Nothing
R/IIS.R
In BiOFI: Feature Identification Between Metabolome and Microbiome in Disease and Health
Defines functions
IIS
Documented in
IIS
#' @title IIS
#' @description
#' A function is used to calculates and identify the key nodes,the formula of IIS is:
#' IIS= nDFS x DFS + nDS x DS + nES x ES + nAS x AS;
#' nDFS,nDS,nES,nAS indict the weight value of DFS,DS,ES and AS, and nDFS + nDS + nES + nAS = 1;
#' DFS means Difference Score;
#' DS means Degree Score;
#' ES means Edge Score;
#' AS means Abundance Score.
#'
#' @param microApath A data frame including microbes and predicted KEGG pathways.
#' @param metaApath A data frame including metabolites with compound ID and predicted KEGG pathways from Meta2pathway function.
#' @param conf A data frame including confounders. Default value is NULL.
#' @param groupInfo A data frame including sample grouping information
#' @param nDFS A value indicated the weight of difference Score in the IIS formula. Default value is 0.4.
#' @param nDS A value indicated the weight of degree score in the IIS formula. Default value is 0.3.
#' @param nES A value indicated the weight of edge score in the IIS formula. Default value is 0.2.
#' @param nAS A value indicated the weight of abundance score in the IIS formula. Default value is 0.1.
#' @param NS A threshold value of IIS to identify the key nodes. Default value is 2.5.
#' @param r A threshold value of correlation coefficient to construct the network. Default value is 0.5.
#' @param p_adjust A threshold value of correlation adjust p value to construct the network. Default value is 0.05.
#' @return IIS_res A list including IIS and other sub-score results.
#' @export IIS
#' @importFrom igraph graph_from_data_frame
#' @importFrom dplyr group_by
#' @importFrom dplyr summarise
#' @import ggrepel
#' @import ppcor
#' @importFrom utils write.csv
#' @importFrom stats cor.test
#' @importFrom stats p.adjust
#' @importFrom stats wilcox.test
#' @examples
#' IIScore <- IIS(microApath = micro.eg, metaApath = metabo.eg,
#' conf = confounder.eg, groupInfo = groupInfo.eg)
IIS<-function(microApath, metaApath, conf = NULL, groupInfo = NULL, nDFS = 0.4,
nDS = 0.3, nES = 0.2, nAS = 0.1, NS = 2.5, r = 0.5, p_adjust = 0.05){
from = NULL;to = NULL
A=microApath
B=metaApath
mmApath=get0("sysdata", envir = asNamespace("BiOFI"))$mm4path
MMData1=data.frame(metaandpath=colnames(B),group=0)
for(i in 1:length(colnames(B))){
if(MMData1$metaandpath[i]%in% unique(mmApath$meta_path)){
MMData1$group[i]='met_path'
}else{
MMData1$group[i]='metabolite'
}
}
if(!'met_path'%in% unique(MMData1$group)){
stop('please tranfer metabolites to pathway using Meta2pathway function')
}
MMData2=data.frame(microandpath=colnames(A),group=0)
for(i in 1:length(colnames(A))){
if(substr(MMData2$microandpath[i],1,3)=='k__'){
MMData2$group[i]='microbe'
}else{MMData2$group[i]='mic_path'}
}
if(length(rownames(groupInfo))==0){
stop('Group information is lack, please check!')
}
colnames(B)=paste(colnames(B),'-M',sep=" ")
mic_meta_cor = matrix(NA, nrow = ncol (A), ncol = ncol (B))
mic_intra_cor=matrix(NA, nrow = ncol (A), ncol = ncol (A))
met_intra_cor=matrix(NA, nrow = ncol (B), ncol = ncol (B))
rownames (mic_meta_cor) = colnames (A)
colnames (mic_meta_cor) = colnames (B)
rownames (mic_intra_cor) = colnames (A)
colnames (mic_intra_cor) = colnames (A)
rownames (met_intra_cor) = colnames (B)
colnames (met_intra_cor) = colnames (B)
mic_meta_cor_p = mic_meta_cor
mic_meta_cor_p.adjust = mic_meta_cor
mic_intra_cor_p=mic_intra_cor
mic_intra_cor_p.adjust=mic_intra_cor
met_intra_cor_p=met_intra_cor
met_intra_cor_p.adjust=met_intra_cor
for (m in colnames (B)) {
if(length(rownames(conf))==0){
mic_meta_cor[,m] = apply (A, MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x),as.numeric(B[,m]), method = "spearman",use = "pairwise.complete.obs")$estimate)
mic_meta_cor_p[,m] = apply (A, MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x),as.numeric(B[,m]), method ="spearman", use = "pairwise.complete.obs")$p.value)
}else{
tryCatch({
mic_meta_cor[,m] = apply (A, MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method = "spearman")$estimate)
mic_meta_cor_p[,m] = apply (A, MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method ="spearman")$p.value)
}, error=function(e) {
})
}
}
mic_meta_cor_p.adjust = apply(mic_meta_cor_p, 2, FUN = function(x){
p.adjust(x,method = "BH")})
del_var <- c()
for (m in 1:ncol (A)) {
del_var <- append(del_var,m)
if(length(rownames(conf))==0){
temp_cor = apply(as.data.frame(A[,-del_var]), MARGIN = 2,FUN = function (x)
cor.test (as.numeric(x), as.numeric(A[,m]), method = "spearman", use = "pairwise.complete.obs")$estimate)
mic_intra_cor [ , m] = c(rep(NA,m),temp_cor)
temp_p = apply(as.data.frame(A[,-del_var]), MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x), as.numeric(A[,m]), method ="spearman", use = "pairwise.complete.obs")$p.value)
mic_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
tryCatch({
if(m!=ncol (A)){
temp_cor = apply (as.data.frame(A[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(A[,m]), conf, method = "spearman")$estimate)
mic_intra_cor[,m] = c(rep(NA,m),temp_cor)
temp_p = apply (as.data.frame(A[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(A[,m]), conf, method ="spearman")$p.value)
