data-raw/biomarker_utilities.R
In sigven/oncoPharmaDB: Molecularly targeted cancer drugs and biomarkers
clean_variant_consequence <- function(variant_df){
variant_df <- variant_df |>
dplyr::mutate(
variant_consequence =
dplyr::if_else(
stringr::str_detect(variant,"Overexpression") &
variant_consequence == "N/A",
"transcript_amplification",
as.character(variant_consequence))) |>
dplyr::mutate(
variant_consequence =
dplyr::if_else(
stringr::str_detect(variant,"Underexpression") &
variant_consequence == "N/A",
"transcript_ablation",
as.character(variant_consequence))) |>
dplyr::mutate(
variant_consequence =
dplyr::if_else(
stringr::str_detect(
variant_consequence,
"^(plus_1_frameshift|minus_1_frameshift|frameshift_)"),
"frameshift_variant",
as.character(variant_consequence))) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
variant_consequence == "" &
stringr::str_detect(variant,"EXON") &
stringr::str_detect(variant,"DELETION"),
"exon_loss_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
variant_consequence == "" &
stringr::str_detect(variant,"Exon") &
stringr::str_detect(variant,"Mutation"),
"exon_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
variant_consequence == "" &
stringr::str_detect(variant,"^(Loss|Loss-of-function)$"),
"loss_of_function_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant,"^(Activating |Gain-of)") &
(is.na(variant_consequence) | nchar(variant_consequence) == 0),
"gain_of_function_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant,"^Wildtype") &
(is.na(variant_consequence) | nchar(variant_consequence) == 0),
"wild_type",
as.character(variant_consequence)
))
return(variant_df)
}
set_alteration_type <- function(variant_df){
variant_df <- variant_df |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(toupper(variant),"EXPRESSION"),
"EXP",as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::case_when(
alteration_type == "EXP" &
stringr::str_detect(
toupper(variant),"OVER") ~ "EXP_OVER",
alteration_type == "EXP" &
stringr::str_detect(
toupper(variant),"UNDER") ~ "EXP_UNDER",
TRUE ~ as.character(alteration_type)
)
) |>
dplyr::mutate(alteration_type = dplyr::if_else(
variant_consequence == "stop_gained",
"MUT_LOF",as.character(alteration_type)
)) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant),"LOSS-OF-FUNCTION|LOSS|TRUNCAT|INACTIVAT") |
stringr::str_detect(
variant_consequence,
"stop_gained,loss_of_function_variant|loss_of_function_variant,stop_gained") |
(stringr::str_detect(
variant_consequence, "loss_of_function_variant") &
toupper(variant) == "MUTATION") |
stringr::str_detect(variant_consequence,"truncation"),
"MUT_LOF",as.character(alteration_type))) |>
dplyr::mutate(alteration_type = dplyr::if_else(
toupper(variant) == "BIALLELIC INACTIVATION",
"MUT_LOF_BIALLELIC",
as.character(alteration_type)
)) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
alteration_type == 'MUT_LOF' &
(stringr::str_detect(
toupper(variant_consequence),"FRAMESHIFT") |
stringr::str_detect(
variant_consequence,"feature_truncation")),
"MUT_LOF_FS",as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant),
"AMPLIFICATION|DELETION|COPY") &
stringr::str_detect(
variant_consequence, "_amplification|_ablation"),
"CNA",as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant_consequence),
"TRANSCRIPT_FUSION|REGION_FUSION") |
stringr::str_detect(toupper(variant),"FUSION") |
(variant_consequence == "" &
stringr::str_detect(variant,"::")),
"FUSION",
as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant),"REARRANGEMENT") &
variant_consequence == "",
"REARRANGEMENT",
as.character(alteration_type)))
return(variant_df)
}
clean_variant_naming <- function(
variant_df, col = "molecular_profile_name"){
variant_df[,"cc_clean"] <- variant_df[,col]
variant_df <- variant_df |>
dplyr::mutate(cc_clean = stringr::str_trim(cc_clean)) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean,"RS[0-9]{5,}"),
stringr::str_replace_all(cc_clean,"RS","rs"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = stringr::str_replace_all(
cc_clean,
paste0(
"(ENST[0-9]{1,}|",
"NM_[0-9]{1,})(\\.[0-9]{1,})?:c\\."),"c."
)) |>
dplyr::mutate(cc_clean = stringr::str_replace_all(
cc_clean,
paste0(
"(ENSP[0-9]{1,}|NP_[0-9]{1,}",
")(\\.[0-9]{1,})?:p\\."),"p."
)) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "^NC_[0-9]{1,}"),
"", as.character(cc_clean)
)) |>
dplyr::mutate(cc_clean = dplyr::case_when(
cc_clean == "D835H/Y" ~ "D835H,D835Y",
cc_clean == "G12/G13" ~ "G12,G13",
cc_clean == "Q157P/R" ~ "Q157P,Q157R",
cc_clean == "S310F/Y" ~ "S310F,S310Y",
cc_clean == "S34Y/F" ~ "S34Y,S34F",
cc_clean == "S893A/T" ~ "S893A,S893T",
TRUE ~ as.character(cc_clean)
)) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "^(CD44s|p16|PRPS1|IKZF1|NUCLEAR) "),
stringr::str_replace_all(
cc_clean, "^(CD44s|p16|PRPS1|IKZF1|NUCLEAR) ", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, " NUCLEAR "),
stringr::str_replace_all(
cc_clean, " NUCLEAR ", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "IVS2\\+1G>A"),
stringr::str_replace_all(
cc_clean, "^IVS2\\+1G>A", "c.444+1G>A"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, " HOMOZYGOSITY"),
stringr::str_replace_all(
cc_clean, " HOMOZYGOSITY", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "deletion and mutation"),
stringr::str_replace_all(
cc_clean, "deletion and mutation", "Mutation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = stringr::str_replace(
cc_clean, " \\(.+\\)$","")) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "\\*$"),
stringr::str_replace_all(cc_clean, "\\*$", "X"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "^\\*"),
stringr::str_replace_all(cc_clean, "^\\*", "X"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "\\* "),
stringr::str_replace_all(cc_clean, "\\* ", "X "),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Rare "),
stringr::str_replace_all(cc_clean, "Rare ", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, " expression"),
stringr::str_replace_all(
cc_clean, " expression", " Expression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "BIALLELIC"),
stringr::str_replace_all(
cc_clean, "BIALLELIC", "Biallelic"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "MUTATION|mutation"),
stringr::str_replace_all(
cc_clean, "MUTATION|mutation", "Mutation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "DOMAIN"),
stringr::str_replace_all(
cc_clean, "DOMAIN", "Domain"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "CONSERVED"),
stringr::str_replace_all(
cc_clean, "CONSERVED", "Conserved"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "DELETION"),
stringr::str_replace_all(
cc_clean, "DELETION", "Deletion"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "FRAMESHIFT|FRAME SHIFT"),
stringr::str_replace_all(
cc_clean, "FRAMESHIFT|FRAME SHIFT", "Frameshift"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "INACTIVATING"),
stringr::str_replace_all(
cc_clean, "INACTIVATING", "Inactivating"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "INACTIVATION"),
stringr::str_replace_all(
cc_clean, "INACTIVATION", "Inactivation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "OVEREXPRESSION"),
stringr::str_replace_all(
cc_clean, "OVEREXPRESSION", "Overexpression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "UNDEREXPRESSION"),
stringr::str_replace_all(
cc_clean, "UNDEREXPRESSION", "Underexpression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "EXPRESSION"),
stringr::str_replace_all(
cc_clean, "EXPRESSION", "Expression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "EXON"),
stringr::str_replace_all(cc_clean, "EXON", "Exon"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "LOSS$"),
stringr::str_replace_all(cc_clean, "LOSS", "Loss"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "AMPLIFICATION"),
stringr::str_replace_all(
cc_clean, "AMPLIFICATION", "Amplification"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "TRUNCATING"),
stringr::str_replace_all(
cc_clean, "TRUNCATING", "Truncating"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "TRUNCATION"),
stringr::str_replace_all(
cc_clean, "TRUNCATION", "Truncation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "TRANSLOCATION|rearrangement"),
stringr::str_replace_all(
cc_clean, "TRANSLOCATION|rearrangement", "Rearrangement"),
as.character(cc_clean)))
if(col != "molecular_profile_name"){
variant_df <- variant_df |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Binding Domain Mutation") &
feature_name == "TP53",
"aa_region:100-288",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "SH2 Domain") &
feature_name == "STAT3",
"aa_region:584-651",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Kinase Domain") &
feature_name == "ERBB2",
"aa_region:722-975",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "TKD ") &
feature_name == "FLT3",
"aa_region:610-942",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Conserved Domain") &
feature_name == "TP53",
paste0(
"aa_region:117-143,aa_region:171-181,",
"aa_region:234-258,aa_region:270-286"),
as.character(cc_clean)))
}
variant_df <- variant_df |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "N/A"),
stringr::str_replace_all(cc_clean, "N/A", ""),
as.character(cc_clean))) |>
dplyr::distinct()
variant_df[,col] <- variant_df[,"cc_clean"]
variant_df[,"cc_clean"] <- NULL
return(variant_df)
}
get_literature_references <- function(biomarker_items){
## set logging layout
lgr::lgr$appenders$console$set_layout(
lgr::LayoutFormat$new(timestamp_fmt = "%Y-%m-%d %T"))
biomarker_items[['literature']] <- as.data.frame(
biomarker_items[['clinical']] |>
dplyr::filter(!is.na(citation_id)) |>
tidyr::separate_rows(citation_id, sep=";") |>
