R/pathinfo.R

In metabolomicsworkbench/MetENP: Metabolite enrichment analysis and their associated enriched pathways

#'Provides pathway or orthology information from reaction data of significant metabolite

#'@param res dataframe having metabolite information along with reaction information as obtained from rxninfo 

#'@param col_name "PATHWAY" or "ORTHOLOGY"

#'@importFrom KEGGREST keggGet

#'@importFrom tidyr uncount

#'@export

#'@examples

#'res_path= pathinfo(res, 'PATHWAY') ###rxn pathway



pathinfo = function(res, col_name)

{

  

  query = as.vector(res$rxn)

  query = unique(query)

  query = query[!is.na(query)]; # remove NAs

  query = unique(query[query != ""]) # exclude empty

  ### pass the argument in list of 10s since keggrest takes 10 inputs

  query_split = split(query,  ceiling(seq_along(query)/10))

 

  

  ### kegg info

  info = llply(query_split, function(x)keggGet(x))

  unlist_info <- unlist(info, recursive = F)

  extract_info <- lapply(unlist_info, '[', c("ENTRY","NAME","PATHWAY","RCLASS","ORTHOLOGY","DEFINITION","EQUATION","ENZYME"))

  dd=do.call(rbind, extract_info)

  df = as.data.frame(dd)

  ## BUG FIX; Added by Sumana Srinivasan 0n 04042022

  ### named all columns to be consistent

  names(df)[1] = 'ENTRY'

  names(df)[2] = 'NAME'

  names(df)[3] = 'PATHWAY'

  names(df)[4] = 'RCLASS'            

  names(df)[5] = 'ORTHOLOGY'

  names(df)[6] = 'EQUATION_more'

  names(df)[7] = 'EQUATION'

  names(df)[8] = 'ENZYME' 

  ## BUG FIX; Added by Sumana Srinivasan 0n 04042022, for rno, the ORTHOLOGY ENZYME column, keggGET returns NA and hence have to be named

  

  ###pathway

  if (col_name == "PATHWAY"){

  path_info = df[,c("ENTRY","NAME","PATHWAY")]

  }else{

    path_info = df[,c("ENTRY","NAME","RCLASS","ORTHOLOGY","EQUATION",'EQUATION_more',"ENZYME")]

  }

  

  r = sapply(path_info[,col_name], paste0, collapse=";")

  path_info[["id"]]=NA

  for (i in 1: length(r))

  {

    path_info$countrxn[i] = length(strsplit(r[[i]], ";")[[1]])

    if (!is.null(names(extract_info[[i]][[col_name]]))){

    path_info[["id"]][i] = do.call(paste, c(as.list(names(extract_info[[i]][[col_name]])), sep = ";"))

    }

  }

  path_info[,col_name] = sapply(path_info[,col_name], paste0, collapse=";")

  #aa2 = aa[rep(seq_len(nrow(aa)), aa$countrxn), ]

  path_info3=path_info %>%

    uncount(weights = countrxn, .id = "n", .remove = F) 

  for (j in 1:nrow(path_info3)){

    

    path_info3[j, col_name]= strsplit(path_info3[,col_name][j], ";")[[1]][path_info3$n[j]]

    path_info3[j, "id"]= strsplit(path_info3[,"id"][j], ";")[[1]][path_info3$n[j]]

  }

  

  path_info3 = path_info3[,setdiff(colnames(path_info3), c("countrxn","n"))]

  

  if (col_name == "PATHWAY"){

    colnames(path_info3)[4] = "pathway_id"

  }

  # else if (col_name == "ORTHOLOGY")

  # {colnames(path_info3)[4] = "orthology_id"}

 

  return(path_info3)

}



##### Test code, Sumana Srinivasan, 05052022 #############

# library(plyr); library(KEGGREST);

# rxn_ids <- c("R00021", "R00086") ## give rxn ids that give NA

# info = llply(rxn_ids, function(x)keggGet(x))

# unlist_info <- unlist(info, recursive = F)

# extract_info <- lapply(unlist_info, '[', c("ENTRY","NAME","PATHWAY","RCLASS","ORTHOLOGY","DEFINITION","EQUATION","ENZYME"))

# dd=do.call(rbind, extract_info)

# df = as.data.frame(dd)