mic_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
mic_intra_cor[,m] = rep(NA,m)
mic_intra_cor_p[,m] = rep(NA,m)
}
}, error=function(e) {
})
}
}
mic_intra_cor_p.adjust = apply(mic_intra_cor_p, 2, FUN = function(x){
p.adjust(x,method = "BH")})
del_var <- c()
for (m in 1:ncol (B)) {
del_var <- append(del_var,m)
if(length(rownames(conf))==0){
temp_cor = apply(as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x),as.numeric(B[,m]), method = "spearman", use = "pairwise.complete.obs")$estimate)
met_intra_cor [ , m] = c(rep(NA,m),temp_cor)
temp_p = apply(as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x), as.numeric(B[,m]), method ="spearman", use = "pairwise.complete.obs")$p.value)
met_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
tryCatch({
if(m!=ncol (B)){
temp_cor = apply (as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method = "spearman")$estimate)
met_intra_cor[,m] = c(rep(NA,m),temp_cor)
temp_p = apply (as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method ="spearman")$p.value)
met_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
met_intra_cor[,m] = rep(NA,m)
met_intra_cor_p[,m] = rep(NA,m)
}
}, error=function(e) {
})
}
}
met_intra_cor_p.adjust = apply(met_intra_cor_p, 2, FUN = function(x){
p.adjust(x,method = "BH")
})
microbes<-matrix(MMData2[MMData2$group=='microbe',]$microandpath)
metabolites<-matrix(paste(MMData1[MMData1$group=='metabolite',]$metaandpath,'-M',sep=" "))
mic_path<-matrix(MMData2[MMData2$group=='mic_path',]$microandpath)
met_path<-matrix(paste(MMData1[MMData1$group=='met_path',]$metaandpath,'-M',sep=" "))
CorrDF <- function(cormat,pmat) {
ut = matrix (TRUE, nrow = nrow (cormat), ncol = ncol (pmat))
data.frame(
from = rownames(cormat)[row(cormat)[ut]],
to = colnames(cormat)[col(cormat)[ut]],
r = (cormat)[ut],
p.adjust = pmat[ut]
)
}
mic_meta_cor_df <- CorrDF(mic_meta_cor,mic_meta_cor_p.adjust)
mic_meta_cor_filtered <- mic_meta_cor_df[which(abs(mic_meta_cor_df$r) > r & mic_meta_cor_df$p.adjust < p_adjust),]
met_intra_cor_df <- CorrDF(met_intra_cor,met_intra_cor_p.adjust)
met_intra_cor_filtered <- met_intra_cor_df[which(abs(met_intra_cor_df$r) > r & met_intra_cor_df$p.adjust < p_adjust),]
mic_intra_cor_df <- CorrDF(mic_intra_cor,mic_intra_cor_p.adjust)
mic_intra_cor_filtered <- mic_intra_cor_df[which(abs(mic_intra_cor_df$r) > r & mic_intra_cor_df$p.adjust < p_adjust),]
mic_meta_cor_filtered <- rbind(mic_meta_cor_filtered,met_intra_cor_filtered)
mic_meta_cor_filtered <- rbind(mic_meta_cor_filtered,mic_intra_cor_filtered)
centrality = "degree"
nodeattrib <- data.frame(nodes = union(mic_meta_cor_filtered$from,mic_meta_cor_filtered$to))
mic_meta_cor_filtered <- mic_meta_cor_filtered
nodeattrib$group <- 0
for (i in as.character(nodeattrib$nodes)){
if (i %in% microbes[,1] == TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Microbe"
}else if(i %in% mic_path[,1] == TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Microbiome_pathway"
}else if(i %in% metabolites[,1] == TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Metabolite"
}else if(i %in% met_path[,1]==TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Metabolome_pathway"
}
}
rownames(nodeattrib) <- nodeattrib$nodes
co_net <- graph_from_data_frame(mic_meta_cor_filtered,directed = F, vertices = nodeattrib)
net_nodes <- rownames(nodeattrib[nodeattrib$group %in% co_net,])
nodes_size = centr_degree(co_net)$res
DS <- as.data.frame(cbind(names(V(co_net)),nodes_size))
colnames(DS) <- c("nodes","size")
DS$size <- as.numeric(DS$size)
DS <- DS[order(DS$size,decreasing = T),]
DS$score <- seq(nrow(DS),1,-1)
DS$group <- 0
for (i in DS$nodes){
if (i %in% microbes[,1]== TRUE){
DS[DS$nodes == i,"group"] <- "Microbe"
}else if(i %in% mic_path[,1] == TRUE){
DS[DS$nodes == i,"group"] <- "Microbiome_pathway"
}else if(i %in% metabolites[,1] == TRUE){
DS[DS$nodes == i,"group"] <- "Metabolite"
}else{
DS[DS$nodes == i,"group"] <- "Metabolome_pathway"
}
}
DS[which(DS$group=="Metabolome_pathway"),]$nodes=gsub(" -M",'',DS[which(DS$group=="Metabolome_pathway"),]$nodes)
DS[which(DS$group=="Metabolite"),]$nodes=gsub(" -M",'',DS[which(DS$group=="Metabolite"),]$nodes)
DS_metabolites <- DS[which(DS$group == "Metabolite"),]
DS_met_pathway <- DS[which(DS$group == "Metabolome_pathway"),]
DS_microbes <- DS[which(DS$group == "Microbe"),]
DS_mic_pathway <- DS[which(DS$group == "Microbiome_pathway"),]
DS_metabolites$score <- seq(nrow(DS_metabolites),1,-1)
DS_metabolites$score <- (DS_metabolites$score - 1)/(nrow(DS_metabolites) - 1) * 2 + 1
DS_met_pathway$score <- seq(nrow(DS_met_pathway),1,-1)
DS_met_pathway$score <- (DS_met_pathway$score - 1)/(nrow(DS_met_pathway) - 1) * 2 + 1
DS_microbes$score <- seq(nrow(DS_microbes),1,-1)
DS_microbes$score <- (DS_microbes$score - 1)/(nrow(DS_microbes) - 1) * 2 + 1
DS_mic_pathway$score <- seq(nrow(DS_mic_pathway),1,-1)
DS_mic_pathway$score <- (DS_mic_pathway$score - 1)/(nrow(DS_mic_pathway) - 1) * 2 + 1
DS<- rbind(DS_metabolites,DS_met_pathway)
DS<-rbind(DS,DS_microbes)
DS<-rbind(DS,DS_mic_pathway)
# groupInfo value ------------------------------------------------------------
group1=groupInfo[which(groupInfo$Group==unique(groupInfo$Group)[1]),]$SampleID