dplyr::mutate(
citation_id = stringr::str_trim(citation_id)) |>
dplyr::filter(
!stringr::str_detect(
tolower(citation_id),
"asco|aacr|abstr|suppl|confex|jci|blood|annonc|caris")) |>
dplyr::distinct() |>
dplyr::mutate(citation_id = dplyr::case_when(
citation_id == "PMC3638050" ~ "PMID:22038996",
citation_id == "PMC3936420" ~ "PMID:24265155",
TRUE ~ as.character(
stringr::str_replace(citation_id,"PMID: ","PMID:"))
)) |>
dplyr::filter(!stringr::str_detect(citation_id, "^[0-9]{1,5}$")) |>
dplyr::filter(citation_id != "PMID:238900088" &
citation_id != "PMID:24569783" &
citation_id != "2556735") |>
dplyr::mutate(citation_id = dplyr::case_when(
stringr::str_detect(citation_id, "^FDA") ~ "FDA",
stringr::str_detect(citation_id, "NCCN") ~ "NCCN",
TRUE ~ as.character(citation_id)
)) |>
dplyr::mutate(citation_id = stringr::str_replace(
citation_id, "PMID:",""
)) |>
dplyr::select(citation_id, evidence_id) |>
dplyr::rename(source_id = citation_id) |>
dplyr::mutate(source_id = as.character(source_id)) |>
dplyr::group_by(source_id) |>
dplyr::summarise(evidence_id = paste(
unique(sort(evidence_id)), collapse = ";"
), .groups = "drop") |>
dplyr::left_join(
literatureVault::vault,
by = "source_id",
relationship = "many-to-many",
multiple = "all"
)
)
biomarker_items[['clinical']]$citation_id <- NULL
not_found <- biomarker_items$literature |>
dplyr::filter(is.na(link)) |>
dplyr::select(source, source_id)
if(NROW(not_found) > 0){
lgr::lgr$info(
"Citations missing in sigven/literatureVault::vault - update needed")
return(not_found)
}else{
lgr::lgr$info(
"Found all literature references in sigven/literatureVault::vault"
)
}
## ignore molecular profiles/biomarkers that lack literature support
biomarkers_with_literature_support <- as.data.frame(
biomarker_items[['literature']] |>
dplyr::select(evidence_id, source_id) |>
tidyr::separate_rows(evidence_id, sep = ";") |>
dplyr::group_by(evidence_id) |>
dplyr::mutate(source_id = paste(
unique(source_id), collapse = ";"
)) |>
dplyr::distinct()
)
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::inner_join(
biomarkers_with_literature_support,
by = "evidence_id", multiple = "all",
relationship = "many-to-many")
return(biomarker_items)
}
#' #' A function that splits an array into chunks of equal size
#' chunk <- function(x,n) split(x, factor(sort(rank(x)%%n)))
#'
#' #' A function that returns a citation with first author, journal and year for a PubMed ID
#' #'
#' #' @param pmid An array of Pubmed IDs
#' #' @param chunk_size Size of PMID chunks
#' #'
#' #' @export
#' #'
#' #' @return citation PubMed citation, with first author, journal and year
#' #'
#' get_citations_pubmed <- function(
#' pmid,
#' pmid_cache_dir = NA,
#' chunk_size = 2){
#'
#' ## make chunk of maximal 400 PMIDs from input array (limit by EUtils)
#' pmid_chunks <- chunk(pmid, ceiling(length(pmid)/chunk_size))
#' j <- 0
#' all_citations <- data.frame()
#' cat('Retrieving PubMed citations for PMID list, total length', length(pmid))
#' cat('\n')
#' while(j < length(pmid_chunks)){
#' #cat(unlist(pmid_chunks),"\n")
#' pmid_chunk <- pmid_chunks[[as.character(j)]]
#' #cat(unlist(pmid_chunk),"\n")
#' #cat('Processing chunk ',j,' with ',length(pmid_chunk),'PMIDS')
#' #cat('\n')
#' pmid_string <- paste(pmid_chunk,collapse = " ")
#' res <- RISmed::EUtilsGet(
#' RISmed::EUtilsSummary(pmid_string, type="esearch", db="pubmed", retmax = 5000)
#' )
#'
#'
#' year <- RISmed::YearPubmed(res)
#' authorlist <- RISmed::Author(res)
#' pmid_list <- RISmed::PMID(res)
#' i <- 1
#' first_author <- c()
#' while(i <= length(authorlist)){
#'
#'
#' if(length(authorlist[[i]]) == 5){
#' first_author <- c(
#' first_author,
#' paste(authorlist[[i]][1,]$LastName," et al.",sep=""))
#' }else{
#' first_author <- c(
#' first_author, as.character("Unknown et al.")
#' )
#' }
#' i <- i + 1
#' }
#' journal <- RISmed::ISOAbbreviation(res)
#' citations <- data.frame('pmid' = as.integer(pmid_list),
#' 'citation' = paste(first_author,year,journal,sep=", "),
#' stringsAsFactors = F)
#' citations$link <- paste0('<a href=\'https://www.ncbi.nlm.nih.gov/pubmed/',
#' citations$pmid,'\' target=\'_blank\'>',
#' citations$citation,'</a>')
#' all_citations <- dplyr::bind_rows(
#' all_citations, citations)
#' j <- j + 1
#' }
#'
#' return(all_citations)
#'
#' }
#'
#'
#' get_literature_links <- function(literature_df) {
#'
#' stopifnot(is.data.frame(literature_df))
#' assertable::assert_colnames(
#' literature_df, c("source","source_id","evidence_id"),
#' quiet = T)
#'
#' pubmed_source_ids <- literature_df |>
#' dplyr::filter(source == "PubMed")
#'
#' other_source_ids <- literature_df |>
#' dplyr::filter(source != "PubMed")
#'
#'
#' if (NROW(pubmed_source_ids) > 0) {
#' pubmed_source_ids <- get_citations_pubmed(
#' pmid = pubmed_source_ids$source_id, chunk_size = 100) |>
#' dplyr::rename(
#' name = citation,
#' ) |>
#' dplyr::mutate(source_id = as.character(pmid)) |>
#' dplyr::select(-pmid) |>
#' dplyr::left_join(
#' pubmed_source_ids, by = "source_id")
#' }
#'
#' if (NROW(other_source_ids) > 0) {
#' other_source_ids <- other_source_ids |>
#' dplyr::filter(source != "PubMed") |>
#' dplyr::mutate(name = source_id) |>
#' dplyr::mutate(link = dplyr::case_when(
#' source == "FDA" ~
#' "<a href='https://www.fda.gov/drugs/resources-information-approved-drugs/oncology-cancer-hematologic-malignancies-approval-notifications' target='_blank'>FDA approvals</a>",
#' source == "NCCN" ~
#' "<a href='https://www.nccn.org/guidelines/category_1' target='_blank'>NCCN guidelines</a>",
#' source == "EMA" ~ "<a href='https://www.ema.europa.eu/en/medicines/field_ema_web_categories%253Aname_field/Human' target='_blank'>European Medicines Agency (EMA)</a>",
#' source == "clinicaltrials.gov" ~ paste0(
#' "<a href='https://clinicaltrials.gov/ct2/show/",
#' source_id, "' target='_blank'>",
#' source_id, "</a>")
#' ))
#' }
#'
#' all_literature_links <- pubmed_source_ids |>
#' dplyr::select(source, source_id,
#' evidence_id,
#' dplyr::everything())
#'
#' if(NROW(other_source_ids) > 0){
#' all_literature_links <-
#' other_source_ids |>
#' dplyr::bind_rows(pubmed_source_ids)
#' }
#'
#' return(all_literature_links)
#'
#'
#' }
get_amino_acid_dictionary <- function() {
aa_df <-
data.frame('one_letter' = names(Biostrings::AMINO_ACID_CODE),
'three_letter' = Biostrings::AMINO_ACID_CODE,
'three_letter_uc' = toupper(Biostrings::AMINO_ACID_CODE),
stringsAsFactors = F) |>
dplyr::filter(one_letter != "B" &
one_letter != "J" &
one_letter != "O" &
one_letter != "U" &
one_letter != "X" &
one_letter != "Z")
rownames(aa_df) <- NULL
return(aa_df)
}
expand_hgvs_terms <- function(var, aa_dict, add_codon_markers = FALSE) {
hits <- c()
if (stringr::str_detect(toupper(var),"([A-Z]{3}[0-9]{1,}(([A-Z]{3}) {0,1})([0-9]{1,})?)") &
!stringr::str_detect(toupper(var),"(DEL|INS|DUP)|=|-|\\*")) {
aa <- stringr::str_to_title(strsplit(var, "[0-9]{1,}")[[1]])
codon <- stringr::str_match(var, "[0-9]{1,}")[[1]]
if (length(aa) == 1) {
alt1 <- paste0(aa[1],codon)
alt2 <- paste0('p.', alt1)
if (aa[1] != "Ter") {
if (aa[1] %in% aa_dict$three_letter) {
aa1 <- seqinr::a(aa[1])
if (!is.na(aa1)) {
alt3 <- paste0(seqinr::a(aa[1]), codon)
hits <- c(alt1, alt2, alt3)
}
}
}else{
alt3 <- paste0("X", codon)
hits <- c(alt1, alt2, alt3)
}
}else{
if (length(aa) == 2) {
alt1 <- stringr::str_replace(
stringr::str_replace(paste0(aa[1],codon, aa[2]),"(FS|Fs)$","fs"),
"ter","Ter")
alt2 <- stringr::str_replace(
stringr::str_replace(paste0('p.', alt1),"(FS|Fs)$","fs"),
"ter","Ter")
if (aa[1] != "Ter" &
aa[2] != "Ter" &
aa[2] != "Fs" &
nchar(aa[1]) == 3 &
nchar(aa[2]) == 3 &
aa[1] %in% aa_dict$three_letter &
aa[2] %in% aa_dict$three_letter) {
alt3 <- paste0(seqinr::a(aa[1]), codon, seqinr::a(aa[2]))
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
else{
if (aa[1] == "Ter") {
alt3 <- paste0("X", codon, seqinr::a(aa[2]))
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}else{
if (aa[2] == "Ter") {
if (aa[1] %in% aa_dict$three_letter) {
alt3 <- paste0(seqinr::a(aa[1]), codon, "X")
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
}
if (aa[2] == "Fs") {
if (aa[1] %in% aa_dict$three_letter) {
alt3 <- paste0(seqinr::a(aa[1]), codon, "fs")
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
}
}
}
}
}
}
if (stringr::str_detect(var,"^([A-Z]{1}[0-9]{1,}[A-Z]{0,1})$") &
!stringr::str_detect(var, "FS|DUP|DEL|INS|fs|dup|del|ins|\\*")) {
aa <- stringr::str_to_title(strsplit(var, "[0-9]{1,}")[[1]])
codon <- stringr::str_match(var, "[0-9]{1,}")[[1]]
if (length(aa) == 1) {
alt1 <- paste0(aa[1],codon)
alt2 <- paste0('p.', alt1)
if (aa[1] != "X" & aa[1] %in% aa_dict$one_letter) {
alt3 <- paste0(seqinr::aaa(aa[1]), codon)
hits <- c(alt1, alt2, alt3)
}else{
alt3 <- paste0("Ter", codon)
hits <- c(alt1, alt2, alt3)
}
}else{
if (length(aa) == 2) {
alt1 <- paste0(aa[1],codon, aa[2])
alt2 <- paste0('p.', alt1)
if (aa[1] != "X" &
aa[2] != "X" &
aa[1] %in% aa_dict$one_letter &
aa[2] %in% aa_dict$one_letter) {
alt3 <- paste0(seqinr::aaa(aa[1]),
codon, seqinr::aaa(aa[2]))
alt4 <- paste0('p.', alt3)
if(aa[2] == aa[1]){
alt5 <- paste0("p.", seqinr::aaa(aa[1]),
codon, "=")
alt6 <- paste0("p.", aa[1],
codon, "=")
hits <- c(alt1, alt2, alt3, alt4, alt5, alt6)
}else{
hits <- c(alt1, alt2, alt3, alt4)
}
}else{
if (aa[1] == "X" & aa[2] %in% aa_dict$one_letter) {
alt3 <- paste0("Ter",
codon, seqinr::aaa(aa[2]))
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}else{
if (aa[2] == "X" & aa[1] %in% aa_dict$one_letter) {
alt3 <- paste0(seqinr::aaa(aa[1]),
codon, "Ter")
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
}
}
}
}
}
if (stringr::str_detect(var,"^C\\.")) {
var <- stringr::str_replace(var,"^C\\.","c.")
