1

Excuse me because I feel like this doubt should be simpler, but I can't find a satisfactory answer.

I have two quantitative bipartite networks (that show the ecological relationships among A-B, and B-C). My problem is that I do not know how join both in order to make a directed quantitative tripartite network (like a typical food web). The B level of each bipartite network have the same vertex composition (in quantitie and vertex names). In addition, the levels A and C can not interact among them. For this reason, the order and direction of the final tripartite network it should be C->B->A

Any suggestions?

Thanks for your attention!

Alejandro López Núñez
  • 25
  • 5
  • I am confused with how you write about directionality... Can you provide a (toy) data example? – Michał Oct 21 '15 at 14:22
  • Hello, At first, I have two bipartite networks: For example: Bipartite Network 1 explores relationships between Plant-Herbivore (A-B). Bipartite Network 2 explore relationships between Herbivore-Predator (B-C). I would like to obtain a new matrix that represent the relationships among A,B,C. Due to predators only feed on herbivores, and herbivores only feed on plants (it is not possible that predators feed on plants or herbivores feed on predators) I need a directed network (like food web). Follow the link to see a toy data. [link] https://goo.gl/Gc9tMF – Alejandro López Núñez Oct 22 '15 at 09:31

1 Answers1

4

How about this:

Cook-up some data (i was hoping you would cook it for me :) )

library(igraph)
set.seed(666)
# herbivore-plant
m_hp <- matrix( rbinom(12, 10, p=0.2), 4, 3)
dimnames(m_hp) <- list(
  consuming=paste0("h", seq(1, nrow(m_hp))),
  consumed=paste0("p", seq(1, ncol(m_hp)))
)
# carnivore-herbivore
m_ch <- matrix( rbinom(20, 10, p=0.2), 5, 4)
dimnames(m_ch) <- list(
  consuming=paste0("c", seq(1, nrow(m_ch))),
  consumed=paste0("h", seq(1, ncol(m_ch)))
)

... so it looks like yours (I presume):

m_hp

##          consumed
## consuming p1 p2 p3
##        h1  3  1  0
##        h2  1  3  1
##        h3  5  5  3
##        h4  1  2  0

m_ch

##          consumed
## consuming h1 h2 h3 h4
##        c1  0  4  0  2
##        c2  1  2  1  2
##        c3  1  2  3  3
##        c4  3  5  1  2
##        c5  0  2  0  2

Now turn them into igraph objects via edgelists

el_hp <- as.data.frame(as.table(m_hp), stringsAsFactors = FALSE)
el_ch <- as.data.frame(as.table(m_ch), stringsAsFactors = FALSE)
el <- rbind(el_hp, el_ch)
g <- graph.data.frame( el[el$Freq != 0 , ]  )
V(g)$type <- substr(V(g)$name, 1, 1)

Adjacency matrix of joint network:

get.adjacency(g, sparse=FALSE, attr="Freq")

##    h1 h2 h3 h4 c2 c3 c4 c1 c5 p1 p2 p3
## h1  0  0  0  0  0  0  0  0  0  3  1  0
## h2  0  0  0  0  0  0  0  0  0  1  3  1
## h3  0  0  0  0  0  0  0  0  0  5  5  3
## h4  0  0  0  0  0  0  0  0  0  1  2  0
## c2  1  2  1  2  0  0  0  0  0  0  0  0
## c3  1  2  3  3  0  0  0  0  0  0  0  0
## c4  3  5  1  2  0  0  0  0  0  0  0  0
## c1  0  4  0  2  0  0  0  0  0  0  0  0
## c5  0  2  0  2  0  0  0  0  0  0  0  0
## p1  0  0  0  0  0  0  0  0  0  0  0  0
## p2  0  0  0  0  0  0  0  0  0  0  0  0
## p3  0  0  0  0  0  0  0  0  0  0  0  0

Graphically

t <- match(V(g)$type, c("p", "h", "c") )
plot(g, vertex.color=t )

enter image description here

or even

l <- layout_with_fr(g, miny=t, maxy=t  )
plot(g, vertex.color=t, layout=l, edge.width=E(g)$Freq)

enter image description here

Michał
  • 2,755
  • 1
  • 17
  • 20