group2=groupInfo[which(groupInfo$Group==unique(groupInfo$Group)[2]),]$SampleID
wil_p.value <- c()
for (i in colnames(B)) {
wil_p.value <- append(wil_p.value, wilcox.test(B[group1,][,i],B[group2,][,i])$p.value)
}
met_log2fc <- c()
for (i in colnames(B)) {
met_log2fc <- append(met_log2fc, log2(mean(B[group2,][,i])/mean(B[group1,][,i])))
}
diff_met <- data.frame(nodes = colnames(B),log2FC=met_log2fc,p.value=wil_p.value)
diff_met<-diff_met[which(diff_met$p.value<0.05),]
wil_p.value <- c()
for (i in colnames(A)) {
wil_p.value <- append(wil_p.value, wilcox.test(as.numeric(A[group1,][,i],A[group2,][,i]))$p.value)
}
mic_log2fc <- c()
for (i in colnames(A)) {
mic_log2fc <- append(mic_log2fc, log2(mean(A[group2,][,i])/mean(A[group1,][,i])))
}
diff_mic<- data.frame(nodes = colnames(A),log2FC=mic_log2fc,p.value=wil_p.value)
diff_mic<-diff_mic[which(diff_mic$p.value<0.05),]
diff_mic_met<-rbind(diff_mic,diff_met)
diff_df<-as.data.frame(diff_mic_met$nodes)
colnames(diff_df)<-"Diff"
mic_meta_cor_filtered$ES <- 1
dfii <- unique(diff_df$Diff)
# n1 <- n2 <- 0
for (i in 1:nrow(mic_meta_cor_filtered)) {
if((mic_meta_cor_filtered$from[i] %in% dfii) & (mic_meta_cor_filtered$to[i] %in% dfii)){
mic_meta_cor_filtered$ES[i] <- 3
# n1 <- n1 + 1
}else{
# n2 <- n2 + 1
mic_meta_cor_filtered$ES[i] <- 1
}
}
from_es <- mic_meta_cor_filtered %>%
group_by(from) %>%
summarise("ES" = sum(ES))
colnames(from_es) <- c("nodes","ES")
to_es <- mic_meta_cor_filtered %>%
group_by(to) %>%
summarise("ES" = sum(ES))
colnames(to_es) <- c("nodes","ES")
mic_met_es <- merge(from_es,to_es,by = "nodes",all = T)
mic_met_es[is.na(mic_met_es)] <- 0
mic_met_es$ES <- mic_met_es$ES.x + mic_met_es$ES.y
mic_met_es <- mic_met_es[,-(2:3)]
ES <- mic_met_es
ES <- ES[order(ES$ES,decreasing = T),]
ES$score <- seq(nrow(ES),1,-1)
ES$group <- 0
ES <- as.data.frame(ES)
for (i in ES$nodes){
if (i %in% microbes[,1] == TRUE){
ES[ES$nodes == i,"group"] <- "Microbe"
}else if(i %in% mic_path[,1] == TRUE){
ES[ES$nodes == i,"group"] <- "Microbiome_pathway"
}else if(i %in% metabolites[,1] == TRUE){
ES[ES$nodes == i,"group"] <- "Metabolite"
}else{
ES[ES$nodes == i,"group"] <- "Metabolome_pathway"
}
}
ES[which(ES$group=="Metabolome_pathway"),]$nodes=gsub(" -M",'',ES[which(ES$group=="Metabolome_pathway"),]$nodes)
ES[which(ES$group=="Metabolite"),]$nodes=gsub(" -M",'',ES[which(ES$group=="Metabolite"),]$nodes)
ES_metabolites <- ES[which(ES$group == "Metabolite"),]
ES_met_pathway <- ES[which(ES$group == "Metabolome_pathway"),]
ES_microbes <- ES[which(ES$group == "Microbe"),]
ES_mic_pathway <- ES[which(ES$group == "Microbiome_pathway"),]
ES_metabolites$score <- seq(nrow(ES_metabolites),1,-1)
ES_metabolites$score <- (ES_metabolites$score - 1)/(nrow(ES_metabolites) - 1) * 2 + 1
ES_met_pathway$score <- seq(nrow(ES_met_pathway),1,-1)
ES_met_pathway$score <- (ES_met_pathway$score - 1)/(nrow(ES_met_pathway) - 1) * 2 + 1
ES_microbes$score <- seq(nrow(ES_microbes),1,-1)
ES_microbes$score <- (ES_microbes$score - 1)/(nrow(ES_microbes) - 1) * 2 + 1
ES_mic_pathway$score <- seq(nrow(ES_mic_pathway),1,-1)
ES_mic_pathway$score <- (ES_mic_pathway$score - 1)/(nrow(ES_mic_pathway) - 1) * 2 + 1
ES<-rbind(ES_metabolites,ES_met_pathway)
ES<-rbind(ES,ES_microbes)
ES<-rbind(ES,ES_mic_pathway)
colnames(B)=gsub(" -M",'',colnames(B))
metabolites_abun <- B[,which(colnames(B) %in% DS_metabolites$nodes)]
metabolites_pathway_abun <- B[,which(colnames(B) %in% DS_met_pathway$nodes )]
AS_metabolites <- apply(metabolites_abun,2,mean)
AS_metabolites <- data.frame(nodes = colnames(metabolites_abun), abundance = AS_metabolites)
AS_met_pathway <- apply(metabolites_pathway_abun,2,mean)
AS_met_pathway <- data.frame(nodes = colnames(metabolites_pathway_abun), abundance = AS_met_pathway)
AS_metabolites <- AS_metabolites[order(AS_metabolites$abundance,decreasing = T),]
AS_metabolites$score <- seq(nrow(AS_metabolites),1,-1)
AS_metabolites$score <- (AS_metabolites$score - 1)/(nrow(AS_metabolites) - 1) * 2 + 1
AS_metabolites$group <- "Metabolite"
AS_met_pathway <- AS_met_pathway[order(AS_met_pathway$abundance,decreasing = T),]
AS_met_pathway$score <- seq(nrow(AS_met_pathway),1,-1)
AS_met_pathway$score <- (AS_met_pathway$score - 1)/(nrow(AS_met_pathway) - 1) * 2 + 1
AS_met_pathway$group <- "MetaBolome_pathway"
microbes_abun <- A[,which(DS_microbes$nodes%in% colnames(A))]
if(length(rownames(DS_microbes))>1){
microbes_abun <- A[,which(colnames(A) %in% DS_microbes$nodes)]
}else{
microbes_abun=as.data.frame(A[,which(colnames(A) %in% DS_microbes$nodes)])
rownames(microbes_abun)=rownames(A)
colnames(microbes_abun)=DS_microbes$nodes
}
if(length(rownames(DS_mic_pathway$nodes))>1){
microbes_pathway_abun <- A[,which(colnames(A) %in% DS_mic_pathway$nodes)]
}else{
microbes_pathway_abun=as.data.frame(A[,which(colnames(A) %in% DS_mic_pathway$nodes)])
rownames(microbes_pathway_abun)=rownames(A)
colnames(microbes_pathway_abun)=DS_mic_pathway$nodes
}
AS_microbes <- apply(microbes_abun,2,mean)
AS_microbes <- data.frame(nodes = colnames(microbes_abun), abundance = AS_microbes)
AS_mic_pathway <- apply(microbes_pathway_abun,2,mean)
AS_mic_pathway <- data.frame(nodes = colnames(microbes_pathway_abun), abundance = AS_mic_pathway)
AS_microbes <- AS_microbes[order(AS_microbes$abundance,decreasing = T),]