#hits <- res
}
if (stringr::str_detect(var,"^P\\.")) {
var <- stringr::str_replace(var,"^P\\.","p.")
#hits <- res
}
if (stringr::str_detect(var,"FS$")) {
var <- stringr::str_replace(var,"FS$","fs")
}
if (stringr::str_detect(var,";") & stringr::str_detect(var,"\\(")) {
var <- tolower(var)
}
if (stringr::str_detect(var,"^([A-Z]{1}[0-9]{1,}[A-Z]{1})$")) {
res1 <- paste0('p.', var)
if(add_codon_markers == TRUE){
res2 <- stringr::str_match(var,"^([A-Z]{1}[0-9]{1,})")[[1]]
hits <- c(hits,res1,res2)
}else{
hits <- c(hits, res1)
}
}
if (stringr::str_detect(var,"FS$")) {
res <- stringr::str_replace(var,"FS$","fs")
hits <- res
}
if (stringr::str_detect(var,"(DEL|Del|INS|Ins|Dup|DUP)") &
!stringr::str_detect(var, "INSERTION|DUPLICATION|NM_|NP_|DELETERIOUS|DELETION|MUTATION|Mutation|Deletion|")) {
var <- stringr::str_replace(var,"(DEL|Del)","del")
var <- stringr::str_replace(var,"(INS|Ins)","ins")
var <- stringr::str_replace(var,"(DUP|Dup)","dup")
if (!(stringr::str_detect(var,"^(c|p|rs|;|[0-9])"))) {
res2 <- paste0('p.', var)
hits <- c(var, res2)
}
}
if (length(hits) == 0 & var != "") {
if (!(stringr::str_detect(var,"^(c|rs|p|[0-9])")) &
stringr::str_detect(var, "[0-9]{1,}") &
!stringr::str_detect(
var,"EXON|;|MUTATION|Mutation|-|NM_|NP_|DELETERIOUS|Deletion|INTRON|DELETION|DUPLICATION")) {
if (stringr::str_detect(var,"FS$|DUP|DELINS|FS\\*")) {
var <- stringr::str_replace(var, "FS","fs")
var <- stringr::str_replace(var, "DUP","dup")
var <- stringr::str_replace(var, "DELINS","delins")
var <- stringr::str_replace(var, "DEL","del")
}
res1 <- paste0('p.',var)
hits <- c(var, res1)
}else{
hits <- c(var)
}
}
hits <- unique(hits)
return(hits)
}
load_civic_biomarkers <- function(
datestamp = '20240709',
compound_synonyms = NULL,
hg38_fasta =
"/Users/sigven/research/DB/hg38/hg38.fa",
crossmap_chainfile_dir =
"/Users/sigven/research/DB/chainFiles",
cache_dir = NA) {
gene_alias <- geneOncoX::get_alias(
cache_dir = cache_dir
)
civic_local_fnames <- list()
for(e in c("VariantSummaries",
"civic_accepted",
"ClinicalEvidenceSummaries",
"MolecularProfileSummaries")){
civic_local_fnames[[e]] <- file.path(
cache_dir,
paste0("civic_", e, "_", datestamp,".tsv")
)
remote_url <- paste0(
"https://civicdb.org/downloads/nightly/nightly-", e, ".tsv")
if(e == "civic_accepted"){
remote_url <- paste0(
"https://civicdb.org/downloads/nightly/nightly-", e, ".vcf")
civic_local_fnames[['vcf_grch37']] <- file.path(
cache_dir,
paste0("civic_", datestamp,"_grch37.vcf")
)
civic_local_fnames[['vcf_grch38']] <- file.path(
cache_dir,
paste0("civic_", datestamp,"_grch38.vcf")
)
}
if(e == "civic_accepted"){
if(!file.exists(civic_local_fnames[['vcf_grch37']])){
download.file(
url = remote_url,
destfile = civic_local_fnames[['vcf_grch37']])
}
}else{
if(!file.exists(civic_local_fnames[[e]])){
download.file(
url = remote_url,
destfile = civic_local_fnames[[e]])
}
}
if(e == 'civic_accepted'){
vcfhelpR::crossmap_vcf(
direction = "hg19Tohg38",
chain_file_directory = crossmap_chainfile_dir,
target_genome_file = hg38_fasta,
source_vcf = civic_local_fnames[['vcf_grch37']],
target_vcf = civic_local_fnames[['vcf_grch38']])
system(paste0('bgzip -dc ',civic_local_fnames[['vcf_grch38']],
'.gz > ',
civic_local_fnames[['vcf_grch38']]))
}
}
aa_dict <- get_amino_acid_dictionary()
variant_coordinates <- data.frame()
for(assembly in c('grch37','grch38')){
skip_lines <- 8
if(assembly == "grch38"){
skip_lines <- 7
}
variant_coordinates <- as.data.frame(
dplyr::bind_rows(
variant_coordinates,
readr::read_tsv(
file = civic_local_fnames[[paste0('vcf_',assembly)]],
skip = skip_lines,
show_col_types = F,
col_names = T,
guess_max = 1000) |>
janitor::clean_names() |>
dplyr::rename(variant_id = id) |>
dplyr::mutate(
vcf_coord = paste0(
assembly, ":",
paste(number_chrom, pos, ref, alt, sep="_"))
) |>
dplyr::select(variant_id, vcf_coord)
)
)
}
## ignore these biomarkers
biomarker_names_skip_regex <-
paste0(
"methylat|phosphoryl|serum lev|mislocali|peri-therap|",
"alternative|cytoplasmic|p-loop|reg_|alternative|polymorphism|",
"loss of heterozygosity|philadelphia|internal|n-terminal")
molecularProfileSummary <- as.data.frame(
data.table::fread(
civic_local_fnames[['MolecularProfileSummaries']],
select = c(1:13), fill = T)) |>
dplyr::rename(variant_id = variant_ids,
molecular_profile_name = name,
molecular_profile_summary = summary) |>
dplyr::select(variant_id,
molecular_profile_id,
molecular_profile_name,
molecular_profile_summary) |>
## ignore profiles with aberration combinations for now
tidyr::separate_rows(variant_id, sep = ",") |>
##dplyr::filter(!stringr::str_detect(variant_id, ",")) |>
dplyr::mutate(variant_id = as.integer(variant_id)) |>
clean_variant_naming(col = "molecular_profile_name") |>
dplyr::filter(!stringr::str_detect(
tolower(molecular_profile_name),
biomarker_names_skip_regex))
clinicalEvidenceSummary <- as.data.frame(
readr::read_tsv(
civic_local_fnames[['ClinicalEvidenceSummaries']],
show_col_types = F, guess_max = 5000)) |>
dplyr::filter(source_type == "PubMed") |>
dplyr::mutate(citation_id = as.character(citation_id)) |>
dplyr::select(molecular_profile_id,
evidence_id,
evidence_civic_url,
disease,
doid,
therapies,
citation_id,
evidence_statement,
evidence_type,
evidence_level,
evidence_direction,
significance,
variant_origin,
rating) |>
dplyr::mutate(citation_id = dplyr::case_when(
citation_id == "27577079" ~ "28947956",
citation_id == "27556863" ~ "28978004",
TRUE ~ as.character(citation_id)
)) |>
dplyr::mutate(disease = dplyr::if_else(
is.na(disease),
"Cancer",
as.character(disease)
)) |>
dplyr::mutate(doid = dplyr::if_else(
is.na(doid),
"162",
as.character(doid)
)) |>
dplyr::rename(disease_ontology_id = doid,
clinical_significance = significance,
evidence_description = evidence_statement,
evidence_url = evidence_civic_url,
cancer_type = disease) |>
dplyr::mutate(
clinical_significance =
dplyr::case_when(
clinical_significance == "Resistance" |
clinical_significance == "Reduced Sensitivity" ~ "Resistance/Non-response",
TRUE ~ as.character(clinical_significance))) |>
dplyr::mutate(disease_ontology_id = paste0(
"DOID:", disease_ontology_id)) |>
dplyr::mutate(
evidence_level =
dplyr::case_when(
evidence_level == 'A' ~ 'A: Validated',
evidence_level == 'B' ~ 'B: Clinical evidence',
evidence_level == 'C' ~ 'C: Case study',
evidence_level == 'D' ~ 'D: Preclinical evidence',
evidence_level == 'E' ~ 'E: Indirect evidence',
TRUE ~ as.character(evidence_level)
)
) |>
dplyr::mutate(
variant_origin =
dplyr::if_else(
variant_origin == "N/A" |
variant_origin == "Unknown",
as.character(NA),
as.character(
stringr::str_replace(
variant_origin,"Rare |Common ",""))))
therapeutic_contexts <- as.data.frame(
clinicalEvidenceSummary |>
dplyr::select(evidence_id, therapies) |>
dplyr::filter(!is.na(therapies) &
nchar(therapies) > 0) |>
tidyr::separate_rows(therapies, sep = ",") |>
dplyr::mutate(therapies = tolower(therapies)) |>
dplyr::left_join(
dplyr::select(compound_synonyms,
alias_lc,
drug_name,
molecule_chembl_id),
by = c("therapies" = "alias_lc"),
multiple = "all",
relationship = "many-to-many"
) |>
dplyr::filter(!is.na(drug_name)) |>
dplyr::select(evidence_id, molecule_chembl_id) |>
dplyr::distinct() |>
dplyr::group_by(evidence_id) |>
dplyr::summarise(molecule_chembl_id = paste(
unique(sort(molecule_chembl_id)), collapse = "|"),
.groups = "drop"
) |>
dplyr::ungroup() |>
dplyr::mutate(molecule_chembl_id = stringr::str_replace(
molecule_chembl_id,"^NA\\||\\|NA$",""
))
)
clinicalEvidenceSummary <- clinicalEvidenceSummary |>
dplyr::left_join(
therapeutic_contexts,
by = "evidence_id", multiple = "all",
relationship = "many-to-many")
variantSummary <- as.data.frame(
data.table::fread(civic_local_fnames[['VariantSummaries']],
select = c(1:25), fill = T)) |>
## Ignore the "Factor" feature type (Kataegis, CK, methylation)
dplyr::filter(feature_type != "Factor") |>
#dplyr::rename(molecular_profile_id = variant_id) |>
dplyr::mutate(alteration_type = "MUT") |>
dplyr::mutate(variant = stringr::str_trim(variant)) |>
dplyr::mutate(variant = dplyr::if_else(
!is.na(variant) &
(variant == "Fusion" | '::' %in% variant),
as.character(feature_name),
