cooccurrence_network.Rmd

---
title: "cooccurrence_network"
output: html_document
---

# packages 
```{r}
library(GGally)
library(network)
library(sna)
library(ggplot2)
library(igraph)
library(cooccur)
```

```{r}
# # random graph
# net = rgraph(10, mode = "graph", tprob = 0.5)
# net = network(net, directed = FALSE)
# 
# # vertex names
# network.vertex.names(net) = letters[1:10]
# 
# ggnet2(net)
# sp_by_sp_matrix <- create.N.matrix(subject_species_matrix_transposed)
# net = network(sp_by_sp_matrix)
# 
# ggnet2(net, label = TRUE, mode = "circle")

sp_by_sp_matrix <- create.N.matrix(t(subject_tipglomcluster_matrix))
net = network(sp_by_sp_matrix)

# cool i think this works?
```

# extracting edge list from cooccur object
```{r}
# set rownames
# subject_species_matrix = subject_species_matrix %>% column_to_rownames("...1")

# no subsetting
# subject_species_matrix_transposed <- t(subject_species_matrix)
cooccur.subjectCluster <- cooccur(mat=t(subject_species_matrix), spp_names = T)

# View(cooccur.subjectSpecies$results)

edgeList <- cooccur.subjectCluster$results[, c("sp1_name", "sp2_name", "obs_cooccur")]
edgeList <- subset(edgeList, obs_cooccur > 20)
# 
# net = network(edgeList, directed C= F)

# ggnet2(net, label = TRUE, size = "degree", label.size = 2, edge.size = "weights") +
#   guides(size = "none")

# subsetting the df to take out some species with <20 observations 
# species_to_exclude <- c("Desulfitobacterium hafniense", "Desulfovibrio desulfuricans cluster 3", "Thermodesulfobium narugense", "unclassified Firmicute Unc01cxc", "Archaeoglobus sulfaticallidus", "Desulfovibrio desulfuricans cluster 2")
# species_to_include <- setdiff(rownames(subject_species_matrix_transposed), species_to_exclude)
# subject_species_matrix_transposed_subset <- subject_species_matrix_transposed[rownames(subject_species_matrix_transposed) %in% species_to_include,]
# cooccur.subjectSpecies <- cooccur(mat=subject_species_matrix_transposed_subset, spp_names = T)
# 
# edgeList <- cooccur.subjectSpecies$results[, c("sp1_name", "sp2_name", "obs_cooccur")]
# edgeList <- subset(edgeList, obs_cooccur > 5)

igraph_net <- graph_from_data_frame(d = edgeList, directed = F)
E(igraph_net)$width <- E(igraph_net)$obs_cooccur/10
V(igraph_net)$label.size <- 0
deg <- degree(igraph_net, mode="all")
V(igraph_net)$size <- 1


radian.rescale <- function(x, start=0, direction=1) {
  c.rotate <- function(x) (x + start) %% (2 * pi) * direction
  c.rotate(scales::rescale(x, c(0, 2 * pi), range(x)))
}

lab.locs <- radian.rescale(x=1:vcount(igraph_net), direction=-1, start=0)



l <- layout_in_circle(igraph_net)
# 
# plot(igraph_net, layout = l, edge.curved=.3, vertex.color = "lavender", vertex.frame.color = "black", vertex.label.cex = 0.7, vertex.label.color = "black", vertex.label.degree = lab.locs, vertex.label.dist=4, edge.alpha = 0.5)

order <- c(1, 2, 3, 4, 7, 5, 6, 8)
igraph_net <- permute(igraph_net, order)

unlabeled_network <- plot(igraph_net, layout = l, edge.curved=.3, vertex.label = NA, edge.alpha = 0.5, vertex.color = "black", vertex.frame.color = "black")


labeled_network <- plot(igraph_net, layout = l, edge.curved=.3, vertex.color = "black", vertex.frame.color = "black", vertex.label.cex = 0.7, vertex.label.color = "black", vertex.label.degree = lab.locs, vertex.label.dist=4, edge.alpha = 0.5)

ggsave(plot = unlabeled_network, file = "unlabeled_network.jpeg", width = 7, height = 7)
ggsave(plot = labeled_network, file = "labeled_network.jpeg", width = 7, height = 7)

# vertex.label.dist = 3
```

# cooccurrence statistics
# FOR JESS 
```{r}
# number of co-occurrence instances
# sum(edgeList$obs_cooccur)

# nuimber of co-occurrence instances in Bilophila

# sum(subset(edgeList$obs_cooccur, edgeList$sp1_name == "Bilophila wadsworthia" | edgeList$sp2_name == "Bilophila wadsworthia"))

# make dataframe for co-occurrence pairs (long) by subject

# subject_species_edge <- subject_species_matrix %>% 
#   data.frame(check.names = F) %>%
#   rownames_to_column("subject_id")

clusters <- unique(c(edgeList$sp1_name, edgeList$sp2_name))

# subject_species_edge_cooccur <- data.frame()

x = map_dfr(1:nrow(subject_species_edge[1:5,]), function(x) {
  subject_id = subject_species_edge[x, "subject_id"]
  print(subject_id)
  
  # out = data.frame(subject = subject_id)
  
  for (sp1_name in colnames(subject_species_edge)) {
    if (sp1_name %in% clusters) {
      
      if (subject_species_edge[x, sp1_name] == 1) {
        # subject_species_edge_cooccur$subject_id <- subject_id
        # subject_species_edge_cooccur$sp1[subject_species_edge_cooccur$subject_id == subject_id] <- column_name
        
        sp1 = sp1_name
        print(sp1)
        #   for (sp2_name in colnames(subject_species_edge)) {
        #     if (sp2_name %in% clusters) {
        #       if (sp2_name != sp1) {
        #         if (subject_species_edge[x, sp2_name] == 1) {
        #           
        #           sp2 = sp2_name
        #           
        #           
        #     }
        #   }
        # }
        # }
      }
      
    }
    # } )
    return(list(subject_id, sp1))
  }
  
})

for (n_row in 1:nrow(subject_species_edge[1:5,])) {
  subject_id <- subject_species_edge[n_row, "subject_id"]
  print(subject_id)
  for (column_name in colnames(subject_species_edge)) {
    if (column_name %in% clusters) {

      if (subject_species_edge[n_row, column_name] == 1) {
        subject_species_edge_cooccur$subject_id <- subject_id
        subject_species_edge_cooccur$sp1[subject_species_edge_cooccur$subject_id == subject_id] <- column_name

        print(column_name)
      #   for (column_name in colnames(subject_species_edge)) {
      #     if (column_name %in% clusters) {
      #       if (subject_species_edge[n_row, column_name] == 1)
      #   }
      # }
    }
  }
  }
}

# View( subject_species_edge_cooccur)

# subject_species_edge_long <- subject_species_edge %>%
#   data.frame() %>%
#   rownames_to_column("subject_id") %>%
#   pivot_longer(cols = !starts_with("subject_id"))

```