AS_microbes$score <- seq(nrow(AS_microbes),1,-1)
AS_microbes$score <- (AS_microbes$score - 1)/(nrow(AS_microbes) - 1) * 2 + 1
AS_microbes$group <- "Microbe"
AS_mic_pathway <- AS_mic_pathway[order(AS_mic_pathway$abundance,decreasing = T),]
AS_mic_pathway$score <- seq(nrow(AS_mic_pathway),1,-1)
AS_mic_pathway$score <- (AS_mic_pathway$score - 1)/(nrow(AS_mic_pathway) - 1) * 2 + 1
AS_mic_pathway$group <- "Microbiome_pathway"
AS<-rbind(AS_metabolites,AS_met_pathway)
AS<-rbind(AS,AS_microbes)
AS<-rbind(AS,AS_mic_pathway)
#####DFS
diff_met$DFS <- sqrt((log10(diff_met$p.value))^2 + (diff_met$log2FC)^2)
diff_mic$DFS <- sqrt((log10(diff_mic$p.value))^2 + (diff_mic$log2FC)^2)
met_diff <- data.frame(nodes = diff_met$nodes, DFS = diff_met$DFS)
mic_diff <- data.frame(nodes = diff_mic$nodes, DFS = diff_mic$DFS)
met_diff$nodes=gsub(' -M','',met_diff$nodes)
DFS_metabolites <- unique(merge(AS_metabolites,met_diff,by="nodes"))
DFS_metabolites$abundance <- DFS_metabolites$DFS
DFS_metabolites <- DFS_metabolites[,-length(DFS_metabolites)]
colnames(DFS_metabolites)[2] <- "DFS"
DFS_metabolites <- DFS_metabolites[order(DFS_metabolites$DFS,decreasing = T),]
DFS_metabolites$score <- seq(nrow(DFS_metabolites),1,-1)
DFS_metabolites$score <- (DFS_metabolites$score - 1)/(nrow(DFS_metabolites) - 1) * 2 + 1
DFS_metabolites$group <- "Metabolite"
DFS_met_pathway <- unique(merge(AS_met_pathway,met_diff,by="nodes"))
DFS_met_pathway$abundance <- DFS_met_pathway$DFS
DFS_met_pathway <- DFS_met_pathway[,-length(DFS_met_pathway)]
colnames(DFS_met_pathway)[2] <- "DFS"
DFS_met_pathway <- DFS_met_pathway[order(DFS_met_pathway$DFS,decreasing = T),]
DFS_met_pathway$score <- seq(nrow(DFS_met_pathway),1,-1)
DFS_met_pathway$score <- (DFS_met_pathway$score - 1)/(nrow(DFS_met_pathway) - 1) * 2 + 1
DFS_met_pathway$group <- "Metabolome_pathway"
DFS_microbes <- unique(merge(AS_microbes,mic_diff,by="nodes"))
DFS_microbes$abundance <- DFS_microbes$DFS
DFS_microbes <- DFS_microbes[,-length(DFS_microbes)]
colnames(DFS_microbes)[2] <- "DFS"
DFS_microbes <- DFS_microbes[order(DFS_microbes$DFS,decreasing = T),]
flag_inf <- is.finite(DFS_microbes$DFS)
flag_inf2 <- is.infinite(DFS_microbes$DFS)
DFS_microbes[flag_inf,]$score <- seq(nrow(DFS_microbes[flag_inf,]),1,-1)
DFS_microbes[flag_inf,]$score <- (DFS_microbes[flag_inf,]$score - 1)/(nrow(DFS_microbes[flag_inf,]) - 1) * 2 + 1
DFS_microbes$group <- "Microbe"
DFS_mic_pathway <- unique(merge(AS_mic_pathway,mic_diff,by="nodes"))
DFS_mic_pathway$abundance <- DFS_mic_pathway$DFS
DFS_mic_pathway <- DFS_mic_pathway[,-length(DFS_mic_pathway)]
colnames(DFS_mic_pathway)[2] <- "DFS"
DFS_mic_pathway <- DFS_mic_pathway[order(DFS_mic_pathway$DFS,decreasing = T),]
flag_inf <- is.finite(DFS_mic_pathway$DFS)
flag_inf2 <- is.infinite(DFS_mic_pathway$DFS)
DFS_mic_pathway[flag_inf,]$score <- seq(nrow(DFS_mic_pathway[flag_inf,]),1,-1)
DFS_mic_pathway[flag_inf,]$score <- (DFS_mic_pathway[flag_inf,]$score - 1)/(nrow(DFS_mic_pathway[flag_inf,]) - 1) * 2 + 1
DFS_mic_pathway$group <- "Microbiome_pathway"
DFS<-rbind(DFS_metabolites,DFS_met_pathway)
DFS<-rbind(DFS,DFS_microbes)
DFS<-rbind(DFS,DFS_mic_pathway)
NS_metabolites <- data.frame(nodes = DS_metabolites$nodes,ns = (nDFS*DFS_metabolites$score + nDS*DS_metabolites$score + nES*ES_metabolites$score + nAS*AS_metabolites$score),group = DS_metabolites$group)
NS_met_pathway <- data.frame(nodes = DS_met_pathway$nodes,ns = (nDFS*DFS_met_pathway$score + nDS*DS_met_pathway$score + nES*ES_met_pathway$score + nAS*AS_met_pathway$score),group = DS_met_pathway$group)
NS_microbes <- data.frame(nodes = DS_microbes$nodes,ns = (nDFS*DFS_microbes$score + nDS*DS_microbes$score + nES*ES_microbes$score + nAS*AS_microbes$score),group = DS_microbes$group)
NS_mic_pathway <- data.frame(nodes = DS_mic_pathway$nodes,ns = (nDFS*DFS_mic_pathway$score + nDS*DS_mic_pathway$score + nES*ES_mic_pathway$score + nAS*AS_mic_pathway$score),group = DS_mic_pathway$group)
NS_metabolites <- NS_metabolites[order(NS_metabolites$ns,decreasing = T),]
NS_met_pathway <- NS_met_pathway[order(NS_met_pathway$ns,decreasing = T),]
NS_microbes <- NS_microbes[order(NS_microbes$ns,decreasing = T),]
NS_mic_pathway <- NS_mic_pathway[order(NS_mic_pathway$ns,decreasing = T),]
for (i in 1:nrow(NS_mic_pathway)) {
NS_mic_pathway[i,1] <- strsplit(NS_mic_pathway[i,1], " \\[")[[1]][1]
}
Node_Score<-rbind(NS_metabolites,NS_met_pathway)
Node_Score<-rbind(Node_Score,NS_microbes)
Node_Score<-rbind(Node_Score,NS_mic_pathway)
NScore=Node_Score[Node_Score$ns>NS,]
mic_meta_cor_filtered$from=gsub(' -M','',mic_meta_cor_filtered$from)
mic_meta_cor_filtered$to=gsub(' -M','',mic_meta_cor_filtered$to)
colnames(mic_meta_cor)=gsub(' -M','',colnames(mic_meta_cor))
colnames(mic_meta_cor_p.adjust)=gsub(' -M','',colnames(mic_meta_cor_p.adjust))
mic_meta_cor=as.data.frame(mic_meta_cor)
mic_meta_cor_p.adjust=as.data.frame(mic_meta_cor_p.adjust)
IIS_res=list("Correlation analysis results"=mic_meta_cor_filtered,'mic_meta_cor'=mic_meta_cor,'mic_meta_cor_p.adjust'=mic_meta_cor_p.adjust,"Dgree_Score"=DS,"Abundance_Score"=AS,"Edge_Score"=ES,"Diff_Score"=DFS,"NScore"=NScore,'Node_Score'=Node_Score)
return(IIS_res)
}
Try the BiOFI package in your browser
Any scripts or data that you put into this service are public.
