Error in Constructing public MS2 datbase using metid

[liaochenlanruo]

Hi Shen,
When I try to construct public MS2 database using metid, some errors happened as below （someone else also suffered this error : R package：metID(六)：代谢物的鉴定）, any idea to deal with that? Thanks!

Error in `dplyr::select()`: 
! Can't subset columns that don't exist. 
✖ Column `Name` doesn't exist. 
Run `rlang::last_trace()` to see where the error occurred. 
> rlang::last_trace() 
<error/vctrs_error_subscript_oob>
 Error in `dplyr::select()`:
 ! Can't subset columns that don't exist. 
✖ Column `Name` doesn't exist.
--- 
Backtrace: 
 ▆  
1. ├─metid::construct_mona_database(...)  
2. │ └─metabolite_info %>% ...
3. ├─dplyr::select(...)  
4. └─dplyr:::select.data.frame(...) 
Run rlang::last_trace(drop = FALSE) to see 18 hidden frames. 
> rlang::last_trace(drop = FALSE) 
<error/vctrs_error_subscript_oob> 
Error in `dplyr::select()`: 
! Can't subset columns that don't exist. 
✖ Column `Name` doesn't exist.
--- 
Backtrace: 
 ▆  
1. ├─metid::construct_mona_database(...)  
2. │ └─metabolite_info %>% ...  
3. ├─dplyr::select(...)  
4. ├─dplyr:::select.data.frame(...)  
5. │ └─tidyselect::eval_select(expr(c(...)), data = .data, error_call = error_call)  
6. │   └─tidyselect:::eval_select_impl(...)  
7. │     ├─tidyselect:::with_subscript_errors(...)  
8. │     │ └─rlang::try_fetch(...)  
9. │     │   └─base::withCallingHandlers(...)  
10. │     └─tidyselect:::vars_select_eval(...)  
11. │       └─tidyselect:::walk_data_tree(expr, data_mask, context_mask)  
12. │         └─tidyselect:::eval_c(expr, data_mask, context_mask)  
13. │           └─tidyselect:::reduce_sels(node, data_mask, context_mask, init = init)  
14. │             └─tidyselect:::walk_data_tree(new, data_mask, context_mask)  
15. │               └─tidyselect:::as_indices_sel_impl(...)  
16. │                 └─tidyselect:::as_indices_impl(...)  
17. │                   └─tidyselect:::chr_as_locations(x, vars, call = call, arg = arg) 
18. │                     └─vctrs::vec_as_location(...)  
19. └─vctrs (local) `<fn>`()  
20.  └─vctrs:::stop_subscript_oob(...)  
21.  └─vctrs:::stop_subscript(...)  
22.  └─rlang::abort(...)

[jaspershen]

Please use massdatabase for the public database construction.
You can install massdatabase here: https://massdatabase.tidymass.org/index.html

data <- 
  massdatabase::read_msp_data("MoNA-export-CASMI_2016.msp", source = "mona")

massdatabase::convert_mona2metid(data = data, path = ".", threads = 3)

[ShawnWx2019]

I encountered the same issue. After parsing step by step, I found that it might be the problem with the .mgf files provided by the MoNA database. For instance, some .mgf files do not provide collision energy, and some positive and negative spectra are not marked properly. For example, there are spaces before "p" labels => p. Therefore, when MetID parses and converts these .mgf files, will report errors.

Then, I modified the construct_mona_database() function to address these two issues.

construct_mona_database2 = function(
    file, only.remain.ms2 = TRUE, path = ".", version = "0.0.1", 
    source = "MoNA", link = "https://mona.fiehnlab.ucdavis.edu/", 
    creater = "Xiaotao Shen", email = "[email protected]", rt = FALSE, 
    threads = 5
    ) {
  mona_database = read_msp_mona(file = file)
  
