miRNA-gene interaction network using miRTarBase and networkD3 in R

code.R

nasza = read.csv("nasza.csv", header=T, stringsAsFactors = F)



library("openxlsx")
library(dplyr)
miRtarbase = openxlsx::read.xlsx("hsa_MTI.xlsx")
refdb = filter(miRtarbase, Support.Type == "Functional MTI" | Support.Type == "Non-Functional MTI")

miR = unique(nasza$hsa.miRNA)
miR = miR[-which(miR=="")]

#korekty
miR[36] = "hsa-miR-106a"

miRy = c(miR,paste0(miR,rep("-5p",length(miR))),paste0(miR,rep("-3p",length(miR))))

target = list()
for(i in 1:length(miRy)) {
  target[[i]] = unique(refdb$Target.Gene[which(refdb$miRNA == miRy[i])])
}

pubtab1 = data.frame(miR = miR, targets = NA)
target2 = list()
for(i in 1:length(miR)) {
  target2[[i]] = unique(c(target[[i]],target[[i+length(miR)]],target[[i+(length(miR)*2)]]))
  pubtab1$targets[i] = paste0(target2[[i]], collapse = ", ")
}

ftemp = table(unlist(target))
ftemp2 = ftemp[order(ftemp, decreasing = T)]
write.csv(ftemp2,"ftemp2.csv")

# Wszystkie, ale nie weak
write.csv(pubtab1,"wszystkie_nieweak.csv")
tab1 = pubtab1


####
refdb = miRtarbase

target = list()
for(i in 1:length(miRy)) {
  target[[i]] = unique(refdb$Target.Gene[which(refdb$miRNA == miRy[i])])
}

pubtab1 = data.frame(miR = miR, targets = NA)
target2 = list()
for(i in 1:length(miR)) {
  target2[[i]] = unique(c(target[[i]],target[[i+length(miR)]],target[[i+(length(miR)*2)]]))
  pubtab1$targets[i] = paste0(target2[[i]], collapse = ", ")
}

# Wszystkie, ale nie weak
write.csv(pubtab1,"wszystkie_plusweak.csv")
tab2 = pubtab1

ftemp = table(unlist(target))
ftemp2 = ftemp[order(ftemp, decreasing = T)]
temp = data.frame(lapply(ftemp2, type.convert), stringsAsFactors=FALSE)

### Venn
library(VennDiagram)

calculate.overlap(target2)

#BiocManager::install("SpidermiR")
miRbase = read.csv("miRNA.csv", stringsAsFactors = F)
czy_istnieja = match(miRy, c(miRbase$Mature1_ID,miRbase$Mature2_ID,miRbase$ID))
miRy_final = miRy[!is.na(czy_istnieja)]
cat(paste0(miRy_final,collapse = "\n"))

list<-SpidermiRdownload_miRNAvalidate(miRy_final)
org<-SpidermiRquery_species(species)
net_shar_prot<-SpidermiRquery_spec_networks(organismID = org[6,],
                                            network = "PATH")
out_net<-SpidermiRdownload_net(net_shar_prot)

list<-SpidermiRdownload_miRNAvalidate(validated)
list_circ<-SpidermiRdownload_miRNAextra_cir(miRNAextra_cir)
mir_pharmaco<-SpidermiRdownload_pharmacomir(pharmacomir=pharmacomir)

geneSymb_net<-SpidermiRprepare_NET(organismID = org[6,],
                                   data = out_net)

refdb = filter(miRtarbase, Support.Type == "Functional MTI" | Support.Type == "Non-Functional MTI")
target = list()
for(i in 1:length(miRy)) {
  target[[i]] = unique(refdb$Target.Gene[which(refdb$miRNA == miRy_final[i])])
}
ftemp = table(unlist(target))
ftemp = ftemp[order(ftemp, decreasing = T)]
temp = as.data.frame(ftemp)
ftemp2 = ftemp[ftemp>5]
write.csv(temp,"forGOrilla.csv")
temp2 = data.frame(gene = NA, miR = NA)
for (i in 1:length(ftemp2)) {
  temp = names(ftemp2)[i]
  for (ii in 1:length(target)) {
    if(!is.na(match(temp,target[[ii]]))) { temp2 = rbind(temp2, c(temp,miRy_final[ii]))} 
  }
}
temp2 = temp2[complete.cases(temp2),]

net = function (data) 
{
  library(networkD3)
  colnames(data) <- c("V1", "V2")
  if (length(grep("hsa", data$V1)) != 0 | length(grep("hsa", 
                                                      data$V2)) != 0) {
    IDs2 <- sort.int(unique(c(data$V1, data$V2)), decreasing = FALSE)
    IDs2 <- data.frame(ID = seq_along(IDs2) - 1, name = IDs2)
    dataIDs2 <- merge(data, IDs2, by.x = "V1", by.y = "name")
    dataIDs2$V1 <- dataIDs2$ID
    dataIDs2$ID <- NULL
    dataIDs2 <- merge(dataIDs2, IDs2, by.x = "V2", by.y = "name")
    dataIDs2$V2 <- dataIDs2$ID
    dataIDs2$ID <- NULL
    dataIDs2 <- dataIDs2[, c("V1", "V2")]
    data2 <- data[order(data$V1, data$V2, decreasing = FALSE), 
                  ]
    dataIDs2 <- dataIDs2[order(dataIDs2$V1, dataIDs2$V2, 
                               decreasing = FALSE), ]
    att <- as.data.frame(sort(unique(unlist(data)), decreasing = FALSE))
    colnames(att)[1] <- "v1"
    att$v2 <- ""
    att$v2 <- replace(att$v2, att$v2 == "", "gene")
    att$v2[(grep("[a-z]", att$v1))] <- "Pharmaco"
    att$v2[grep("hsa", att$v1)] <- "miRNA"
    att$v2[grep("orf", att$v1)] <- "gene"
    att <- as.data.frame(att[order(att$v2, decreasing = FALSE), 
                             ])
    i <- sapply(att, is.factor)
    att[i] <- lapply(att[i], as.character)
    colnames(att)[1] <- "name"
    colnames(att)[2] <- "Group"
    attr2 <- merge(att, IDs2)
    attr2 <- attr2[order(attr2$ID), ]
    ColourScal <- "d3.scaleOrdinal()\n    .domain([\"gene\", \"Pharmaco\",\"miRNA\"])\n    .range([\"#000000\", \"#37bf30\" , \"#37bf30\"]);"
    return(forceNetwork(Links = dataIDs2, Nodes = attr2, 
                        Source = "V1", Target = "V2", NodeID = "name", Group = "Group", 
                        height = 1500, width = 1500, opacity = 1, zoom = F, fontFamily = "Calibri",
                        legend = F, opacityNoHover = FALSE, colourScale = JS(ColourScal), bounded = F, charge = -2000,
                        #radiusCalculation = JS("Math.sqrt(d.nodesize)+6"),
                        #linkDistance = JS("function(d){return d.value * 10}"),
                        #linkWidth = JS("function(d) { return Math.sqrt(d.value); }"),
                        fontSize = 26))
  }
  if (length(grep("hsa", data$V1)) == 0) {
    .SpidermiRvisualize_gene(data)
  }
}

net(data=temp2)