BiOFI documentation built on Sept. 22, 2022, 9:07 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions IIS
Documented in IIS
#' @title IIS
#' @description
#' A function is used to calculates and identify the key nodes,the formula of IIS is:
#' IIS= nDFS x DFS + nDS x DS + nES x ES + nAS x AS;
#' nDFS,nDS,nES,nAS indict the weight value of DFS,DS,ES and AS, and nDFS + nDS + nES + nAS = 1;
#' DFS means Difference Score;
#' DS means Degree Score;
#' ES means Edge Score;
#' AS means Abundance Score.
#'
#' @param microApath A data frame including microbes and predicted KEGG pathways.
#' @param metaApath A data frame including metabolites with compound ID and predicted KEGG pathways from Meta2pathway function.
#' @param conf A data frame including confounders. Default value is NULL.
#' @param groupInfo A data frame including sample grouping information
#' @param nDFS A value indicated the weight of difference Score in the IIS formula. Default value is 0.4.
#' @param nDS A value indicated the weight of degree score in the IIS formula. Default value is 0.3.
#' @param nES A value indicated the weight of edge score in the IIS formula. Default value is 0.2.
#' @param nAS A value indicated the weight of abundance score in the IIS formula. Default value is 0.1.
#' @param NS A threshold value of IIS to identify the key nodes. Default value is 2.5.
#' @param r A threshold value of correlation coefficient to construct the network. Default value is 0.5.
#' @param p_adjust A threshold value of correlation adjust p value to construct the network. Default value is 0.05.
#' @return IIS_res A list including IIS and other sub-score results.
#' @export IIS
#' @importFrom igraph graph_from_data_frame
#' @importFrom dplyr group_by
#' @importFrom dplyr summarise
#' @import ggrepel
#' @import ppcor
#' @importFrom utils write.csv
#' @importFrom stats cor.test
#' @importFrom stats p.adjust
#' @importFrom stats wilcox.test
#' @examples
#' IIScore <- IIS(microApath = micro.eg, metaApath = metabo.eg,
#' conf = confounder.eg, groupInfo = groupInfo.eg)
IIS<-function(microApath, metaApath, conf = NULL, groupInfo = NULL, nDFS = 0.4,
nDS = 0.3, nES = 0.2, nAS = 0.1, NS = 2.5, r = 0.5, p_adjust = 0.05){
from = NULL;to = NULL
A=microApath
B=metaApath
mmApath=get0("sysdata", envir = asNamespace("BiOFI"))$mm4path
MMData1=data.frame(metaandpath=colnames(B),group=0)
for(i in 1:length(colnames(B))){
if(MMData1$metaandpath[i]%in% unique(mmApath$meta_path)){
MMData1$group[i]='met_path'
}else{
MMData1$group[i]='metabolite'
}
}
if(!'met_path'%in% unique(MMData1$group)){
stop('please tranfer metabolites to pathway using Meta2pathway function')
}
MMData2=data.frame(microandpath=colnames(A),group=0)
for(i in 1:length(colnames(A))){
if(substr(MMData2$microandpath[i],1,3)=='k__'){
MMData2$group[i]='microbe'
}else{MMData2$group[i]='mic_path'}
}
if(length(rownames(groupInfo))==0){
stop('Group information is lack, please check!')
}
colnames(B)=paste(colnames(B),'-M',sep=" ")
mic_meta_cor = matrix(NA, nrow = ncol (A), ncol = ncol (B))
mic_intra_cor=matrix(NA, nrow = ncol (A), ncol = ncol (A))
met_intra_cor=matrix(NA, nrow = ncol (B), ncol = ncol (B))
rownames (mic_meta_cor) = colnames (A)
colnames (mic_meta_cor) = colnames (B)
rownames (mic_intra_cor) = colnames (A)
colnames (mic_intra_cor) = colnames (A)
rownames (met_intra_cor) = colnames (B)
colnames (met_intra_cor) = colnames (B)
mic_meta_cor_p = mic_meta_cor
mic_meta_cor_p.adjust = mic_meta_cor
mic_intra_cor_p=mic_intra_cor
mic_intra_cor_p.adjust=mic_intra_cor
met_intra_cor_p=met_intra_cor
met_intra_cor_p.adjust=met_intra_cor
for (m in colnames (B)) {
if(length(rownames(conf))==0){
mic_meta_cor[,m] = apply (A, MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x),as.numeric(B[,m]), method = "spearman",use = "pairwise.complete.obs")$estimate)
mic_meta_cor_p[,m] = apply (A, MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x),as.numeric(B[,m]), method ="spearman", use = "pairwise.complete.obs")$p.value)
}else{
tryCatch({
mic_meta_cor[,m] = apply (A, MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method = "spearman")$estimate)
mic_meta_cor_p[,m] = apply (A, MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method ="spearman")$p.value)
}, error=function(e) {
})
}
}
mic_meta_cor_p.adjust = apply(mic_meta_cor_p, 2, FUN = function(x){
p.adjust(x,method = "BH")})
del_var <- c()
for (m in 1:ncol (A)) {
del_var <- append(del_var,m)
if(length(rownames(conf))==0){
temp_cor = apply(as.data.frame(A[,-del_var]), MARGIN = 2,FUN = function (x)
cor.test (as.numeric(x), as.numeric(A[,m]), method = "spearman", use = "pairwise.complete.obs")$estimate)
mic_intra_cor [ , m] = c(rep(NA,m),temp_cor)
temp_p = apply(as.data.frame(A[,-del_var]), MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x), as.numeric(A[,m]), method ="spearman", use = "pairwise.complete.obs")$p.value)
mic_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
tryCatch({
if(m!=ncol (A)){
temp_cor = apply (as.data.frame(A[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(A[,m]), conf, method = "spearman")$estimate)
mic_intra_cor[,m] = c(rep(NA,m),temp_cor)
temp_p = apply (as.data.frame(A[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(A[,m]), conf, method ="spearman")$p.value)