  #> Issue1: set rownames of mona_database[[i]]$info
    mona_database = purrr::map(mona_database,function(x){
    x$info = data.frame(
      row.names = x$info$info,
      value = x$info$value
    )
    db = list(info = x$info,
              spec = x$spec)
    return(db)
    })
    
    all_metabolite_names = purrr::map(mona_database, function(x) {
        rownames(x$info)
    }) %>% unlist() %>% unique()
    metabolite_info = mona_database %>% purrr::map(function(x) {
        x = as.data.frame(x$info)
        new_x = x[, 1]
        names(new_x) = rownames(x)
        new_x = new_x[all_metabolite_names]
        names(new_x) = all_metabolite_names
        new_x
    }) %>% do.call(rbind, .) %>% as.data.frame()
    colnames(metabolite_info) = all_metabolite_names
    if (only.remain.ms2) {
        remain_idx = which(metabolite_info$Spectrum_type == "MS2")
        metabolite_info = metabolite_info[remain_idx, ]
        mona_database = mona_database[remain_idx]
    }
    metabolite_info = 
      metabolite_info %>% 
      dplyr::select(
        Compound.name = Name, 
        mz = ExactMass, Formula,
        MoNA.ID = `DB#`, 
        dplyr::everything()
        )
    metabolite_info = 
      metabolite_info %>% 
      dplyr::mutate(
        Lab.ID = paste("MoNA", seq_len(nrow(metabolite_info)), sep = "_"), 
        RT = NA, 
        CAS.ID = NA, 
        HMDB.ID = NA, 
        KEGG.ID = NA, 
        mz.pos = NA, 
        mz.neg = NA, 
        Submitter = "MoNA", 
        Family = NA, 
        Sub.pathway = NA, 
        Note = NA) %>% 
      dplyr::select(
        Lab.ID, 
        Compound.name, 
        mz, 
        RT, 
        CAS.ID, 
        HMDB.ID, 
        KEGG.ID, 
        Formula, 
        mz.pos, 
        mz.neg, 
        Submitter, 
        Family, 
        Sub.pathway, 
        Note, 
        dplyr::everything()
        )
    #> Issue2: Collision_energy
     if(!"Collision_energy"%in%colnames(metabolite_info)) {
       metabolite_info$Collision_energy = NA
       }
    
    metabolite_info$Collision_energy[is.na(metabolite_info$Collision_energy)] = "not_available"
    metabolite_info$Collision_energy[metabolite_info$Collision_energy == ""] = "not_available"
    
    #> Issue3: Ion_mode
    metabolite_info = 
    metabolite_info %>% 
    mutate(Ion_mode = 
      case_when(
        str_detect(Ion_mode,regex("P",ignore_case = T)) ~ "P",
        str_detect(Ion_mode,regex("N",ignore_case = T)) ~ "N"
      )
    )
    
    positive_idx = which(metabolite_info$Ion_mode == "P")
    negative_idx = which(metabolite_info$Ion_mode == "N")
    Spectra.positive = mona_database[positive_idx]
    Spectra.negative = mona_database[negative_idx]
    names(Spectra.positive) = metabolite_info$Lab.ID[positive_idx]
    names(Spectra.negative) = metabolite_info$Lab.ID[negative_idx]
    Spectra.positive = purrr::map2(.x = Spectra.positive, .y = metabolite_info$Collision_energy[positive_idx], 
        .f = function(x, y) {
            x = x$spec
            x = list(x)
            names(x) = y
            x
        })
    Spectra.negative = purrr::map2(.x = Spectra.negative, .y = metabolite_info$Collision_energy[negative_idx], 
        .f = function(x, y) {
            x = x$spec
            x = list(x)
            names(x) = y
            x
        })
    database.info <- list(Version = version, Source = source, 
        Link = link, Creater = creater, Email = email, RT = rt)
    spectra.info <- as.data.frame(metabolite_info)
    rm(list = "metabolite_info")
    Spectra <- list(Spectra.positive = Spectra.positive, Spectra.negative = Spectra.negative)
    database <- new(Class = "databaseClass", database.info = database.info, 
        spectra.info = spectra.info, spectra.data = Spectra)
    database@database.info$RT <- ifelse(all(is.na(database@spectra.info$RT)), 
        FALSE, TRUE)
    message(crayon::bgRed("All done!\n"))
    return(database)
}