as.character(variant)
)) |>
dplyr::mutate(variant_aliases = dplyr::if_else(
gene == "BRAF" & variant == "V600",
paste(variant_aliases, "V640", sep = ","),
as.character(variant_aliases)
)) |>
dplyr::mutate(variant_aliases = dplyr::if_else(
!is.na(hgvs_descriptions) &
hgvs_descriptions != "NA" &
nchar(hgvs_descriptions) > 0,
paste(variant_aliases, hgvs_descriptions, sep=","),
as.character(variant_aliases))) |>
dplyr::select(-c("hgvs_descriptions")) |>
dplyr::mutate(variant_aliases = stringr::str_replace(
variant_aliases, "^,", ""
)) |>
tidyr::separate_rows(variant_aliases, sep = ",") |>
tidyr::separate_rows(variant, sep=",") |>
clean_variant_naming(col = "variant") |>
clean_variant_naming(col = "variant_aliases") |>
dplyr::rename(variant_consequence = variant_types,
molecular_profile_id =
single_variant_molecular_profile_id,
variant_alias = variant_aliases) |>
clean_variant_consequence() |>
#dplyr::filter(variant != "Alu insertion") |>
dplyr::filter(
!stringr::str_detect(
tolower(variant),
biomarker_names_skip_regex)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
feature_type == "Fusion" &
nchar(variant_alias) == 0 |
stringr::str_detect(
variant_alias,"multiple partners"),
paste0(
stringr::str_replace(
feature_name,"v::|::v","")," Fusion"),
as.character(variant_alias)
)) |>
dplyr::mutate(variant = dplyr::if_else(
nchar(variant_alias) > 0 &
feature_type == "Fusion" &
stringr::str_detect(variant,"^e[0-9]{1,}::"),
stringr::str_replace(variant_alias," E[0-9]{1,}-E[0-9]{1,}$",""),
as.character(variant))) |>
dplyr::mutate(gene = dplyr::if_else(
feature_type == "Fusion",
as.character(variant),
as.character(gene)
)) |>
set_alteration_type() |>
dplyr::select(variant_id,
molecular_profile_id,
variant_consequence,
variant,
alteration_type,
gene,
variant_alias) |>
dplyr::mutate(
variant = stringr::str_replace_all(
variant,
paste0(
"((Inactivating|Activating|Deleterious|",
"Promoter|Truncating|TKD) Mutation)|Loss-of-function"),
"Mutation"
)
) |>
dplyr::mutate(variant_alias = paste(
variant_alias, variant, sep=","
)) |>
tidyr::separate_rows(
variant_alias, sep=","
) |>
dplyr::filter(
!stringr::str_detect(
tolower(variant_alias),
biomarker_names_skip_regex)) |>
dplyr::filter(
(!(stringr::str_detect(variant,"H$") &
stringr::str_detect(toupper(variant_alias),"TYR$"))) &
(!(stringr::str_detect(variant,"Y$") &
stringr::str_detect(toupper(variant_alias),"HIS$")))
) |>
dplyr::distinct() |>
dplyr::mutate(
variant_name_primary = dplyr::case_when(
alteration_type == "CNA" |
(alteration_type == "MUT_LOF" & !endsWith(variant,"X")) |
stringr::str_detect(
variant,
"Mutation|Wildtype|Splice|Copy Number|Deletion|Amplification") |
stringr::str_detect(alteration_type,"^EXP") ~
paste0(gene," ",stringr::str_to_title(variant)),
variant_consequence == "transcript_fusion" |
(nchar(variant_consequence) == 0 & variant == "FUSION") ~
paste0(gene, " Fusion"),
variant == "REARRANGEMENT" ~
paste0(gene, " Rearrangement"),
TRUE ~ as.character(
paste0(gene, " ", variant)
))
) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant,"^aa_region:"),
as.character("missense_variant"),
as.character(variant_consequence))) |>
dplyr::filter(!nchar(variant_alias) == 0) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "FS","fs")
) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "DELINS","delins")
) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "DEL","del")
) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "INS$","ins")
)
variants_expanded <- list()
variants_expanded[['mut']] <- data.frame()
for (i in unique(variantSummary$variant_id)) {
vrows <- variantSummary[variantSummary$variant_id == i,]
variant_entries <- unique(c(vrows$variant,
vrows$variant_alias))
gene <- unique(vrows$gene)
variant_consequence <- unique(vrows$variant_consequence)
alteration_type <- unique(vrows$alteration_type)
variant_name_primary <- unique(vrows$variant_name_primary)
all_aliases <- c()
if (stringr::str_detect(alteration_type,"MUT")) {
for (e in variant_entries) {
#cat(e,'\n')
clean_aliases <- expand_hgvs_terms(e, aa_dict = aa_dict)
for (c in clean_aliases) {
all_aliases <- unique(c(all_aliases, c))
}
}
for (variant_alias in all_aliases) {
df <- data.frame(
'variant_id' = i,
'symbol' = gene,
'variant_name_primary' = variant_name_primary,
'variant_alias' = variant_alias,
'alteration_type' = alteration_type,
'variant_consequence' = variant_consequence,
stringsAsFactors = F)
variants_expanded[['mut']] <- dplyr::bind_rows(
variants_expanded[['mut']], df
)
}
}
}
variants_expanded[['mut']] <- variants_expanded[['mut']] |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
variant_alias,
"^(p\\.)?(([A-Z]{1}[0-9]{1,})|([A-Z]{1}[a-z]{2}[0-9]{1,}))$"),
as.character("CODON"),
as.character(alteration_type)
)) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"FS"),
stringr::str_replace(variant_alias,"FS","fs"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"DEL") &
!stringr::str_detect(variant_alias, "DELETION|DELETERIOUS"),
stringr::str_replace(variant_alias,"DEL","del"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"INS") &
!stringr::str_detect(variant_alias,"INSERTION"),
stringr::str_replace(variant_alias,"INS","ins"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"TER|ter") &
!stringr::str_detect(
tolower(variant_alias),
"alteration"),
stringr::str_replace(variant_alias,"ter|TER","Ter"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"DUP") &
!stringr::str_detect(variant_alias,"DUPLICATION"),
stringr::str_replace(variant_alias,"DUP","dup"),
as.character(variant_alias)
)
) |>
dplyr::filter(
variant_alias != "p.=" &
!stringr::str_detect(variant_alias, "^p\\.\\(") &
!stringr::str_detect(variant_alias, "^p\\.Exon") &
!stringr::str_detect(variant_alias, "ENSP|ENST") &
!stringr::str_detect(variant_alias, "SPLICE"))
aa_dict <- get_amino_acid_dictionary()
for (i in 1:nrow(aa_dict)) {
aa <- aa_dict[i,"three_letter_uc"]
aa_three_lc <- aa_dict[i, "three_letter"]
if (nrow(variants_expanded[['mut']][!is.na(variants_expanded[['mut']]$variant_alias) &
stringr::str_detect(variants_expanded[['mut']]$variant_alias,aa),]) > 0) {
variants_expanded[['mut']][!is.na(variants_expanded[['mut']]$variant_alias) &
stringr::str_detect(variants_expanded[['mut']]$variant_alias, aa),]$variant_alias <-
stringr::str_replace_all(variants_expanded[['mut']][!is.na(variants_expanded[['mut']]$variant_alias) &
stringr::str_detect(variants_expanded[['mut']]$variant_alias, aa),]$variant_alias,
aa,
aa_three_lc)
}
}
variants_expanded[['mut']] <- variants_expanded[['mut']] |>
dplyr::mutate(alias_type = "hgvsp") |>
dplyr::mutate(variant_alias = stringr::str_trim(variant_alias)) |>
dplyr::mutate(alias_type = dplyr::case_when(
stringr::str_detect(variant_alias, "^rs[0-9]{1,}$") ~ "dbsnp",
stringr::str_detect(variant_alias, "^aa_region") ~ "aa_region",
stringr::str_detect(variant_alias, "^c\\.[0-9]{1,}") ~ "hgvsc",
stringr::str_detect(variant_alias, "^Exon ") ~ "exon",
stringr::str_detect(
variant_alias,
"Mutation|Copy|Wildtype|Loss|Truncation|Alteration|Frameshift") ~ "other_gene",
TRUE ~ as.character(alias_type)
)) |>
dplyr::left_join(variant_coordinates,
by = "variant_id",
relationship = "many-to-many") |>
dplyr::mutate(variant_alias = dplyr::if_else(
!is.na(vcf_coord),
paste(variant_alias, vcf_coord, sep=","),
as.character(variant_alias)
)) |>
tidyr::separate_rows(
variant_alias, sep = ","
) |>
dplyr::select(-vcf_coord) |>
dplyr::distinct() |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch37"),
"genomic_grch37",
as.character(alias_type)
)) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch38"),
"genomic_grch38",
as.character(alias_type)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch37"),
stringr::str_replace(variant_alias, "grch37:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch38"),
stringr::str_replace(variant_alias, "grch38:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(alteration_type = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic"),
"MUT",
as.character(alteration_type)
)) |>
dplyr::filter(
symbol != "TERT" |
(symbol == "TERT" &
!stringr::str_detect(variant_alias, "Cys|^p\\.") &
alteration_type != "CODON")) |>
