microeco
microeco copied to clipboard
ChiLiubio
Convert net_construct or net_analyse object to trans_network or cal_network class
Dear @ChiLiubio,
I have already constructed my network plot using NetCoMi with measure as 'spieceasi' and I was wondering if it is possible to import the net_construct or net_analyse object from NetCoMi into trans_network or cal_network class.
The reason is I am interested in constructing the correlation heatmap as well as using the cal_sum_links to 'quantify' the interaction amongst the different taxa as seen in your package
I could directly create a new network plot as per your tutorial but that would result in a new network and would not be compatible to my existing network.
Hi @chrissy005
Good idea! There is only one tricky step. The reason is netConstruct function does not return an network object with igraph class. This is necessary in trans_network class. So you should create one based on the netConstruct analysis results by using a function like graph_from_adjacency_matrix in igraph package. Here is an example of all the operations.
# network example from igraph
adjm <- matrix(sample(0:1, 100, replace = TRUE, prob = c(0.9, 0.1)), ncol = 10)
g1 <- graph_from_adjacency_matrix(adjm)
# must assign names
V(g1)$name <- as.character(as_ids(V(g1)))
# please install microeco v1.1.0
library(microeco)
testnet <- trans_network$new()
testnet$cal_network(network_method = NULL)
# tax_table is necessary for the summary when you use cal_sum_links function. Please manually assign it
testnet$tax_table <- your_microtable$tax_table
# g1 to your trans_network object
testnet$res_network <- g1
testnet$cal_module()
testnet$get_node_table()
testnet$cal_sum_links(taxa_level = "Phylum")
Please feel free to tell me if something is wrong.
Best, Chi
Dear @ChiLiubio ,
I followed as you suggested but I run into an error at the last step while calculating links: "Error in data.frame(t(sapply(1:ecount(network), function(x) ends(network, : arguments imply differing number of rows: 9428, 0"
Could you attach your data and script so that I can find out the reason? To save your data, please use save function as the tutorial shows (https://chiliubio.github.io/microeco_tutorial/notes.html#save-function).
Dear @ChiLiubio ,
Sorry for the delay. I also encountered a few other problems with other functions so I have consolidated then and listing it here.
1) Unable to get edge_links and cal_sum_links for network object constructed using Netcomi and converted to igraph.
The codes I have used are as follows:
Compute layout: graph3 <- igraph::graph_from_adjacency_matrix(ASV_net_spieceasi_mb$adjaMat1, weighted = TRUE) must assign names: V(graph3)$name <- as.character(as_ids(V(graph3)))
testnet <- trans_network$new(dataset = microeco.dataset, cor_method = NULL, taxa_level = "OTU")
testnet$cal_network(network_method = NULL)
testnet$tax_table <- microeco.dataset$tax_table
g1 to your trans_network object testnet$res_network <- graph3
testnet$cal_module(method = "cluster_fast_greedy", module_name_prefix = "M")
testnet$cal_network_attr() testnet$res_network_attr
testnet$get_node_table(node_roles = TRUE) testnet$res_node_table
####no edges were returned. I suspect this could be a problem. #testnet$get_edge_table() #testnet$res_edge_table
testnet$get_adjacency_matrix() testnet$res_adjacency_matrix
plot the node classification in terms of the within-module connectivity and among-module connectivity. testnet$plot_taxa_roles(use_type = 1, add_label = TRUE)
plot node roles with phylum information testnet$plot_taxa_roles(use_type = 2)
testnet$cal_eigen() testnet$res_eigen
testnet_correlation <- trans_env$new(dataset = microeco.dataset, add_data = metadata.micronet[, 24:31])
testnet_correlation$cal_cor(add_abund_table = testnet$res_eigen)
This also returns a null object.and a network plot
testnet$plot_network(node_color = "module")
This line returns an error "Error in data.frame(t(sapply(1:ecount(network), function(x) ends(network, :
arguments imply differing number of rows: 10306, 0" testnet$cal_sum_links(taxa_level = "Phylum")
microeco.dataset.1.zip ASV_net_spieceasi_mb_1.zip [com.rarefied.min.int.exclude.1.3.rooted.1.zip]
2) I tried constructing a network with the microeco package for comparison. However while I look at the plot using igraph, I do not see the 16 clusters by colors and it also returns a null object.
The codes and data used are as follows
microeco_transnetwork_spieceasi$cal_module(method = "cluster_fast_greedy", module_name_prefix = "M")
calculate network attributes microeco_transnetwork_spieceasi$cal_network_attr() microeco_transnetwork_spieceasi$res_network_attr
get node properties, calculate node roles, i.e. Module hubs, Network hubs, Connectors and Peripherals microeco_transnetwork_spieceasi$get_node_table(node_roles = TRUE) microeco_transnetwork_spieceasi$res_node_table
Get the edge property table, including connected nodes, label and weight. microeco_transnetwork_spieceasi$get_edge_table() microeco_transnetwork_spieceasi$res_edge_table
Get the adjacency matrix from the network graph microeco_transnetwork_spieceasi$get_adjacency_matrix() microeco_transnetwork_spieceasi$res_adjacency_matrix
Now, we show the eigengene analysis of modules. microeco_transnetwork_spieceasi$cal_eigen() microeco_transnetwork_spieceasi$res_eigen
default parameter represents using igraph plot.igraph function microeco_transnetwork_spieceasi$plot_network(method = "igraph", layout = layout_with_kk,node_color = "module")
Nonetheless, here I am able to plot the sum_links microeco_transnetwork_spieceasi$plot_sum_links(plot_pos = TRUE, plot_num = 12, color_values = RColorBrewer::brewer.pal(12, "Paired"))
Could you also advise on how I can save the circos plot created here? I tried it in as follows but it returned a null object.
pdf("test.pdf") microeco_transnetwork_spieceasi$plot_sum_links(plot_pos = TRUE, plot_num = 12, color_values = RColorBrewer::brewer.pal(12, "Paired")) dev.off()
microeco_transnetwork_spieceas.1.zip
Hi.
