KEGG API download of pathways and 'Offline' MR BACOn CAD-serum analyses

Offline/R code/MR serum data offline processing functions v1.R

@@ -0,0 +1,10 @@
+#### functions for offline processing
+
+serum_metabolite <- function(metabolite) {
+  path <- "data/metab_info/"
+  tissue <- "serum"
+  path <- paste(path, tissue, "_map.txt", sep = "")
+  f <- read.delim(path, header = T, stringsAsFactors = F)
+  met <- f$metid[f$metabolite == metabolite]
+  return(met)
+}

Offline/R code/MR serum data offline processing v1.R

@@ -0,0 +1,68 @@
+### offline MR processing for serum
+
+rm(list=ls())
+
+library(devtools)
+install_github("MRCIEU/TwoSampleMR")
+library(TwoSampleMR)
+
+setwd("C:/Users/cavin/Desktop/MR_BACOn/Offline/R code/")
+
+source("MR serum data offline processing functions v1.R")
+
+path <- "../../ShinyApp/data/mr_data/pre-clumped/serum/"
+
+metab_info <- read.delim("../../ShinyApp/data/metab_info/serum_map.txt", header = T, stringsAsFactors = F)
+
+#### get all the files and names in metab info
+mfiles <- list.files("../../ShinyApp/data/mr_data/pre-clumped/serum/")
+mnames <- gsub("serum_","",mfiles)
+mnames <- gsub("\\.txt","",mnames)
+
+metab_info <- subset(metab_info, metid%in%mnames)
+
+pvalue <- 0.05
+
+out.matrix <- c()
+out.single <- c()
+for(i in c(1:nrow(metab_info)))
+{
+  print(paste("Doing metab",i,"out of",nrow(metab_info)))
+  metab <- metab_info$metabolite[i]
+  mid <- metab_info$metid[i]
+  f <- mfiles[grep(mid,mfiles)]
+
+  expos_data <- read.delim(paste0(path,f), header = T, stringsAsFactors = F)
+  outcome_read <- tryCatch(read_outcome_data(filename = "../../ShinyApp/data/mr_data/cad_p05.txt", snps = expos_data$SNP, sep = "\t"),error=function(e){return(1)})
+
+  if(!is.null(dim(outcome_read)))
+  {
+    harmed_data <- harmonise_data(exposure_dat = expos_data, outcome_dat = outcome_read)
+    ### forcing mr_keep to TRUE forces keeping in all SNPs even those which are ambiguous; 
+    ###   use only if sure about strand alignment, otherwise comment out
+    harmed_data$mr_keep <- TRUE 
+
+    mr.result <- mr(harmed_data)
+    mr.result$outcome <- "CAD"
+    mr.result$exposure <- metab
+    mr.result$lci <- mr.result$b - qnorm(0.975,0,1)*mr.result$se
+    mr.result$uci <- mr.result$b + qnorm(0.975,0,1)*mr.result$se
+    mr.result <- mr.result[,-which(names(mr.result)%in%c("id.exposure","id.outcome"))]
+
+    out.matrix <- rbind(out.matrix, mr.result)
+
+    mr.single <- mr_singlesnp(harmed_data)
+    mr.single$outcome <- "CAD"
+    mr.single$exposure <- metab
+    mr.single$lci <- mr.single$b - qnorm(0.975,0,1)*mr.single$se
+    mr.single$uci <- mr.single$b + qnorm(0.975,0,1)*mr.single$se
+    mr.single <- mr.single[,-which(names(mr.single)%in%c("id.exposure","id.outcome","samplesize"))]
+
+    out.single <- rbind(out.single, mr.single)
+  }
+
+}
+
+write.table(out.matrix, file="../Offline Results/CAD serum metablites MR results.csv", col.names=TRUE, row.names=FALSE, quote=FALSE, sep=",")
+write.table(out.single, file="../Offline Results/CAD serum metablites MR results single SNP.csv", col.names=TRUE, row.names=FALSE, quote=FALSE, sep=",")
+