dplyr::mutate(variant_alias = dplyr::if_else(
symbol == "TERT" & variant_alias == "-146C>T",
paste(variant_alias,
"grch37:5_1295250_G_A,grch38:5_1295135_G_A",
sep=","),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
symbol == "MET" &
variant_alias == "Exon 14 Skipping Mutation",
paste(
variant_alias,
"c.3028+1G>T",
"c.3028+1G>C",
"c.3028+1G>A",
"c.3028+2T>C",
"c.3028+2T>G",
"c.3082+1G>T",
"c.3082+1G>C",
"c.3082+1G>A",
"c.3082+2T>C",
"c.3082+2T>G",
sep=","),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
symbol == "TERT" &
variant_alias == "-124C>T",
paste(variant_alias,
"grch37:5_1295228_G_A,grch38:5_1295113_G_A",
sep=","),
as.character(variant_alias)
)) |>
tidyr::separate_rows(
variant_alias, sep = ","
) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch37"),
"genomic_grch37",
as.character(alias_type)
)) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch38"),
"genomic_grch38",
as.character(alias_type)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch37"),
stringr::str_replace(variant_alias, "grch37:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch38"),
stringr::str_replace(variant_alias, "grch38:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(alteration_type = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic"),
"MUT",
as.character(alteration_type)
)) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^c\\."),
"hgvsc",
as.character(alias_type)
)) |>
dplyr::mutate(variant_exon = dplyr::if_else(
stringr::str_detect(variant_alias, "Exon"),
stringr::str_replace_all(
variant_alias,
"Exon |( (Frameshift|((Skipping )?Mutation)|Deletion|Insertion))",""),
as.character(NA)
)) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,"18-21",
"18|19|20|21"),
as.character(variant_exon))
) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,
"1-2","1|2"),
as.character(variant_exon))
) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,
"10 OR 21","10|21"),
as.character(variant_exon))
) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,
"10 AND 21","10&21"),
as.character(variant_exon))
) |>
dplyr::filter(
is.na(variant_exon)
| (variant_exon != "9 AND 20" &
!stringr::str_detect(
variant_exon,"Deletion"))) |>
dplyr::distinct()
## Make sure alteration types for genomic coordinates
## are correct
variants_nongenomic <- variants_expanded[['mut']] |>
dplyr::filter(
!stringr::str_detect(
alias_type, "genomic_"
)
) |>
dplyr::group_by(
variant_id
) |>
dplyr::summarise(
alteration_type =
paste(unique(alteration_type), collapse=";")
) |>
dplyr::mutate(
alteration_type = dplyr::if_else(
stringr::str_detect(alteration_type, ";"),
stringr::str_replace(
alteration_type, ";?CODON;?",""
),
as.character(alteration_type)
)
)
variants_expanded_mut1 <- variants_expanded[['mut']] |>
dplyr::filter(!stringr::str_detect(
alias_type, "genomic"))
variants_expanded_mut2 <- variants_expanded[['mut']] |>
dplyr::filter(stringr::str_detect(
alias_type, "genomic"))
variants_expanded_mut2$alteration_type <- NULL
variants_expanded_mut2 <- variants_expanded_mut2 |>
dplyr::left_join(variants_nongenomic, by = "variant_id",
relationship = "many-to-many")
variants_expanded[['mut']] <- variants_expanded_mut1 |>
dplyr::bind_rows(variants_expanded_mut2)
## Combine the two dataframes
variants_expanded[['other']] <- as.data.frame(
variantSummary |>
dplyr::filter(!stringr::str_detect(alteration_type,"MUT")) |>
dplyr::select(variant_id, variant, variant_consequence,
variant_alias, variant_name_primary, alteration_type,
gene) |>
dplyr::rename(symbol = gene) |>
dplyr::mutate(
alias_type = "other_gene"
) |>
dplyr::distinct()
)
attr(variants_expanded[['other']],"groups") <- NULL
gene_aliases <- gene_alias$records |>
dplyr::filter(.data$n_primary_map == 1 &
source == "NCBI") |>
dplyr::select(alias, symbol, entrezgene) |>
dplyr::distinct()
biomarker_items <- list()
biomarker_items[['variant']] <-
dplyr::bind_rows(
variants_expanded[['mut']],
variants_expanded[['other']]) |>
dplyr::select(variant_id, variant_alias,
alias_type, dplyr::everything()) |>
dplyr::rename(gene = symbol) |>
dplyr::left_join(
gene_aliases,
by = c("gene" = "alias"),
relationship = "many-to-one"
) |>
dplyr::mutate(variant_consequence = dplyr::case_when(
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "delins([A-Z]{1,4})+$") ~ "inframe_insertion",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "ins([A-Z]{1,4})+$") ~ "inframe_insertion",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "[0-9]{1,}?del([A-Z]{1,4})+$") ~ "inframe_deletion",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "fs") ~ "frameshift_variant",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "^Mutation$") ~ "protein_altering_variant",
nchar(variant_consequence) == 0 &
alias_type == "hgvsp" ~ "missense_variant",
TRUE ~ as.character(variant_consequence)
)) |>
dplyr::mutate(variant_alias = dplyr::case_when(
(is.na(variant_alias) | nchar(variant_alias) == 0) &
!is.na(gene) &
alias_type == "other_gene" &
stringr::str_detect(gene,"-") ~ gene,
(is.na(variant_alias) | nchar(variant_alias) == 0) &
!is.na(gene) &
alias_type == "other_gene" &
!stringr::str_detect(gene,"-") ~
paste0(gene, " Fusion"),
TRUE ~ as.character(variant_alias)
)) |>
dplyr::distinct()
biomarker_items[['clinical']] <-
clinicalEvidenceSummary |>
dplyr::full_join(
molecularProfileSummary, by = "molecular_profile_id",
multiple = "all", relationship = "many-to-many") |>
dplyr::select(molecular_profile_id, molecular_profile_name,
molecular_profile_summary,
variant_id, dplyr::everything()) |>
dplyr::rename(therapeutic_context = therapies) |>
dplyr::mutate(evidence_id = paste0("EID",evidence_id)) |>
dplyr::filter(evidence_id != "EIDNA") |>
dplyr::mutate(biomarker_source = "civic",
biomarker_source_datestamp = datestamp,
biomarker_entity = T) |>
dplyr::inner_join(dplyr::select(
biomarker_items[['variant']], variant_id),
by = "variant_id",
multiple = "all", relationship = "many-to-many") |>
dplyr::distinct() |>
dplyr::arrange(evidence_level, molecular_profile_id)
biomarker_items[['clinical']] <- map_biomarker_phenotypes(
biomarker_items[['clinical']],
cache_dir = cache_dir)
biomarker_items <- get_literature_references(biomarker_items)
if(is.data.frame(biomarker_items)){
if(unique(colnames(biomarker_items) == c("source","source_id")) == T){
lgr::lgr$error("literature missing ")
return(biomarker_items)
}
}
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::mutate(molecular_profile_type = "Single") |>
dplyr::mutate(molecular_profile_type = dplyr::case_when(
stringr::str_detect(molecular_profile_name, " AND|and|And ") ~
paste0(
"Combo_", stringr::str_count(variant_id, " AND|and|And ")[1] + 1),
stringr::str_detect(molecular_profile_name, " OR|or|Or ") ~
paste0(
"Combo_Opt_", stringr::str_count(variant_id, ",")[1] + 1),
TRUE ~ molecular_profile_type
)) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
biomarker_entity,
molecular_profile_id,
molecular_profile_name,
molecular_profile_summary,
molecular_profile_type,
variant_id,
source_id,
evidence_id,
variant_origin,
primary_site,
cancer_type,
disease_ontology_id,
do_name,
efo_id,
efo_name,
cui,
cui_name,
dplyr::everything()
)
biomarker_items[['variant']] <- as.data.frame(
biomarker_items[['variant']] |>
dplyr::mutate(biomarker_source = "civic",
biomarker_source_datestamp = datestamp) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
variant_name_primary,
alteration_type,
alias_type,
variant_consequence,
variant_exon,
gene,
symbol,
entrezgene,
dplyr::everything()
) |>
dplyr::inner_join(
dplyr::select(biomarker_items[['clinical']],
variant_id, molecular_profile_id),
by = "variant_id",
relationship = "many-to-many"
) |>
dplyr::distinct() |>
dplyr::select(-c("variant")) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
variant_id,
molecular_profile_id,
dplyr::everything()
)
)
return(biomarker_items)
}
load_cgi_biomarkers <- function(compound_synonyms = NULL,
datestamp = '20221027',
cache_dir = NA) {
aa_dict <- get_amino_acid_dictionary()
gene_alias <- geneOncoX::get_alias(
cache_dir = cache_dir
)
cancer_type_abbreviations <-
read.table(
file = file.path(cache_dir, "cgi_cancer_subtypes.tsv"),