mic_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
mic_intra_cor[,m] = rep(NA,m)
mic_intra_cor_p[,m] = rep(NA,m)
}
}, error=function(e) {
})
}
}
mic_intra_cor_p.adjust = apply(mic_intra_cor_p, 2, FUN = function(x){
p.adjust(x,method = "BH")})
del_var <- c()
for (m in 1:ncol (B)) {
del_var <- append(del_var,m)
if(length(rownames(conf))==0){
temp_cor = apply(as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x),as.numeric(B[,m]), method = "spearman", use = "pairwise.complete.obs")$estimate)
met_intra_cor [ , m] = c(rep(NA,m),temp_cor)
temp_p = apply(as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
cor.test (as.numeric(x), as.numeric(B[,m]), method ="spearman", use = "pairwise.complete.obs")$p.value)
met_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
tryCatch({
if(m!=ncol (B)){
temp_cor = apply (as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method = "spearman")$estimate)
met_intra_cor[,m] = c(rep(NA,m),temp_cor)
temp_p = apply (as.data.frame(B[,-del_var]), MARGIN = 2, FUN = function (x)
ppcor::pcor.test (as.numeric(x),as.numeric(B[,m]), conf, method ="spearman")$p.value)
met_intra_cor_p[,m] = c(rep(NA,m),temp_p)
}else{
met_intra_cor[,m] = rep(NA,m)
met_intra_cor_p[,m] = rep(NA,m)
}
}, error=function(e) {
})
}
}
met_intra_cor_p.adjust = apply(met_intra_cor_p, 2, FUN = function(x){
p.adjust(x,method = "BH")
})
microbes<-matrix(MMData2[MMData2$group=='microbe',]$microandpath)
metabolites<-matrix(paste(MMData1[MMData1$group=='metabolite',]$metaandpath,'-M',sep=" "))
mic_path<-matrix(MMData2[MMData2$group=='mic_path',]$microandpath)
met_path<-matrix(paste(MMData1[MMData1$group=='met_path',]$metaandpath,'-M',sep=" "))
CorrDF <- function(cormat,pmat) {
ut = matrix (TRUE, nrow = nrow (cormat), ncol = ncol (pmat))
data.frame(
from = rownames(cormat)[row(cormat)[ut]],
to = colnames(cormat)[col(cormat)[ut]],
r = (cormat)[ut],
p.adjust = pmat[ut]
)
}
mic_meta_cor_df <- CorrDF(mic_meta_cor,mic_meta_cor_p.adjust)
mic_meta_cor_filtered <- mic_meta_cor_df[which(abs(mic_meta_cor_df$r) > r & mic_meta_cor_df$p.adjust < p_adjust),]
met_intra_cor_df <- CorrDF(met_intra_cor,met_intra_cor_p.adjust)
met_intra_cor_filtered <- met_intra_cor_df[which(abs(met_intra_cor_df$r) > r & met_intra_cor_df$p.adjust < p_adjust),]
mic_intra_cor_df <- CorrDF(mic_intra_cor,mic_intra_cor_p.adjust)
mic_intra_cor_filtered <- mic_intra_cor_df[which(abs(mic_intra_cor_df$r) > r & mic_intra_cor_df$p.adjust < p_adjust),]
mic_meta_cor_filtered <- rbind(mic_meta_cor_filtered,met_intra_cor_filtered)
mic_meta_cor_filtered <- rbind(mic_meta_cor_filtered,mic_intra_cor_filtered)
centrality = "degree"
nodeattrib <- data.frame(nodes = union(mic_meta_cor_filtered$from,mic_meta_cor_filtered$to))
mic_meta_cor_filtered <- mic_meta_cor_filtered
nodeattrib$group <- 0
for (i in as.character(nodeattrib$nodes)){
if (i %in% microbes[,1] == TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Microbe"
}else if(i %in% mic_path[,1] == TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Microbiome_pathway"
}else if(i %in% metabolites[,1] == TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Metabolite"
}else if(i %in% met_path[,1]==TRUE){
nodeattrib[nodeattrib$nodes == i,"group"] <- "Metabolome_pathway"
}
}
rownames(nodeattrib) <- nodeattrib$nodes
co_net <- graph_from_data_frame(mic_meta_cor_filtered,directed = F, vertices = nodeattrib)
net_nodes <- rownames(nodeattrib[nodeattrib$group %in% co_net,])
nodes_size = centr_degree(co_net)$res
DS <- as.data.frame(cbind(names(V(co_net)),nodes_size))
colnames(DS) <- c("nodes","size")
DS$size <- as.numeric(DS$size)
DS <- DS[order(DS$size,decreasing = T),]
DS$score <- seq(nrow(DS),1,-1)
DS$group <- 0
for (i in DS$nodes){
if (i %in% microbes[,1]== TRUE){
DS[DS$nodes == i,"group"] <- "Microbe"
}else if(i %in% mic_path[,1] == TRUE){
DS[DS$nodes == i,"group"] <- "Microbiome_pathway"
}else if(i %in% metabolites[,1] == TRUE){
DS[DS$nodes == i,"group"] <- "Metabolite"
}else{
DS[DS$nodes == i,"group"] <- "Metabolome_pathway"
}
}
DS[which(DS$group=="Metabolome_pathway"),]$nodes=gsub(" -M",'',DS[which(DS$group=="Metabolome_pathway"),]$nodes)
DS[which(DS$group=="Metabolite"),]$nodes=gsub(" -M",'',DS[which(DS$group=="Metabolite"),]$nodes)
DS_metabolites <- DS[which(DS$group == "Metabolite"),]
DS_met_pathway <- DS[which(DS$group == "Metabolome_pathway"),]
DS_microbes <- DS[which(DS$group == "Microbe"),]
DS_mic_pathway <- DS[which(DS$group == "Microbiome_pathway"),]
DS_metabolites$score <- seq(nrow(DS_metabolites),1,-1)
DS_metabolites$score <- (DS_metabolites$score - 1)/(nrow(DS_metabolites) - 1) * 2 + 1
DS_met_pathway$score <- seq(nrow(DS_met_pathway),1,-1)
DS_met_pathway$score <- (DS_met_pathway$score - 1)/(nrow(DS_met_pathway) - 1) * 2 + 1
DS_microbes$score <- seq(nrow(DS_microbes),1,-1)
DS_microbes$score <- (DS_microbes$score - 1)/(nrow(DS_microbes) - 1) * 2 + 1
DS_mic_pathway$score <- seq(nrow(DS_mic_pathway),1,-1)
DS_mic_pathway$score <- (DS_mic_pathway$score - 1)/(nrow(DS_mic_pathway) - 1) * 2 + 1
DS<- rbind(DS_metabolites,DS_met_pathway)
DS<-rbind(DS,DS_microbes)
DS<-rbind(DS,DS_mic_pathway)
# groupInfo value ------------------------------------------------------------
group1=groupInfo[which(groupInfo$Group==unique(groupInfo$Group)[1]),]$SampleID
