GetPriorKnoweldge.R

## ---------------------------
##
## Script name: GetPriorKnowledge
##
## Purpose of script: Create gene-metabolite sets for pathway enrichment analysis.
##
## Author: Christina Schmidt
##
## Date Created: 2024-01-21
##
## Copyright (c) Christina Schmidt
## Email:
##
## ---------------------------
##
## Notes:
##
##
## ---------------------------


##########################################################################################
### ### ### Get KEGG prior knowledge ### ### ###
##########################################################################################

#' Imports KEGG pathways into the environment
#'
#' @title KEGG
#' @description Import and process KEGG.
#' @importFrom utils read.csv
#' @importFrom OmnipathR kegg_link kegg_conv
#' @return A data frame containing the KEGG pathways for ORA.
#'
#' @examples
#' KEGG_Pathways <- MetaProViz::LoadKEGG()
#'
#' @export
#'
LoadKEGG <- function(){
  ## ------------ Create log file ----------- ##
  MetaProViz_Init()

  logger::log_info("Load KEGG.")

  #------------------------------------------------------------------
  # Remove Metabolites
  to_remove <-
    list(
      ### 1. Ions should be removed
      Remove_Ions = c("Calcium cation","Potassium cation","Sodium cation","H+","Cl-", "Fatty acid", "Superoxide","H2O", "CO2", "Copper", "Fe2+", "Magnesium cation", "Fe3+",  "Zinc cation", "Nickel", "NH4+"),
      ### 2. Unspecific small molecules
      Remove_Small = c("Nitric oxide","Hydrogen peroxide", "Superoxide","H2O", "CO2", "Hydroxyl radical", "Ammonia", "HCO3-",  "Oxygen", "Diphosphate", "Reactive oxygen species", "Nitrite", "Nitrate", "Hydrogen", "RX", "Hg")
    ) %>%
    unlist()


  # remotes::install_github("saezlab/OmnipathR@devel")
  KEGG_H <- OmnipathR::kegg_link('compound', 'pathway') %>%
    dplyr::mutate(across(everything(), ~stringr::str_replace(., '^\\w+:', ''))) %>%
    purrr::set_names(c('pathway', 'compound')) %>%
    dplyr::inner_join(OmnipathR::kegg_list('pathway'), by = c(pathway = 'id')) %>%
    dplyr::inner_join(OmnipathR::kegg_list('compound'), by = c(compound = 'id')) %>%
    dplyr::inner_join(
      OmnipathR::kegg_conv('compound', 'pubchem') %>%
        dplyr::mutate(across(everything(), ~stringr::str_replace(., '^\\w+:', ''))),
      by = c(compound = 'id_b')
    ) %>%
    dplyr::rename(pathway_name = name.x, compound_name = name.y, pubchem = id_a)%>%
    dplyr::mutate(
      compound_names = stringr::str_split(compound_name, '; '),
      compound_name = purrr::map_chr(compound_names, ~extract(.x, 1L))
    ) %>%
    dplyr::filter(
      purrr::map_lgl(
        compound_names,
        ~intersect(.x, to_remove) %>% length() == 0
      )
    )%>%
    dplyr::rename(term = pathway_name, Metabolite = compound_name, MetaboliteID = compound, Description = pathway) #Update vignettes and remove rename


  #Return into environment
  assign("KEGG_Pathways", KEGG_H, envir=.GlobalEnv)
  #return(KEGG_H) change all of this to return
}



##########################################################################################
### ### ### Load Hallmark prior knowledge ### ### ###
##########################################################################################
#'
#' @title Prior Knowledge Import
#' @description Import and process .csv file to create Prior Knowledge.
#' @importFrom readr read_csv cols
#' @return A data frame containing the Prior Knowledge.
#' @export
#'
LoadHallmarks <- function() {
  ## ------------ Create log file ----------- ##
  MetaProViz_Init()

  # Read the .csv files
  Hallmark <- system.file("extdata", "Hallmarks.csv.gz", package = "MetaProViz")
  Hallmark <- readr::read_csv(Hallmark, col_types = readr::cols())

  # Return into environment
  assign("Hallmark_Pathways", Hallmark, envir=.GlobalEnv)
}

##########################################################################################
### ### ### Load Gaude Metabolic Signature prior knowledge ### ### ###
##########################################################################################
#'
#' @title Prior Knowledge Import
#' @description Import and process .csv file to create Prior Knowledge.
#' @importFrom readr read_csv cols
#' @return A data frame containing the Prior Knowledge.
#' @export
#'
LoadGaude <- function() {
  ## ------------ Create log file ----------- ##
  MetaProViz_Init()