sep = "\t", header = T,
stringsAsFactors = F, quote = "",
na.strings = "",fill = T) |>
dplyr::mutate(
disease_ontology_id = paste0(
"DOID:",disease_ontology_id))
## SNVs/InDels
## SKIP/IGNORE mutations that are
# 1) not explicitly mapped to protein alterations (MUT)
# 2) marked with uncertainty, rechecking etc (Comments)
# 3) involved in toxicity
# 4) curated from abstracts etc (not PMIDs or guidelines)
# 5) Non-malignant/Non-cancer conditions
## NEW
cgi_biomarkers <- as.data.frame(
readr::read_tsv(
file.path(cache_dir, "cgi_biomarkers_20221017.tsv"),
col_names = T, show_col_types = F, guess_max = 1000) |>
janitor::clean_names() |>
dplyr::mutate(
biomarker =
stringr::str_replace_all(
biomarker,
"\\.",""
)) |>
dplyr::select(-c(curator, curation_date, targeting,
drug_family, drug_status, drug, tcgi_included,
metastatic_tumor_type,
primary_tumor_type_full_name, assay_type)) |>
dplyr::distinct() |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "12,13,59,61,117,146",
"G12,G13,A59,Q61,K117,A146")) |>
dplyr::mutate(alteration_type = stringr::str_replace(
alteration_type, "EXPR", "EXP"
)) |>
dplyr::mutate(alteration_type = stringr::str_replace(
alteration_type, "FUS", "FUSION"
)) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "\\*","X")) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "618,620,634,768,791,891,918",
"C618,C620,C634,E768,Y791,S891,M918")) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "12,13","G12,G13")) |>
dplyr::mutate(biomarker = stringr::str_replace(
biomarker, "undexpression","Underexpression")) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, ",537,538,",",Y537,D538,")) |>
dplyr::rename(
molecular_profile = biomarker
) |>
dplyr::mutate(molecular_profile = stringr::str_replace_all(
molecular_profile, ";|(, )", ","
)) |>
dplyr::mutate(molecular_profile = stringr::str_replace_all(
molecular_profile, " (\\+|-)$", ""
)) |>
dplyr::filter(
!stringr::str_detect(source, "^(ENA|ESMO|ASCO|AACR)")
) |>
tidyr::separate_rows(
primary_tumor_type, sep = ";") |>
dplyr::left_join(
cancer_type_abbreviations,
by = c("primary_tumor_type" = "cgi_abbreviation"),
relationship = "many-to-many") |>
dplyr::mutate(molecular_profile_id = dplyr::row_number()) |>
dplyr::mutate(molecular_profile_type = dplyr::if_else(
stringr::str_detect(molecular_profile, " \\+ "),
"Combo",
"Single"
)) |>
dplyr::mutate(molecular_profile_name = molecular_profile) |>
tidyr::separate_longer_delim(molecular_profile, delim = " + ")
)
combo_index <- cgi_biomarkers |>
dplyr::group_by(molecular_profile_id) |>
dplyr::reframe(combo_index = dplyr::row_number()) |>
dplyr::select(-molecular_profile_id)
cgi_biomarkers <- as.data.frame(
cgi_biomarkers |>
dplyr::bind_cols(combo_index) |>
dplyr::mutate(molecular_profile_type = dplyr::if_else(
molecular_profile_type == "Combo",
paste(
molecular_profile_type, combo_index, sep="_"),
as.character(molecular_profile_type)
)) |>
dplyr::select(-combo_index) |>
dplyr::mutate(alteration2 =
stringr::str_replace_all(
stringr::str_match(
molecular_profile, "\\([A-Z0-9]+(,[A-Z0-9]+){0,}\\)")[,1],
"\\)|\\(","")
) |>
dplyr::mutate(alteration_type = dplyr::case_when(
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " amplification") ~ "CNA",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " deletion") &
!stringr::str_detect(molecular_profile, "inframe") ~ "CNA",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " fusion") ~ "FUSION",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(tolower(molecular_profile), "overexpression") ~ "EXP_OVER",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(tolower(molecular_profile), "underexpression") ~ "EXP_OVER",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " expression") ~ "EXP",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(alteration_type,"MUT") &
stringr::str_detect(tolower(molecular_profile), "oncogenic|\\(") ~ "MUT_ONC",
TRUE ~ as.character(alteration_type)
)) |>
tidyr::separate(gene, c("gene1","gene2"), sep = ";") |>
dplyr::mutate(gene = dplyr::if_else(
!is.na(gene1) & !is.na(gene2) &
startsWith(molecular_profile, gene2),
as.character(gene2),
as.character(gene1)
)) |>
dplyr::select(-c(gene1,gene2)) |>
dplyr::select(molecular_profile, gene, dplyr::everything()) |>
dplyr::mutate(alteration_type = dplyr::case_when(
alteration_type == "BIA" ~ "MUT_LOF_BIALLELIC",
stringr::str_detect(
molecular_profile, " wildtype") ~ "WT",
stringr::str_detect(
molecular_profile, " oncogenic") ~ "MUT_ONC",
stringr::str_detect(
tolower(molecular_profile), "overexpression") ~ "EXP_OVER",
stringr::str_detect(
tolower(molecular_profile), "underexpression") ~ "EXP_UNDER",
stringr::str_detect(
molecular_profile, " fusion$") ~ "FUSION",
TRUE ~ as.character(alteration_type)
)) |>
dplyr::mutate(variant_consequence = dplyr::case_when(
alteration_type == "WT" ~ as.character(NA),
alteration_type == "CNA" &
stringr::str_detect(
alteration,":amp") ~ "transcript_amplification",
alteration_type == "CNA" &
stringr::str_detect(
alteration,":del") ~ "transcript_ablation",
alteration_type == "EXP_OVER" ~ "transcript_amplification",
alteration_type == "FUSION" ~ "transcript_fusion",
alteration_type == "EXP_UNDER" ~ "transcript_ablation",
alteration_type == "MUT_LOF_BIALLELIC" ~ "loss_of_function_variant",
alteration_type == "MUT_ONC" ~ "protein_altering_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
!stringr::str_detect(molecular_profile,"inframe|splice|insertion|deletion") ~ "missense_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"splice donor") ~ "splice_donor_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"splice acceptor") ~ "splice_acceptor_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"inframe") &
stringr::str_detect(molecular_profile,"insertion") &
stringr::str_detect(molecular_profile, "deletion") ~ "inframe_deletion,inframe_insertion",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"inframe") &
stringr::str_detect(molecular_profile, "deletion") ~ "inframe_deletion",
molecular_profile == "FLT3-ITD" ~ "inframe_insertion",
molecular_profile == "PDGFRA exon 18 mutations" ~ "inframe_deletion,inframe_insertion,missense_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"inframe") &
stringr::str_detect(molecular_profile, "insertion") ~ "inframe_insertion",
(alteration_type == "MUT") &
stringr::str_detect(molecular_profile, " exon ") &
stringr::str_detect(alteration, "inframe_insertion") ~ "inframe_insertion",
(alteration_type == "MUT") &
stringr::str_detect(molecular_profile, " exon ") &
stringr::str_detect(alteration, "inframe_deletion") ~ "inframe_deletion",
TRUE ~ as.character(NA)
)) |>
dplyr::rename(
variant_alias = alteration2
) |>
dplyr::mutate(variant_alias = dplyr::case_when(
alteration_type == "EXP" ~ "Expression",
stringr::str_detect(molecular_profile, "wildtype") ~ "Wildtype",
is.na(variant_alias) &
(alteration_type == "MUT" |
alteration_type == "MUT_ONC" |
alteration_type == "MUT_LOF_BIALLELIC") ~ "Mutation",
alteration_type == "CNA" &
variant_consequence == "transcript_amplification" ~ "Amplification",
alteration_type == "CNA" &
variant_consequence == "transcript_ablation" ~ "Deletion",
alteration_type == "FUSION" ~ "Fusion",
alteration_type == "EXP_UNDER" &
variant_consequence == "transcript_ablation" ~ "Underexpression",
alteration_type == "EXP_OVER" &
variant_consequence == "transcript_amplification" ~ "Overexpression",
TRUE ~ as.character(variant_alias)
)) |>
dplyr::mutate(gene = dplyr::if_else(
alteration_type == "FUSION" &
stringr::str_detect(molecular_profile, "-"),
stringr::str_replace(
stringr::str_replace(molecular_profile, " fusion",""),
"ABL1-BCR","BCR-ABL1"),
as.character(gene))
) |>
dplyr::mutate(gene = dplyr::case_when(
gene == "MLL2" ~ "KMT2D",
TRUE ~ as.character(gene)
)) |>
dplyr::mutate(variant_origin = "Somatic") |>
dplyr::mutate(variant_origin = dplyr::if_else(
!is.na(comments) &
stringr::str_detect(tolower(comments),"germline"),
"Germline",
as.character(variant_origin)
)) |>
dplyr::mutate(evidence_type = 'Predictive',
therapeutic_context = drug_full_name,
clinical_significance = NA) |>
dplyr::mutate(
clinical_significance =
dplyr::case_when(