group2=groupInfo[which(groupInfo$Group==unique(groupInfo$Group)[2]),]$SampleID
wil_p.value <- c()
for (i in colnames(B)) {
wil_p.value <- append(wil_p.value, wilcox.test(B[group1,][,i],B[group2,][,i])$p.value)
}
met_log2fc <- c()
for (i in colnames(B)) {
met_log2fc <- append(met_log2fc, log2(mean(B[group2,][,i])/mean(B[group1,][,i])))
}
diff_met <- data.frame(nodes = colnames(B),log2FC=met_log2fc,p.value=wil_p.value)
diff_met<-diff_met[which(diff_met$p.value<0.05),]
wil_p.value <- c()
for (i in colnames(A)) {
wil_p.value <- append(wil_p.value, wilcox.test(as.numeric(A[group1,][,i],A[group2,][,i]))$p.value)
}
mic_log2fc <- c()
for (i in colnames(A)) {
mic_log2fc <- append(mic_log2fc, log2(mean(A[group2,][,i])/mean(A[group1,][,i])))
}
diff_mic<- data.frame(nodes = colnames(A),log2FC=mic_log2fc,p.value=wil_p.value)
diff_mic<-diff_mic[which(diff_mic$p.value<0.05),]
diff_mic_met<-rbind(diff_mic,diff_met)
diff_df<-as.data.frame(diff_mic_met$nodes)
colnames(diff_df)<-"Diff"
mic_meta_cor_filtered$ES <- 1
dfii <- unique(diff_df$Diff)
# n1 <- n2 <- 0
for (i in 1:nrow(mic_meta_cor_filtered)) {
if((mic_meta_cor_filtered$from[i] %in% dfii) & (mic_meta_cor_filtered$to[i] %in% dfii)){
mic_meta_cor_filtered$ES[i] <- 3
# n1 <- n1 + 1
}else{
# n2 <- n2 + 1
mic_meta_cor_filtered$ES[i] <- 1
}
}
from_es <- mic_meta_cor_filtered %>%
group_by(from) %>%
summarise("ES" = sum(ES))
colnames(from_es) <- c("nodes","ES")
to_es <- mic_meta_cor_filtered %>%
group_by(to) %>%
summarise("ES" = sum(ES))
colnames(to_es) <- c("nodes","ES")
mic_met_es <- merge(from_es,to_es,by = "nodes",all = T)
mic_met_es[is.na(mic_met_es)] <- 0
mic_met_es$ES <- mic_met_es$ES.x + mic_met_es$ES.y
mic_met_es <- mic_met_es[,-(2:3)]
ES <- mic_met_es
ES <- ES[order(ES$ES,decreasing = T),]
ES$score <- seq(nrow(ES),1,-1)
ES$group <- 0
ES <- as.data.frame(ES)
for (i in ES$nodes){
if (i %in% microbes[,1] == TRUE){
ES[ES$nodes == i,"group"] <- "Microbe"
}else if(i %in% mic_path[,1] == TRUE){
ES[ES$nodes == i,"group"] <- "Microbiome_pathway"
}else if(i %in% metabolites[,1] == TRUE){
ES[ES$nodes == i,"group"] <- "Metabolite"
}else{
ES[ES$nodes == i,"group"] <- "Metabolome_pathway"
}
}
ES[which(ES$group=="Metabolome_pathway"),]$nodes=gsub(" -M",'',ES[which(ES$group=="Metabolome_pathway"),]$nodes)
ES[which(ES$group=="Metabolite"),]$nodes=gsub(" -M",'',ES[which(ES$group=="Metabolite"),]$nodes)
ES_metabolites <- ES[which(ES$group == "Metabolite"),]
ES_met_pathway <- ES[which(ES$group == "Metabolome_pathway"),]
ES_microbes <- ES[which(ES$group == "Microbe"),]
ES_mic_pathway <- ES[which(ES$group == "Microbiome_pathway"),]
ES_metabolites$score <- seq(nrow(ES_metabolites),1,-1)
ES_metabolites$score <- (ES_metabolites$score - 1)/(nrow(ES_metabolites) - 1) * 2 + 1
ES_met_pathway$score <- seq(nrow(ES_met_pathway),1,-1)
ES_met_pathway$score <- (ES_met_pathway$score - 1)/(nrow(ES_met_pathway) - 1) * 2 + 1
ES_microbes$score <- seq(nrow(ES_microbes),1,-1)
ES_microbes$score <- (ES_microbes$score - 1)/(nrow(ES_microbes) - 1) * 2 + 1
ES_mic_pathway$score <- seq(nrow(ES_mic_pathway),1,-1)
ES_mic_pathway$score <- (ES_mic_pathway$score - 1)/(nrow(ES_mic_pathway) - 1) * 2 + 1
ES<-rbind(ES_metabolites,ES_met_pathway)
ES<-rbind(ES,ES_microbes)
ES<-rbind(ES,ES_mic_pathway)
colnames(B)=gsub(" -M",'',colnames(B))
metabolites_abun <- B[,which(colnames(B) %in% DS_metabolites$nodes)]
metabolites_pathway_abun <- B[,which(colnames(B) %in% DS_met_pathway$nodes )]
AS_metabolites <- apply(metabolites_abun,2,mean)
AS_metabolites <- data.frame(nodes = colnames(metabolites_abun), abundance = AS_metabolites)
AS_met_pathway <- apply(metabolites_pathway_abun,2,mean)
AS_met_pathway <- data.frame(nodes = colnames(metabolites_pathway_abun), abundance = AS_met_pathway)
AS_metabolites <- AS_metabolites[order(AS_metabolites$abundance,decreasing = T),]
AS_metabolites$score <- seq(nrow(AS_metabolites),1,-1)
AS_metabolites$score <- (AS_metabolites$score - 1)/(nrow(AS_metabolites) - 1) * 2 + 1
AS_metabolites$group <- "Metabolite"
AS_met_pathway <- AS_met_pathway[order(AS_met_pathway$abundance,decreasing = T),]
AS_met_pathway$score <- seq(nrow(AS_met_pathway),1,-1)
AS_met_pathway$score <- (AS_met_pathway$score - 1)/(nrow(AS_met_pathway) - 1) * 2 + 1
AS_met_pathway$group <- "MetaBolome_pathway"
microbes_abun <- A[,which(DS_microbes$nodes%in% colnames(A))]
if(length(rownames(DS_microbes))>1){
microbes_abun <- A[,which(colnames(A) %in% DS_microbes$nodes)]
}else{
microbes_abun=as.data.frame(A[,which(colnames(A) %in% DS_microbes$nodes)])
rownames(microbes_abun)=rownames(A)
colnames(microbes_abun)=DS_microbes$nodes
}
if(length(rownames(DS_mic_pathway$nodes))>1){
microbes_pathway_abun <- A[,which(colnames(A) %in% DS_mic_pathway$nodes)]
}else{
microbes_pathway_abun=as.data.frame(A[,which(colnames(A) %in% DS_mic_pathway$nodes)])
rownames(microbes_pathway_abun)=rownames(A)
colnames(microbes_pathway_abun)=DS_mic_pathway$nodes
}
AS_microbes <- apply(microbes_abun,2,mean)
AS_microbes <- data.frame(nodes = colnames(microbes_abun), abundance = AS_microbes)
AS_mic_pathway <- apply(microbes_pathway_abun,2,mean)
AS_mic_pathway <- data.frame(nodes = colnames(microbes_pathway_abun), abundance = AS_mic_pathway)
AS_microbes <- AS_microbes[order(AS_microbes$abundance,decreasing = T),]