  # Read the .csv files
  MetabolicSig <- system.file(
      "extdata",
      "Compilled_MetabolicSig_2025-01-07.csv.gz",
      package = "MetaProViz"
  )
  MetabolicSig <- readr::read_csv(MetabolicSig, col_types = readr::cols())

  # Return into environment
  assign("Gaude_Pathways", MetabolicSig, envir=.GlobalEnv)
}

##########################################################################################
### ### ### Load RaMP prior knowledge ### ### ###
##########################################################################################
#'
#' @title Prior Knowledge Import
#' @param version \emph{Optional: } Version of the RaMP database loaded from OmniPathR. \strong{default: "2.5.4"}
#' @param SaveAs_Table \emph{Optional: } File types for the analysis results are: "csv", "xlsx", "txt". \strong{Default = "csv"}
#' @param FolderPath {Optional:} String which is added to the resulting folder name \strong{default: NULL}

#' @description Import and process file to create Prior Knowledge.
#'
#' @importFrom  OmnipathR ramp_table
#' @importFrom rappdirs user_cache_dir
#' @importFrom dplyr filter select group_by summarise mutate
#' @importFrom stringr str_remove
#'
#' @return A data frame containing the Prior Knowledge.
#'
#' @examples
#' ChemicalClass <- MetaProViz::LoadRAMP()
#'
#' @export
#'
LoadRAMP <- function(version = "2.5.4",
                     SaveAs_Table="csv",
                     FolderPath=NULL){
  ## ------------ Create log file ----------- ##
  MetaProViz_Init()

  ## ------------ Folder ----------- ##
  if(is.null(SaveAs_Table)==FALSE){
    Folder <- SavePath(FolderName= "PriorKnowledge",
                       FolderPath=FolderPath)

    SubFolder <- file.path(Folder, "MetaboliteSet")
    if (!dir.exists(SubFolder)) {dir.create(SubFolder)}
  }


  ######################################################
  #Get the directory and filepath of cache results of R
  directory <- rappdirs::user_cache_dir()#get chache directory
  File_path <-paste(directory, "/RaMP-ChemicalClass_Metabolite.rds", sep="")

  if(file.exists(File_path)==TRUE){# First we will check the users chache directory and weather there are rds files with KEGG_pathways already:
    HMDB_ChemicalClass <- readRDS(File_path)
    message("Cached file loaded from: ", File_path)
  }else{# load from OmniPath
  # Get RaMP via OmnipathR and extract ClassyFire classes
  Structure <- OmnipathR::ramp_table( "metabolite_class" , version = version)
  Class <- OmnipathR::ramp_table( "chem_props" , version = version)

  HMDB_ChemicalClass <- merge(Structure, Class[,c(1:3,10)], by="ramp_id", all.x=TRUE)%>%
    dplyr::filter(stringr::str_starts(class_source_id, "hmdb:"))%>% # Select HMDB only!
    dplyr::filter(stringr::str_starts(chem_source_id, "hmdb:"))%>% # Select HMDB only!
    dplyr::select(-c("chem_data_source", "chem_source_id"))%>%
    tidyr::pivot_wider(
      names_from = class_level_name, # Use class_level_name as the new column names
      values_from = class_name,      # Use class_name as the values for the new columns
      values_fn = list(class_name = ~paste(unique(.), collapse = ", ")) # Combine duplicate values
    )%>%
    dplyr::group_by(across(-common_name))%>%
    dplyr::summarise(
      common_name = paste(unique(common_name), collapse = "; "), # Combine all common names into one
      .groups = "drop"  # Ungroup after summarising
    )%>%
    dplyr::mutate(class_source_id = stringr::str_remove(class_source_id, "^hmdb:"))%>% # Remove 'hmdb:' prefix
    dplyr::select(class_source_id, common_name, ClassyFire_class, ClassyFire_super_class, ClassyFire_sub_class) # Reorder columns

  #Save the results as an RDS file in the Cache directory of R
  if(!dir.exists(directory)) {dir.create(directory)}
  saveRDS(HMDB_ChemicalClass, file = paste(directory, "/RaMP-ChemicalClass_Metabolite.rds", sep=""))

