Tutorial.Rmd

---
title: "Tutorial"
author: "Marco Mello & Renata Muylaert"
date: "23/09/2020"
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```


## Summary

1. [Get ready](#ready)
2. [Figure 21.1](#fig1)
3. [Figure 21.2](#fig2)
4. [Figure 21.3](#fig3)
5. [Figure 21.4](#fig4)


## 1. Get ready {#ready}

Load the required packages:

```{r}
library("igraph")
library("ggplot2")
library("reshape2")
```


## 2. Figure 21.1 {#fig1}

Import the data:

```{r}
studies = read.delim("Analysis/studies.txt", row.names=1)
```


Plot the barplot:

```{r, fig.width = 17, fig.height = 5}
par(mfrow = c(1,1), mar = c(6,7,1,1), xpd=TRUE)
counts <- table(studies$Phyllostomidae, studies$Links)
barplot(counts, main="",
        xlab="Topic",
        ylab = "Number of studies",
        col=gray.colors(length(rownames(counts))), 
        border = F, 
        beside=TRUE,
        las = 0,
        mgp = c(4,1,0),
        cex.lab=3,
        cex.axis=2, 
        cex.main=3, 
        cex.sub=3,
        cex.names=1.5)
legend(x = 1,
       y = 8, 
       title = "Phyllostomidae as a model?",
       cex = 1.5,
       bty="n",
       fill=gray.colors(length(rownames(counts))), 
       legend=rownames(counts))
par(mfrow = c(1,1))
```


## 3. Draw Figure 21.2{#fig2}

Use the same data as in Fig.21.1:

```{r}
head(studies, 3)
```


Plot the histogram:

```{r, fig.width = 12, fig.height = 6}
par(mfrow = c(1,1), mar = c(5,7,1,3))
hist(studies$Year,
     breaks=5,
     col = adjustcolor("grey", alpha.f = .5),
     border = F,
     xlab = "Year",
     ylab = "Number of studies",
     main = "",
     mgp = c(4,1,0),
     cex.lab=3,
     cex.axis=2, 
     cex.main=3, 
     cex.sub=3)
par(mfrow = c(1,1))
```


## 4. Draw Figure 21.3{#fig3}

Import the data:

```{r}
softmat = as.matrix(read.delim("Analysis/software.txt", row.names=1))
```


Inspect the object:

```{r}
class(softmat)
softmat
```


Create an igraph object:

```{r}
softnet <- graph_from_incidence_matrix(softmat, weighted = NULL)
```

Inspect the object:

```{r}
class(softnet)
softnet
E(softnet)
V(softnet)
```


Name the two sets of vertices: rows will be "software" and columns, "references":

```{r}
V(softnet)$set = ifelse(V(softnet)$type == FALSE, "software", "reference")
V(softnet)$set
```


Set the layout for the graph:

```{r}
lref <- layout.fruchterman.reingold(softnet)
```


Set the edge curvatures:

```{r}
curvesref = curve_multiple(softnet)
```


Set the edge mode and width:

```{r}
E(softnet)$arrow.mode = 0
E(softnet)$width = 1
```


Set the node shapes:

```{r}
V(softnet)$shape = V(softnet)$set
V(softnet)$shape = gsub("reference","circle",V(softnet)$shape)
V(softnet)$shape = gsub("software","square",V(softnet)$shape)
```


Set the node colors:

```{r}
V(softnet)$color = V(softnet)$set
V(softnet)$color = gsub("reference","grey10",V(softnet)$color)
V(softnet)$color = gsub("software","grey50",V(softnet)$color)
```