association == "Resistant" ~ "Resistance/Non-response",
association == "No Responsive" ~ "Resistance/Non-response",
association == "Responsive" ~ "Sensitivity/Response",
stringr::str_detect(association, "Toxicity") ~ "Toxicity",
TRUE ~ as.character("Other"))) |>
dplyr::rename(evidence_level_raw = evidence_level) |>
dplyr::mutate(mapping_rank = 1) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"guidelines"),
"A: FDA/NCCN/ELN guidelines",
as.character(NA))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Late trials"),
"B1: Clinical evidence: late trials",
as.character(evidence_level))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Early trials"),
"B2: Clinical evidence: early trials",
as.character(evidence_level))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Case report"),
"C: Case study",
as.character(evidence_level))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Pre-clinical"),
"D: Preclinical evidence",
as.character(evidence_level))) |>
dplyr::filter(!is.na(evidence_level_raw)) |>
dplyr::select(-comments) |>
dplyr::mutate(alteration = dplyr::if_else(
molecular_profile == "KIT mutation in exon 9,11,13,14 or 17",
"KIT:449-514,550-592,627-664,664-714,788-828",
as.character(alteration)
)) |>
dplyr::mutate(alteration = dplyr::if_else(
stringr::str_detect(molecular_profile, " exon |-ITD$"),
stringr::str_replace_all(
alteration,
"KIT:|([A-Z0-9]{3,}::consequence::(skipping_mutation|(inframe_(insertion|variant|deletion))):)",
""
),
as.character(alteration)
)) |>
dplyr::mutate(alteration = dplyr::if_else(
stringr::str_detect(molecular_profile, " exon |-ITD$"),
stringr::str_replace_all(
alteration, "::(inframe_insertion|missense_variant):", ""),
as.character(alteration)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(molecular_profile, " exon |-ITD$") &
!is.na(alteration),
as.character(alteration),
as.character(variant_alias)
)) |>
tidyr::separate_longer_delim(variant_alias, delim = ",") |>
dplyr::distinct() |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias, "-"),
paste0("aa_region:", variant_alias),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(
molecular_profile, "^(IL7R inframe insertion \\(237)"),
"aa_region:237-255",
as.character(variant_alias)
)) |>
dplyr::left_join(
dplyr::select(gene_alias$records,
alias, symbol, entrezgene),
by = c("gene" = "alias"), multiple = "all",
relationship = "many-to-many"
) |>
dplyr::mutate(alteration = stringr::str_replace_all(
alteration,"__\\.","::v")
) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant_alias, "X$"),
"stop_gained",
as.character(variant_consequence)
)) |>
dplyr::mutate(alteration_type = dplyr::if_else(
stringr::str_detect(
variant_consequence,
"stop_gained|splice_donor|splice_acceptor"),
"MUT_LOF",
as.character(alteration_type)
))
)
unique_variants <- as.data.frame(
cgi_biomarkers |>
dplyr::filter(!is.na(variant_alias) &
!is.na(gene) &
!is.na(alteration_type)) |>
dplyr::select(gene, symbol, entrezgene, variant_alias,
alteration_type, variant_consequence) |>
dplyr:: arrange(symbol, alteration_type, variant_alias) |>
dplyr::distinct() |>
dplyr::mutate(variant_id = dplyr::row_number()) |>
dplyr::mutate(variant_name_primary = dplyr::case_when(
!is.na(symbol) ~ paste(symbol, variant_alias),
is.na(symbol) &
!is.na(gene) ~ paste(gene, variant_alias),
TRUE ~ as.character(NA)
))
)
cgi_variants <- data.frame()
for (i in 1:nrow(unique_variants)) {
alt <- unique_variants[i,"variant_alias"]
terms <- expand_hgvs_terms(alt, aa_dict = aa_dict)
for (t in terms) {
df <- data.frame(
'variant_id' =
unique_variants[i,"variant_id"],
'variant_alias' = t,
'variant_name_primary' = unique_variants[i,"variant_name_primary"],
'gene' = unique_variants[i,"gene"],
'entrezgene' = unique_variants[i,"entrezgene"],
'symbol' = unique_variants[i,"symbol"],
'variant_consequence' = unique_variants[i,"variant_consequence"],
'alteration_type' = unique_variants[i,"alteration_type"],
stringsAsFactors = F)
cgi_variants <- cgi_variants |>
dplyr::bind_rows(df)
}
}
cgi_variants <- cgi_variants |>
dplyr::mutate(alias_type = "hgvsp") |>
dplyr::mutate(alias_type = dplyr::case_when(
stringr::str_detect(variant_alias, "aa") ~ "aa_region",
stringr::str_detect(variant_alias, "^(Mut|Wild|Exp|Overexp|Underexp|Fus|Ampli|Delet)") ~ "other_gene",
TRUE ~ as.character(alias_type)
)) |>
dplyr::mutate(variant_exon = NA) |>
dplyr::mutate(biomarker_source = "cgi") |>
dplyr::mutate(biomarker_source_datestamp = datestamp) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
!is.na(alteration_type) & !is.na(variant_alias) &
stringr::str_detect(
variant_alias,
"^(([A-Z]{1}[0-9]{1,})|([A-Z]{1}[a-z]{2}[0-9]{1,}))$"),
as.character("CODON"),
as.character(alteration_type)
)) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
variant_name_primary,
alteration_type,
alias_type,
variant_consequence,
variant_exon,
gene,
symbol,
entrezgene,
) |>
dplyr::distinct()
cgi_clinical <- cgi_biomarkers |>
## get variant id's per molecular profile
dplyr::inner_join(
dplyr::select(
cgi_variants, variant_id,
gene, alias_type, variant_alias),
by = c("gene", "variant_alias"),
relationship = "many-to-many"
) |>
dplyr::select(-c(entrezgene, variant_consequence, variant_alias,
alias_type, gene, symbol, alteration_type,
alteration)) |>
dplyr::distinct() |>
dplyr::mutate(therapeutic_context = dplyr::if_else(
stringr::str_detect(
therapeutic_context,
"(ib|ide|ole|ant|stat|ine|us|one|mab|in|el|[0-9])( | )\\("),
stringr::str_replace(
therapeutic_context,
"( | )\\(.*\\)( )?$",""),
as.character(therapeutic_context)
))
therapeutic_contexts <- as.data.frame(
cgi_clinical |>
dplyr::select(molecular_profile_id, therapeutic_context) |>
dplyr::filter(!is.na(therapeutic_context) &
nchar(therapeutic_context) > 0) |>
tidyr::separate_rows(therapeutic_context, sep = " \\+ ") |>
dplyr::mutate(therapeutic_context = tolower(therapeutic_context)) |>
dplyr::left_join(
dplyr::select(compound_synonyms,
alias_lc,
drug_name,
molecule_chembl_id),
by = c("therapeutic_context" = "alias_lc"),
multiple = "all", relationship = "many-to-many"
) |>
dplyr::filter(!is.na(drug_name)) |>
dplyr::select(molecular_profile_id, molecule_chembl_id) |>
dplyr::distinct() |>
dplyr::group_by(molecular_profile_id) |>
dplyr::summarise(molecule_chembl_id = paste(
unique(sort(molecule_chembl_id)), collapse="|"),
.groups = "drop"
) |>
dplyr::ungroup() |>
dplyr::mutate(molecule_chembl_id = stringr::str_replace(
molecule_chembl_id,"^NA\\||\\|NA$",""
))
)
cgi_clinical <- cgi_clinical |>
dplyr::left_join(
therapeutic_contexts, by = "molecular_profile_id",
multiple = "all", relationship = "many-to-many") |>
dplyr::mutate(evidence_url = "https://www.cancergenomeinterpreter.org/biomarkers") |>
dplyr::mutate(biomarker_source = "cgi") |>
dplyr::mutate(biomarker_source_datestamp = datestamp) |>
dplyr::mutate(evidence_description = NA,
evidence_id = paste0("CGI", molecular_profile_id),
molecular_profile_summary = NA,
rating = NA,
biomarker_entity = T) |>
dplyr::select(-c(association, evidence_level_raw,
primary_tumor_type, drug_full_name)) |>
dplyr::rename(cancer_type = disease_name,
citation_id = source)
cgi_clinical <- map_biomarker_phenotypes(
biomarkers_clinical = cgi_clinical,
cache_dir = cache_dir) |>
dplyr::distinct()
biomarker_items <- list()
biomarker_items[['clinical']] <- cgi_clinical
biomarker_items[['clinical']]$molecular_profile <- NULL
biomarker_items <- get_literature_references(biomarker_items)
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
biomarker_entity,
molecular_profile_id,
molecular_profile_name,
molecular_profile_summary,
molecular_profile_type,
variant_id,
source_id,
evidence_id,
variant_origin,
primary_site,
cancer_type,
disease_ontology_id,
do_name,
efo_id,
efo_name,
cui,
cui_name,
dplyr::everything()
) |>
dplyr::distinct()
biomarker_items[['variant']] <- as.data.frame(
cgi_variants |>
dplyr::inner_join(
dplyr::select(
biomarker_items$clinical,
variant_id
), by = "variant_id",
relationship = "many-to-many"
) |>
dplyr::distinct()
)
return(biomarker_items)
}
map_biomarker_phenotypes <- function(biomarkers_clinical = NULL,
cache_dir = NA) {
cancer_term_map <-
phenOncoX::get_terms(cache_dir = cache_dir)$records |>