  }

  ##-------------- Save and return
  DF_List <- list("ChemicalClass_MetabSet"=HMDB_ChemicalClass)
  suppressMessages(suppressWarnings(
    SaveRes(InputList_DF= DF_List,#This needs to be a list, also for single comparisons
            InputList_Plot= NULL,
            SaveAs_Table=SaveAs_Table,
            SaveAs_Plot=NULL,
            FolderPath= SubFolder,
            FileName= "ChemicalClass",
            CoRe=FALSE,
            PrintPlot=FALSE)))

  # Return into environment
  assign("ChemicalClass_MetabSet", HMDB_ChemicalClass, envir=.GlobalEnv)
}


##########################################################################################
### ### ### Get Metabolite Pathways using Cosmos prior knowledge ### ### ###
##########################################################################################
#'
#' Function to add metabolite HMDB IDs to existing genesets based on cosmosR prior knowledge
#'
#' @param Input_GeneSet Dataframe with two columns for source (=term) and Target (=gene), e.g. Hallmarks.
#' @param SettingsInfo \emph{Optional: }  Column name of Target in Input_GeneSet. \strong{Default = c(Target="gene")}
#' @param PKName \emph{Optional: } Name of the prior knowledge resource. \strong{default: NULL}
#' @param SaveAs_Table \emph{Optional: } File types for the analysis results are: "csv", "xlsx", "txt". \strong{Default = "csv"}
#' @param FolderPath {Optional:} String which is added to the resulting folder name \strong{default: NULL}
#'
#' @export

Make_GeneMetabSet <- function(Input_GeneSet,
                              SettingsInfo=c(Target="gene"),
                              PKName= NULL,
                              SaveAs_Table = "csv",
                              FolderPath = NULL){

  ## ------------ Create log file ----------- ##
  MetaProViz_Init()

  logger::log_info("Make_GeneMetabSet.")

  ## ------------ Check Input files ----------- ##
  # 1. The input data:
  if(is.data.frame(Input_GeneSet)==FALSE){
    stop("`Input_GeneSet` must be of class data.frame with columns for source (=term) and Target (=gene). Please check your input")
  }
  # 2. Target:
  if("Target" %in% names(SettingsInfo)){
    if(SettingsInfo[["Target"]] %in% colnames(Input_GeneSet)== FALSE){
      stop("The ", SettingsInfo[["Target"]], " column selected as Conditions in SettingsInfo was not found in Input_GeneSet. Please check your input.")
    }
  }else{
    stop("Please provide a column name for the Target in SettingsInfo.")
  }

  if(is.null(PKName)){
    PKName <- "GeneMetabSet"
  }

  ## ------------ Folder ----------- ##
  if(is.null(SaveAs_Table)==FALSE){
    Folder <- SavePath(FolderName= "PriorKnowledge",
                                    FolderPath=FolderPath)

    SubFolder <- file.path(Folder, "MetaboliteSet")
    if (!dir.exists(SubFolder)) {dir.create(SubFolder)}
  }


  ######################################################
  ##-------------- Cosmos PKN
  #load the network from cosmos
  data("meta_network", package = "cosmosR")
  meta_network <- meta_network[which(meta_network$source != meta_network$target),]

  #adapt to our needs extracting the metabolites:
  meta_network_metabs <- meta_network[grepl("Metab__", meta_network$source) | grepl("Metab__HMDB", meta_network$target),-2]#extract entries with metabolites in source or Target
  meta_network_metabs <- meta_network_metabs[grepl("Gene", meta_network_metabs$source) | grepl("Gene", meta_network_metabs$target),]#extract entries with genes in source or Target

  #Get reactant and product
  meta_network_metabs_reactant <-  meta_network_metabs[grepl("Metab__HMDB", meta_network_metabs$source),]%>% dplyr::rename("metab"=1, "gene"=2)
  meta_network_metabs_products <-  meta_network_metabs[grepl("Metab__HMDB", meta_network_metabs$target),]%>% dplyr::rename("gene"=1, "metab"=2)

  meta_network_metabs <- as.data.frame(rbind(meta_network_metabs_reactant, meta_network_metabs_products))
  meta_network_metabs$gene <- gsub("Gene.*__","",meta_network_metabs$gene)
  meta_network_metabs$metab <- gsub("_[a-z]$","",meta_network_metabs$metab)
  meta_network_metabs$metab <- gsub("Metab__","",meta_network_metabs$metab)

  ##--------------metalinks transporters
  #Add metalinks transporters to Cosmos PKN

  ##-------------- Combine with Input_GeneSet
  #add pathway names --> File that can be used for metabolite pathway analysis
  MetabSet <- merge(meta_network_metabs,Input_GeneSet, by.x="gene", by.y=SettingsInfo[["Target"]])

