>%
  mutate(color = if_else(weight > 0 & t_value < 0, '2', color)) %>%
  mutate(color = if_else(weight < 0 & t_value > 0, '2', color)) %>%
  mutate(color = if_else(weight < 0 & t_value < 0, '1', color))

ggplot(df, aes(x = weight, y = t_value, color = color)) + geom_point() +
  scale_colour_manual(values = c("red","royalblue3","grey")) +
  geom_label_repel(aes(label = ID)) + 
  theme_minimal() +
  theme(legend.position = "none") +
  geom_vline(xintercept = 0, linetype = 'dotted') +
  geom_hline(yintercept = 0, linetype = 'dotted') +
  ggtitle(pathway)


```

The pathway seems to be active since the majority of target genes with positive
weights have positive t-values (1st quadrant), and the majority of the ones with
negative weights have negative t-values (3d quadrant).

# Session information

```{r session_info, echo=FALSE}
options(width = 120)
sessioninfo::session_info()
```