AS_microbes$score <- seq(nrow(AS_microbes),1,-1)
AS_microbes$score <- (AS_microbes$score - 1)/(nrow(AS_microbes) - 1) * 2 + 1
AS_microbes$group <- "Microbe"
AS_mic_pathway <- AS_mic_pathway[order(AS_mic_pathway$abundance,decreasing = T),]
AS_mic_pathway$score <- seq(nrow(AS_mic_pathway),1,-1)
AS_mic_pathway$score <- (AS_mic_pathway$score - 1)/(nrow(AS_mic_pathway) - 1) * 2 + 1
AS_mic_pathway$group <- "Microbiome_pathway"
AS<-rbind(AS_metabolites,AS_met_pathway)
AS<-rbind(AS,AS_microbes)
AS<-rbind(AS,AS_mic_pathway)
#####DFS
diff_met$DFS <- sqrt((log10(diff_met$p.value))^2 + (diff_met$log2FC)^2)
diff_mic$DFS <- sqrt((log10(diff_mic$p.value))^2 + (diff_mic$log2FC)^2)
met_diff <- data.frame(nodes = diff_met$nodes, DFS = diff_met$DFS)
mic_diff <- data.frame(nodes = diff_mic$nodes, DFS = diff_mic$DFS)
met_diff$nodes=gsub(' -M','',met_diff$nodes)
DFS_metabolites <- unique(merge(AS_metabolites,met_diff,by="nodes"))
DFS_metabolites$abundance <- DFS_metabolites$DFS
DFS_metabolites <- DFS_metabolites[,-length(DFS_metabolites)]
colnames(DFS_metabolites)[2] <- "DFS"
DFS_metabolites <- DFS_metabolites[order(DFS_metabolites$DFS,decreasing = T),]
DFS_metabolites$score <- seq(nrow(DFS_metabolites),1,-1)
DFS_metabolites$score <- (DFS_metabolites$score - 1)/(nrow(DFS_metabolites) - 1) * 2 + 1
DFS_metabolites$group <- "Metabolite"
DFS_met_pathway <- unique(merge(AS_met_pathway,met_diff,by="nodes"))
DFS_met_pathway$abundance <- DFS_met_pathway$DFS
DFS_met_pathway <- DFS_met_pathway[,-length(DFS_met_pathway)]
colnames(DFS_met_pathway)[2] <- "DFS"
DFS_met_pathway <- DFS_met_pathway[order(DFS_met_pathway$DFS,decreasing = T),]
DFS_met_pathway$score <- seq(nrow(DFS_met_pathway),1,-1)
DFS_met_pathway$score <- (DFS_met_pathway$score - 1)/(nrow(DFS_met_pathway) - 1) * 2 + 1
DFS_met_pathway$group <- "Metabolome_pathway"
DFS_microbes <- unique(merge(AS_microbes,mic_diff,by="nodes"))
DFS_microbes$abundance <- DFS_microbes$DFS
DFS_microbes <- DFS_microbes[,-length(DFS_microbes)]
colnames(DFS_microbes)[2] <- "DFS"
DFS_microbes <- DFS_microbes[order(DFS_microbes$DFS,decreasing = T),]
flag_inf <- is.finite(DFS_microbes$DFS)
flag_inf2 <- is.infinite(DFS_microbes$DFS)
DFS_microbes[flag_inf,]$score <- seq(nrow(DFS_microbes[flag_inf,]),1,-1)
DFS_microbes[flag_inf,]$score <- (DFS_microbes[flag_inf,]$score - 1)/(nrow(DFS_microbes[flag_inf,]) - 1) * 2 + 1
DFS_microbes$group <- "Microbe"
DFS_mic_pathway <- unique(merge(AS_mic_pathway,mic_diff,by="nodes"))
DFS_mic_pathway$abundance <- DFS_mic_pathway$DFS
DFS_mic_pathway <- DFS_mic_pathway[,-length(DFS_mic_pathway)]
colnames(DFS_mic_pathway)[2] <- "DFS"
DFS_mic_pathway <- DFS_mic_pathway[order(DFS_mic_pathway$DFS,decreasing = T),]
flag_inf <- is.finite(DFS_mic_pathway$DFS)
flag_inf2 <- is.infinite(DFS_mic_pathway$DFS)
DFS_mic_pathway[flag_inf,]$score <- seq(nrow(DFS_mic_pathway[flag_inf,]),1,-1)
DFS_mic_pathway[flag_inf,]$score <- (DFS_mic_pathway[flag_inf,]$score - 1)/(nrow(DFS_mic_pathway[flag_inf,]) - 1) * 2 + 1
DFS_mic_pathway$group <- "Microbiome_pathway"
DFS<-rbind(DFS_metabolites,DFS_met_pathway)
DFS<-rbind(DFS,DFS_microbes)
DFS<-rbind(DFS,DFS_mic_pathway)
NS_metabolites <- data.frame(nodes = DS_metabolites$nodes,ns = (nDFS*DFS_metabolites$score + nDS*DS_metabolites$score + nES*ES_metabolites$score + nAS*AS_metabolites$score),group = DS_metabolites$group)
NS_met_pathway <- data.frame(nodes = DS_met_pathway$nodes,ns = (nDFS*DFS_met_pathway$score + nDS*DS_met_pathway$score + nES*ES_met_pathway$score + nAS*AS_met_pathway$score),group = DS_met_pathway$group)
NS_microbes <- data.frame(nodes = DS_microbes$nodes,ns = (nDFS*DFS_microbes$score + nDS*DS_microbes$score + nES*ES_microbes$score + nAS*AS_microbes$score),group = DS_microbes$group)
NS_mic_pathway <- data.frame(nodes = DS_mic_pathway$nodes,ns = (nDFS*DFS_mic_pathway$score + nDS*DS_mic_pathway$score + nES*ES_mic_pathway$score + nAS*AS_mic_pathway$score),group = DS_mic_pathway$group)
NS_metabolites <- NS_metabolites[order(NS_metabolites$ns,decreasing = T),]
NS_met_pathway <- NS_met_pathway[order(NS_met_pathway$ns,decreasing = T),]
NS_microbes <- NS_microbes[order(NS_microbes$ns,decreasing = T),]
NS_mic_pathway <- NS_mic_pathway[order(NS_mic_pathway$ns,decreasing = T),]
for (i in 1:nrow(NS_mic_pathway)) {
NS_mic_pathway[i,1] <- strsplit(NS_mic_pathway[i,1], " \\[")[[1]][1]
}
Node_Score<-rbind(NS_metabolites,NS_met_pathway)
Node_Score<-rbind(Node_Score,NS_microbes)
Node_Score<-rbind(Node_Score,NS_mic_pathway)
NScore=Node_Score[Node_Score$ns>NS,]
mic_meta_cor_filtered$from=gsub(' -M','',mic_meta_cor_filtered$from)
mic_meta_cor_filtered$to=gsub(' -M','',mic_meta_cor_filtered$to)
colnames(mic_meta_cor)=gsub(' -M','',colnames(mic_meta_cor))
colnames(mic_meta_cor_p.adjust)=gsub(' -M','',colnames(mic_meta_cor_p.adjust))
mic_meta_cor=as.data.frame(mic_meta_cor)
mic_meta_cor_p.adjust=as.data.frame(mic_meta_cor_p.adjust)
IIS_res=list("Correlation analysis results"=mic_meta_cor_filtered,'mic_meta_cor'=mic_meta_cor,'mic_meta_cor_p.adjust'=mic_meta_cor_p.adjust,"Dgree_Score"=DS,"Abundance_Score"=AS,"Edge_Score"=ES,"Diff_Score"=DFS,"NScore"=NScore,'Node_Score'=Node_Score)
return(IIS_res)
}
Try the BiOFI package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.