  #combine with pathways --> File that can be used for combined pathway analysis (metabolites and gene t.vals)
  GeneMetabSet <- unique(as.data.frame(rbind(Input_GeneSet%>%dplyr::rename("feature"=SettingsInfo[["Target"]]), MetabSet[,-1]%>%dplyr::rename("feature"=1))))


  ##------------ Select metabolites only
  MetabSet <-  GeneMetabSet %>%
    filter(grepl("HMDB", feature))


  ##-------------- Save and return
  DF_List <- list("GeneMetabSet"=GeneMetabSet,
                  "MetabSet"=MetabSet)
  suppressMessages(suppressWarnings(
    SaveRes(InputList_DF= DF_List,#This needs to be a list, also for single comparisons
                         InputList_Plot= NULL,
                         SaveAs_Table=SaveAs_Table,
                         SaveAs_Plot=NULL,
                         FolderPath= SubFolder,
                         FileName= PKName,
                         CoRe=FALSE,
                         PrintPlot=FALSE)))

  return(invisible(DF_List))
}



##########################################################################################
### ### ### Load MetaLinksDB prior knowledge ### ### ###
##########################################################################################
#'
#' Function to
#'
#' @param types Desired edge types. Options are: "lr", "pd", where 'lr' stands for 'ligand-receptor' and 'pd' stands for 'production-degradation'.\strong{default: NULL}
#' @param cell_location Desired metabolite cell locations. Pass selection using c("Select1", "Select2", "Selectn"). View options setting "?". Options are: "Cytoplasm", "Endoplasmic reticulum", "Extracellular", "Lysosome" , "Mitochondria", "Peroxisome", "Membrane", "Nucleus", "Golgi apparatus" , "Inner mitochondrial membrane". \strong{default: NULL}
#' @param tissue_location Desired metabolite tissue locations. Pass selection using c("Select1", "Select2", "Selectn"). View options setting "?". Options are: "Placenta", "Adipose Tissue","Bladder", "Brain", "Epidermis","Kidney", "Liver", "Neuron", "Pancreas", "Prostate", "Skeletal Muscle", "Spleen", "Testis", "Thyroid Gland", "Adrenal Medulla", "Erythrocyte","Fibroblasts", "Intestine", "Ovary", "Platelet", "All Tissues", "Semen", "Adrenal Gland", "Adrenal Cortex", "Heart", "Lung", "Hair", "Eye Lens", "Leukocyte", Retina", "Smooth Muscle", "Gall Bladder", "Bile",  "Bone Marrow", "Blood", "Basal Ganglia", "Cartilage". \strong{default: NULL}
#' @param biospecimen_location Desired metabolite biospecimen locations.Pass selection using c("Select1", "Select2", "Selectn").View options setting "?".  "Blood", "Feces", "Saliva", "Sweat", "Urine", "Breast Milk", "Cellular Cytoplasm", "Cerebrospinal Fluid (CSF)", "Amniotic Fluid" , "Aqueous Humour", "Ascites Fluid", "Lymph", "Tears", "Breath", "Bile", "Semen", "Pericardial Effusion".\strong{default: NULL}
#' @param disease Desired metabolite diseases.Pass selection using c("Select1", "Select2", "Selectn"). View options setting "?". \strong{default: NULL}
#' @param pathway Desired metabolite pathways.Pass selection using c("Select1", "Select2", "Selectn"). View options setting "?".\strong{default: NULL}
#' @param hmdb_ids Desired HMDB IDs.Pass selection using c("Select1", "Select2", "Selectn"). View options setting "?".\strong{default: NULL}
#' @param uniprot_ids Desired UniProt IDs.Pass selection using c("Select1", "Select2", "Selectn"). View options setting "?".\strong{default: NULL}
#' @param SaveAs_Table \emph{Optional: } File types for the analysis results are: "csv", "xlsx", "txt". \strong{Default = "csv"}
#' @param FolderPath \emph{Optional:} Path to the folder the results should be saved at. \strong{default: NULL}
#'
#' @importFrom DBI dbListTables dbDisconnect dbGetQuery
#' @importFrom OmnipathR metalinksdb_sqlite
#' @importFrom logger log_info
#' @importFrom dplyr mutate
#' @export
LoadMetalinks <- function(types = NULL,
                          cell_location = NULL,
                          tissue_location = NULL,
                          biospecimen_location = NULL,
                          disease = NULL,
                          pathway = NULL,
                          hmdb_ids = NULL,
                          uniprot_ids = NULL,
                          SaveAs_Table="csv",
                          FolderPath=NULL){
  ## ------------ Create log file ----------- ##
  MetaProViz_Init()

  logger::log_info("MetaLinksDB.")