Draw the graph with node colors by class:

```{r, fig.width = 8, fig.height = 6}
par(mfrow = c(1,1))
plot(softnet, 
     vertex.color = V(softnet)$color, 
     vertex.frame.color = V(softnet)$color, 
     #vertex.shape = vertexshape, 
     vertex.size = 12,
     #vertex.size = V(softnet)$size,
     vertex.label = V(softnet)$name,
     vertex.label.color = "white",
     vertex.label.cex = .5,
     edge.color = adjustcolor("grey70", alpha.f = .5),
     #edge.color = "#FFFD00", 
     #edge.curved = curvesref,
     edge.curved = 0.2,
     edge.width = 7,
     #layout = layout_in_circle,
     layout = lref) 
legend(x = 0.7,y = -0.6, legend = c("References", "Software"), 
       pch = c(19,15), title="Legend", 
       text.col = "gray20", title.col = "black", box.lwd = 0,
       cex = 2, col=c("grey10", "grey50"))
par(mfrow = c(1,1))
```



## 5. Draw Figure 21.4{#fig4}

Import the data:

```{r}
sarmentomat = as.matrix(read.delim("Analysis/sarmento.txt", row.names=1))
```


Inspect the object:

```{r}
class(sarmentomat)
sarmentomat
```


Create an igraph object:

```{r}
sarmentonet <- graph_from_incidence_matrix(sarmentomat, 
                                           directed = F, weighted = TRUE) 
```

Inspect the object:

```{r}
class(sarmentonet)
sarmentonet
E(sarmentonet)
V(sarmentonet)
```


Specify which nodes represent which taxonomic groups:

```{r}
V(sarmentonet)$set[1:12] = "Bats"
V(sarmentonet)$set[13:22] = "Birds"
V(sarmentonet)$set[23:89] = "Plants"
```


Set the layout for the graphs:

```{r}
lsar <- layout_nicely(sarmentonet)
```


Set the edge curvatures:

```{r}
curvessar = curve_multiple(sarmentonet)
```


Set the edge mode and width:

```{r}
E(sarmentonet)$arrow.mode = 0
E(sarmentonet)$width = E(sarmentonet)$weight/5+1
```


Calculate the Louvain modularity (resolution = 1.0):

```{r}
sarmentonet.lou = cluster_louvain(sarmentonet)
```

Import the "diamond" vertex shape:

```{r}
source("Analysis/MyDiamond.R")
```

Set the vertex shapes:

```{r}
V(sarmentonet)$shape = V(sarmentonet)$set
V(sarmentonet)$shape = gsub("Bats","diamond",V(sarmentonet)$shape)
V(sarmentonet)$shape = gsub("Birds","square",V(sarmentonet)$shape)
V(sarmentonet)$shape = gsub("Plants","circle",V(sarmentonet)$shape)
```


Set the colors for the nodes and clouds based on the modularity analysis:

```{r}
colrs <- gray.colors(length(sarmentonet.lou), start = 0.3,
                     end = 0.6, gamma = 1.5, alpha = NULL)
V(sarmentonet)$color <- colrs[sarmentonet.lou$membership]
clouds = gray.colors(length(sarmentonet.lou), start = 0.7,
                     end = 0.9, gamma = 1.5, alpha = 0.1)
```


Plot the graph with node colors by Louvain modularity (grey tones) + clouds around modules:

```{r, fig.width = 8, fig.height = 6}
par(mfrow=c(1,1),mar=c(1,1,1,5))
plot(sarmentonet.lou,
     sarmentonet,
     col = V(sarmentonet)$color,
     mark.border="lightgrey", 
     mark.col=clouds,
     vertex.size=7.5,
     vertex.label=V(sarmentonet)$set,
     vertex.label.color="white",
     vertex.label.cex=.5,
     edge.color = adjustcolor("grey", alpha.f = .5),
     edge.curved=0.3,
     edge.width = 3,
     layout=lsar)
legend(x = 0.9,y = 1.0, legend = c("Bats", "Birds", "Plants"),
       pch = c(18,15,19),  title="Taxon",
       text.col = "gray20", title.col = "black",
       box.lwd = 0, cex = 2, col=c("grey", "grey", "grey"))
par(mfrow=c(1,1))
```