When I load ASV_net_spieceasi_mb.RData you provided, the session only has dataset, which seems like the example dataset in microeco package, not the ASV_net_spieceasi_mb presented in the step.
load("ASV_net_spieceasi_mb.RData")
# if you use save function, it should be like
save(ASV_net_spieceasi_mb, file = "ASV_net_spieceasi_mb.RData")
The figure from plot_sum_links function is not the traditional plot in R. It is based on the js. So please save it from the Rstudio window directly, not the script.
Dear @ChiLiubio ,
sorry about providing wrong files.
Attaching the new ones here.
com.rarefied.min.int.exclude.1.3.rooted.1.zip microeco.dataset.1.zip ASV_net_spieceasi_mb_1.zip
Let me know if these are ok.
Hi. I found that cal_sum_links require the label of edges to partition the links into positive and negative. I will update the package later. To temporarily solve this, we can assign the labels into network: E(graph3)$label <- ifelse(E(graph3)$weight > 0, '+', '-'). The following is all I run.
library(microeco)
library(magrittr)
library(igraph)
load("ASV_net_spieceasi_mb.RData")
load("microeco.dataset.RData")
tmp <- ASV_net_spieceasi_mb$adjaMat1
# 1 of diagonal should be removed
diag(tmp) <- 0
graph3 <- igraph::graph_from_adjacency_matrix(tmp, weighted = TRUE)
# must assign names:
V(graph3)$name <- as.character(as_ids(V(graph3)))
E(graph3)$label <- ifelse(E(graph3)$weight > 0, '+', '-')
testnet <- trans_network$new(dataset = microeco.dataset,
cor_method = NULL, taxa_level = "OTU")
testnet$cal_network(network_method = NULL)
testnet$tax_table <- microeco.dataset$tax_table
# g1 to your trans_network object
testnet$res_network <- graph3
testnet$cal_module(method = "cluster_fast_greedy",
module_name_prefix = "M")
testnet$cal_network_attr()
testnet$res_network_attr
testnet$get_node_table(node_roles = TRUE)
# testnet$res_node_table
####no edges were returned. I suspect this could be a problem.
#testnet$get_edge_table()
#testnet$res_edge_table
# testnet$get_adjacency_matrix()
# testnet$res_adjacency_matrix
# plot the node classification in terms of the within-module connectivity and among-module connectivity.
testnet$plot_taxa_roles(use_type = 1, add_label = TRUE)
# plot node roles with phylum information
testnet$plot_taxa_roles(use_type = 2)
testnet$cal_sum_links(taxa_level = "Phylum")
@ChiLiubio, Thank you for your help. It is all solved now.
However, I am still struggling to adjust the font size on the circos plot while plotting the sum_links.
Also, the only option to save it form the plots panel is as png and not pdf.
Yes. The reason is it is made by the js, not R, for the interactive vis. So I will consider implement another visualization way. Thanks.
@ChiLiubio , I understand that I could actually adjust the font size by passing arguments to the chorddiag function while plotting the sum_links. I also see this in the manual:
"Method plot_sum_links(): Plot the summed linkages among taxa using chorddiag package https://github.com/mattflor/chorddiag. Usage: trans_network$plot_sum_links( plot_pos = TRUE, plot_num = NULL, color_values = RColorBrewer::brewer.pal(8, "Dark2"), ... ) Arguments: plot_pos default TRUE; If TRUE, plot the summed positive linkages; If FALSE, plot the summed negative linkages. plot_num default NULL; number of taxa presented in the plot. color_values default RColorBrewer::brewer.pal(8, "Dark2"); colors palette for taxa. ... parameters pass to chorddiag::chorddiag function."
Could you provide an example on how may I pass additional arguments using '...' to chorddiag?
Also, would it be possible to plot the circos plot without the edge numbers? i.e. I would like to get rid off the numbers on the outside of the circular plot?
Hi. The parameter passing means you can use the parameters from chorddiag fundtion in the plot_sum_links function like this:
library(chorddiag)
?chorddiag
# if you want to use categorynameFontsize, you can directly use it in the plot_sum_links
your_trans_network$plot_sum_links(categorynameFontsize = 40)
For the edge number issue, please first check whether chorddiag has a param to remove it. Another way to do this is to extract the data coming from the cal_sum_links function and visualize that directly use the chorddiag function.
Hi. New microeco version has been released and solved the adjustment of plot in plot_sum_links function. Another static plot based on circilize package has been implemented and can be easily adjusted and saved into local file.
Hello @ChiLiubio ,
thank you for attending to that request.
So all I need to know is to install the new version of microeco and save the plot as pdf?
Or are there any other specific codes?