  #------------------------------------------------------------------
  if((any(c(types, cell_location, tissue_location, biospecimen_location, disease, pathway, hmdb_ids, uniprot_ids)=="?"))==FALSE){
  #Check Input parameters



    }

  #Python version enables the user to add their own link to the database dump (probably to obtain a specific version. Lets check how the link was generated and see if it would make sense for us to do the same.)
  # --> At the moment arbitrary!
  # We could provide the user the ability to point to their own path were they already dumpled/stored QA version of metalinks they like to use!

  ## ------------ Folder ----------- ##
  if(is.null(SaveAs_Table)==FALSE){
    Folder <- SavePath(FolderName= "PriorKnowledge",
                                    FolderPath=FolderPath)

    SubFolder <- file.path(Folder, "MetaboliteSet")
    if (!dir.exists(SubFolder)) {dir.create(SubFolder)}
  }

  con <- OmnipathR::metalinksdb_sqlite()
  on.exit(DBI::dbDisconnect(con))
  #------------------------------------------------------------------
  #Query the database for a specific tables
  tables <- DBI::dbListTables(con)

  TablesList <- list()
  for(table in tables){
    query <- paste("SELECT * FROM", table)
    data <- DBI::dbGetQuery(con, query)
    TablesList[[table]] <- data
  }

  # Close the connection

  MetalinksDB <- TablesList[["edges"]]#extract the edges table
  #------------------------------------------------------------------
  # Answer questions about the database
  #if any parameter is ? then return the data
  if(any(c(types, cell_location, tissue_location, biospecimen_location, disease, pathway, hmdb_ids, uniprot_ids)=="?")){
    Questions <- which(c(types, cell_location, tissue_location, biospecimen_location, disease, pathway, hmdb_ids, uniprot_ids)=="?")
    #Check tables where the user has questions
    if(length(Questions)>0){
      for(i in Questions){
        if(i==1){
          print("Types:")
          print(unique(MetalinksDB$type))
        }
        if(i==2){
          print("Cell Location:")
          CellLocation <- TablesList[["cell_location"]]
          print(unique(CellLocation$cell_location))
        }
        if(i==3){
          print("Tissue Location:")
          TissueLocation <- TablesList[["tissue_location"]]
          print(unique(TissueLocation$tissue_location))
        }
        if(i==4){
          print("Biospecimen Location:")
          BiospecimenLocation <- TablesList[["biospecimen_location"]]
          print(unique(BiospecimenLocation$biospecimen_location))
        }
        if(i==5){
          print("Disease:")
          Disease <- TablesList[["disease"]]
          print(unique(Disease$disease))
        }
        if(i==6){
          print("Pathway:")
          Pathway <- TablesList[["pathway"]]
          print(unique(Pathway$pathway))
        }
        if(i==7){
          print("HMDB IDs:")
          print(unique(MetalinksDB$hmdb))
        }
        if(i==8){
          print("UniProt IDs:")
          print(unique(MetalinksDB$uniprot))
        }
      }
    }
    message("No result is returned unless correct options for your selections are used. `?` is not a valid option, but only returns you the list of options.")
    return()
  }

  #------------------------------------------------------------------
  #Extract specific connections based on parameter settings. If any parameter is not NULL, filter the data:
  ## types
  if(!is.null(types)){
    MetalinksDB <- MetalinksDB[MetalinksDB$type %in% types,]
  }

  ## cell_location
  if(!is.null(cell_location)){
    CellLocation <- TablesList[["cell_location"]]
    CellLocation <- CellLocation[CellLocation$cell_location %in% cell_location,]#Filter the cell location

    #Get unique HMDB IDs
    CellLocation_HMDB <- unique(CellLocation$hmdb)

    #Only keep selected HMDB IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$hmdb %in%  CellLocation_HMDB,]
  }

  ## tissue_location
  if(!is.null(tissue_location)){# "All Tissues"?
    TissueLocation <- TablesList[["tissue_location"]]
    TissueLocation <- TissueLocation[TissueLocation$tissue_location %in% tissue_location,]#Filter the tissue location

    #Get unique HMDB IDs
    TissueLocation_HMDB <- unique(TissueLocation$hmdb)

