In dongzhuoer/hgnc: map common gene ID (accession) to HUGO gene symbol

knitr::opts_knit$set(root.dir = here::here(''))
knitr::opts_chunk$set(collapse = T)

clean

dir(here::here('data'), full.names = T) %>% file.remove()
rm(list = ls(envir = globalenv(), all = T))

Process raw data

hgnc_complete_set.txt.gz comes from https://www.genenames.org/cgi-bin/statistics#complete

gene2unigene.tsv comes from ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene2unigene

hgnc_all <- readr::read_tsv('data-raw/hgnc_complete_set.txt.gz', T, libzhuoer::cols_char()) 
hgnc <- hgnc_all %>% dplyr::filter(status == 'Approved')

hugo_symbol <- hgnc$symbol;



entrez2symbol <- dplyr::select(hgnc, entrez = entrez_id, symbol) %>% hgnc::qualify_map() %T>% print

ensembl2symbol <- dplyr::select(hgnc, ensembl = ensembl_gene_id, symbol) %>% hgnc::qualify_map() %T>% print

hgnc$ensembl_gene_id %>% stringr::str_extract('[^\\d]+') %>% na.omit() %>% unique

entrez_or_symbol2symbol <- dplyr::bind_rows(
    tibble::tibble(entrez_or_symbol = hugo_symbol, symbol = hugo_symbol), 
    dplyr::rename(entrez2symbol, entrez_or_symbol = 'entrez')
) %>% hgnc::qualify_map() %T>% print



genbank2symbol <- dplyr::bind_rows(
    dplyr::select(hgnc, symbol, genbank = 'ena'),
    dplyr::select(hgnc, symbol, genbank = 'refseq_accession')
) %>% hgnc::melt_map('symbol', 'genbank', '\\|') %>% dplyr::select(2:1) %T>% 
    {message('collapsing ', length(stringr::str_subset(.[[1]], '\\.\\d+$')), ' transcripts to genes')} %T>% 
    {message('If this number is too bug,'); .} %T>% 
    {message('check whether different transcripts of the same gene have different symbols')} %>%
    dplyr::mutate_at(1, . %>% stringr::str_remove('\\.\\d+$')) %>% hgnc::qualify_map() %T>% print

unigene2entrez <- 'data-raw/gene2unigene.tsv.gz' %>% 
    readr::read_tsv(T, libzhuoer::cols_char()) %>% 
    dplyr::filter_at(2, dplyr::any_vars(stringr::str_detect(., '^Hs\\.'))) %>% 
    dplyr::select(2, 1) %>% hgnc::qualify_map() %>%
    dplyr::select(unigene = 1, entrez = 2) %T>% print

usethis::use_data(hugo_symbol, entrez2symbol, ensembl2symbol, genbank2symbol, entrez_or_symbol2symbol, unigene2entrez, overwrite = T);

devtools::test()     # test the new data
roxygen2::roxygenize() # you may also have edited data documentation

system('R CMD INSTALL --no-multiarch --with-keep.source .')
devtools::reload()   # now you can use the new data in current R session

mapply(
    function(symbol, name) {
        coverage <- symbol %>% unique %>% length %>% {./length(hgnc::hugo_symbol)};
        message(name, ' covers ', format(coverage, F, 6), ' symbols')
    },
    list(hgnc::entrez2symbol$symbol, hgnc::ensembl2symbol$symbol, hgnc::genbank2symbol$symbol, 
         hgnc::as_symbol_from_unigene(hgnc::unigene2entrez$unigene)),
    c('entrez  ', 'ensembl  ', 'genbank', 'unigene ')
) -> dev.null