    #Only keep selected HMDB IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$hmdb %in%  TissueLocation_HMDB,]
  }

  ## biospecimen_location
  if(!is.null(biospecimen_location)){
    BiospecimenLocation <- TablesList[["biospecimen_location"]]
    BiospecimenLocation <- BiospecimenLocation[BiospecimenLocation$biospecimen_location %in% biospecimen_location,]#Filter the biospecimen location

    #Get unique HMDB IDs
    BiospecimenLocation_HMDB <- unique(BiospecimenLocation$hmdb)

    #Only keep selected HMDB IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$hmdb %in%  BiospecimenLocation_HMDB,]
  }

  ## disease
  if(!is.null(disease)){
    Disease <- TablesList[["disease"]]
    Disease <- Disease[Disease$disease %in% disease,]#Filter the disease

    #Get unique HMDB IDs
    Disease_HMDB <- unique(Disease$hmdb)

    #Only keep selected HMDB IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$hmdb %in%  Disease_HMDB,]
  }

  ## pathway
  if(!is.null(pathway)){
    Pathway <- TablesList[["pathway"]]
    Pathway <- Pathway[Pathway$pathway %in% pathway,]#Filter the pathway

    #Get unique HMDB IDs
    Pathway_HMDB <- unique(Pathway$hmdb)

    #Only keep selected HMDB IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$hmdb %in%  Pathway_HMDB,]
  }

  ## hmdb_ids
  if(!is.null(hmdb_ids)){
    #Only keep selected HMDB IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$hmdb %in%  hmdb_ids,]
  }

  ## uniprot_ids
  if(!is.null(uniprot_ids)){
    #Only keep selected UniProt IDs
    MetalinksDB <- MetalinksDB[MetalinksDB$uniprot %in%  uniprot_ids,]
  }

  #------------------------------------------------------------------
  #Add other ID types:
  ## Metabolite Name
  MetalinksDB <- merge(MetalinksDB, TablesList[["metabolites"]], by="hmdb", all.x=TRUE)

  ## Gene Name
  MetalinksDB <- merge(MetalinksDB, TablesList[["proteins"]], by="uniprot", all.x=TRUE)


  ## Rearrange columns:
  MetalinksDB <- MetalinksDB[,c(2,10:12, 1, 13:14, 3:9)]%>%
    dplyr::mutate(type = dplyr::case_when(
      type == "lr" ~ "Ligand-Receptor",
      type == "pd" ~ "Production-Degradation",
      TRUE ~ type  # this keeps the original value if it doesn't match any condition
    ))%>%
    dplyr::mutate(mode_of_regulation = dplyr::case_when(
    mor == -1 ~ "Inhibiting",
    mor == 1 ~ "Activating",
    mor == 0 ~ "Binding",
    TRUE ~ as.character(mor)  # this keeps the original value if it doesn't match any condition
  ))
  #--------------------------------------------------------
  #Remove metabolites that are not detectable by mass spectrometry


  #------------------------------------------------------------------
  #Decide on useful selections term-metabolite for MetaProViz.
  #MetalinksDB_Pathways <- merge(MetalinksDB[, c(1:3)], TablesList[["pathway"]], by="hmdb", all.x=TRUE)

  MetalinksDB_Type <-MetalinksDB[, c(1:3, 5:7,13, 15)]%>%
    subset(!is.na(protein_type))%>%
    dplyr::mutate(term = gsub('\"', '', protein_type))%>%
    unite(term_specific, c("term", "type"), sep = "_", remove = FALSE)%>%
    dplyr::select(-"type", -"protein_type")

  #------------------------------------------------------------------
  #Save results in folder
  ##-------------- Save and return
  DF_List <- list("MetalinksDB"=MetalinksDB,
                  "MetalinksDB_Type"=MetalinksDB_Type)
  suppressMessages(suppressWarnings(
    SaveRes(InputList_DF= DF_List,#This needs to be a list, also for single comparisons
                         InputList_Plot= NULL,
                         SaveAs_Table=SaveAs_Table,
                         SaveAs_Plot=NULL,
                         FolderPath= SubFolder,
                         FileName= "MetaLinksDB",
                         CoRe=FALSE,
                         PrintPlot=FALSE)))

  #Return into environment
  return(invisible(DF_List))

}

##########################################################################################
### ### ### Helper: Load compound lists of ions, xenobiotics, cofactors (not detectable by LC-MS), etc ### ### ###
##########################################################################################

# This is needed to remove ions, xenobiotics, cofactors, etc. from the metabolite list for any of the functions above.