dplyr::select(do_id, efo_id, efo_name, do_name,
cui, cui_name, primary_site) |>
dplyr::filter(!is.na(do_id) & !is.na(cui)) |>
dplyr::rename(disease_ontology_id = do_id) |>
dplyr::distinct()
cancer_aux_pheno_maps <-
phenOncoX::get_aux_maps(cache_dir = cache_dir)
umls_terms_all <-
cancer_aux_pheno_maps$records$umls$concept |>
dplyr::filter(main_term == T) |>
dplyr::select(cui, cui_name)
do_terms_all <-
cancer_aux_pheno_maps$records$do |>
dplyr::select(
cui, do_id, do_name
) |>
dplyr::distinct()
efo_terms_all <-
cancer_aux_pheno_maps$records$efo$efo2xref |>
dplyr::select(
cui, efo_id, efo_name, primary_site) |>
dplyr::distinct()
biomarkers_all_phenotypes <- biomarkers_clinical |>
dplyr::left_join(
cancer_term_map, by = "disease_ontology_id",
multiple = "all", relationship = "many-to-many")
## if conditions in biomarkers is not found in phenOncoX terms
## match against individual dictionaries
missing_do_phenotypes <- biomarkers_all_phenotypes |>
dplyr::filter(is.na(cui)) |>
dplyr::select(-c(cui, cui_name, primary_site,
efo_id, do_name, efo_name)) |>
dplyr::inner_join(
do_terms_all,
by = c("disease_ontology_id" = "do_id"),
multiple = "all", relationship = "many-to-many") |>
dplyr::left_join(
umls_terms_all, by = "cui",
multiple = "all", relationship = "many-to-many") |>
dplyr::left_join(
efo_terms_all,
by = "cui", multiple = "all",
relationship = "many-to-many") |>
dplyr::distinct() |>
dplyr::mutate(primary_site = dplyr::case_when(
do_name == "cancer" ~ as.character(NA),
is.na(primary_site) & stringr::str_detect(do_name,"brain") ~ "CNS/Brain",
is.na(primary_site) & stringr::str_detect(do_name,"breast") ~ "Breast",
is.na(primary_site) & stringr::str_detect(do_name,"colon|rectum") ~ "Colon/Rectum",
is.na(primary_site) & stringr::str_detect(do_name,"gastric|stomach|esophag") ~ "Esophagus/Stomach",
is.na(primary_site) & stringr::str_detect(do_name,"prostate") ~ "Prostate",
is.na(primary_site) & stringr::str_detect(do_name,"pancrea") ~ "Pancreas",
is.na(primary_site) & stringr::str_detect(do_name,"lung") ~ "Lung",
is.na(primary_site) & stringr::str_detect(do_name,"myeloid") ~ "Myeloid",
is.na(primary_site) & stringr::str_detect(do_name,"lymphoma") ~ "Lymphoid",
is.na(primary_site) & stringr::str_detect(do_name,"ovary|ovarian") ~ "Ovary",
TRUE ~ as.character(primary_site)
))
biomarkers_phenotype_mapped <-
dplyr::filter(biomarkers_all_phenotypes, !is.na(cui)) |>
dplyr::bind_rows(missing_do_phenotypes)
return(biomarkers_phenotype_mapped)
}
load_custom_fusion_db <- function() {
other_translocations <- as.data.frame(read.table(
file = "data-raw/biomarkers/translocation_patterns.tsv", header = T,
sep = "\t", stringsAsFactors = F) |>
dplyr::select(ngram_pattern, biomarker) |>
dplyr::rename(alias = ngram_pattern,
variant = biomarker) |>
dplyr::mutate(biomarker_entity = T,
symbol = variant,
alias = toupper(alias)) |>
tidyr::separate_rows(alias, sep = "\\|") |>
dplyr::mutate(biomarker_source_db = "fusion_customdb") |>
dplyr::distinct()
)
other_translocations_lc <- as.data.frame(read.table(
file = "data-raw/biomarkers/translocation_patterns.tsv", header = T,
sep = "\t", stringsAsFactors = F) |>
dplyr::select(ngram_pattern, biomarker) |>
dplyr::rename(alias = ngram_pattern,
variant = biomarker) |>
dplyr::mutate(biomarker_entity = T,
symbol = variant) |>
tidyr::separate_rows(alias, sep = "\\|") |>
dplyr::mutate(biomarker_source_db = "fusion_customdb") |>
dplyr::distinct()
)
biomarker_items <- list()
biomarker_items[['variant']] <- other_translocations |>
dplyr::bind_rows(other_translocations_lc) |>
dplyr::distinct() |>
dplyr::rename(variant_alias = alias,
gene = symbol) |>
dplyr::mutate(alteration_type = "FUSION",
variant_consequence = "transcript_fusion",
variant_id = variant,
variant_name_primary = variant) |>
dplyr::mutate(alias_type = "other_gene") |>
dplyr::select(variant_id, variant_name_primary,
variant_alias, alias_type,
gene, alteration_type, variant_consequence)
return(biomarker_items)
}
load_depmap_fusions <- function(db_datestamp = "24Q4"){
# Load DepMap fusions
depmap_data <- list()
depmap_data[['fusions']] <- as.data.frame(read.csv(
file = "data-raw/depmap/OmicsFusionFiltered.csv", header = T))
depmap_data[['models']] <- as.data.frame(read.csv(
file = "data-raw/depmap/Model.csv", header = T)) |>
dplyr::select(
ModelID, CellLineName, OncotreeLineage,
OncotreePrimaryDisease, OncotreeCode,
Age, Sex, PrimaryOrMetastasis, SampleCollectionSite,
SourceType
)
return(depmap_data)
}
load_mitelman_db <- function(cache_dir = NA,
db_datestamp = "20241015") {
# Load Mitelman database
# dos2unix -q -n MBCA.TXT.DATA MBCA.TXT
morphology <- read.table(
file = file.path(
cache_dir, "mitelmandb", "KODER.TXT.DATA"),
sep = "\t", stringsAsFactors = F,
header = T, fill = T) |>
dplyr::filter(KodTyp == "MORPH" & nchar(Kod) > 0) |>
dplyr::mutate('morphology_code' = as.character(Kod),
'morphology_name' = Benamning) |>
dplyr::select(morphology_code, morphology_name) |>
dplyr::distinct()
topography <- read.table(
file = file.path(
cache_dir, "mitelmandb", "KODER.TXT.DATA"),
sep = "\t", stringsAsFactors = F,
header = T, fill = T) |>
dplyr::filter(KodTyp == "TOP" & nchar(Kod) > 0) |>
dplyr::mutate('topography_code' = as.character(Kod),
'topography_name' = Benamning) |>
dplyr::select(topography_code, topography_name) |>
dplyr::distinct()
pmid_data <- readr::read_tsv(
file = file.path(
cache_dir, "mitelmandb", "REF.TXT.DATA"),
quote = "none", show_col_types = F, guess_max = 1000000) |>
dplyr::select(RefNo, Pubmed) |>
dplyr::filter(nchar(Pubmed) > 0) |>
dplyr::mutate(reference_id = as.character(RefNo)) |>
dplyr::distinct() |>
dplyr::rename(source_id = Pubmed) |>
dplyr::select(reference_id, source_id) |>
dplyr::distinct()
fusion_event_data <- as.data.frame(readr::read_tsv(
file = file.path(
cache_dir, "mitelmandb", "MBCA.TXT.DATA"),
show_col_types = F, guess_max = 100000)) |>
dplyr::filter(stringr::str_detect(GeneShort,"::")) |>
dplyr::rename(variant = GeneShort,
karyotype = KaryShort,
reference_id = RefNo,
morphology_code = Morph,
topography_code = Topo) |>
dplyr::mutate(reference_id = as.character(reference_id),
morphology_code = as.character(morphology_code)) |>
dplyr::select(MolClin,
reference_id,
morphology_code,
topography_code,
variant, karyotype) |>
tidyr::separate_rows(variant, sep = ",") |>
dplyr::left_join(pmid_data, by = "reference_id",
multiple = "all",
relationship = "many-to-many") |>
dplyr::left_join(morphology, by = "morphology_code",
multiple = "all",
relationship = "many-to-many") |>
dplyr::left_join(topography, by = "topography_code",
multiple = "all",
relationship = "many-to-many") |>
dplyr::filter(stringr::str_detect(variant,"::")) |>
dplyr::mutate(variant = stringr::str_replace(
variant, "\\+", "")) |>
dplyr::mutate(evidence_id = dplyr::row_number()) |>
dplyr::mutate(evidence_id = paste0("MITDB_",evidence_id)) |>
dplyr::filter(stringr::str_count(variant,",") ==
stringr::str_count(karyotype,",")) |>
tidyr::separate_rows(c(variant,karyotype),
sep = ",") |>
dplyr::mutate(topography_name2 = dplyr::case_when(
is.na(topography_name) &
stringr::str_detect(
tolower(morphology_name), "leukemia|myeloma") ~ "Myeloid",
is.na(topography_name) &
stringr::str_detect(
tolower(morphology_name), "lymphoma") ~ "Lymphoid",
topography_name == "Ovary" ~ "Ovary/Fallopian Tube",
topography_name == "Fallopian tube" ~ "Ovary/Fallopian Tube",
topography_name == "Uterus, corpus" ~ "Uterus",
stringr::str_detect(morphology_name,
paste0(
"Nasal cavity/Paranasal sinuses|Nasopharynx|",
"LarynxOral cavity|Oro- and hypopharynx|",
"Larynx|Teeth")) ~ "Head and Neck",
topography_name == "Gallbladder/Biliary system" ~ "Biliary Tract",
topography_name == "Soft tissue" ~ "Soft Tissue",
topography_name == "Brain" ~ "CNS/Brain",
morphology_name == "Mature B-cell neoplasm, NOS" ~ "Lymphoid",
topography_name == "Brain stem" ~ "CNS/Brain",
topography_name == "Cerebellum" ~ "CNS/Brain",
topography_name == "Vagina" ~ "Vulva/Vagina",
topography_name == "Skeleton" ~ "Bone",
stringr::str_detect(
tolower(morphology_name),
"myelodysplastic|myeloproliferative") ~ "Myeloid",
stringr::str_detect(
tolower(morphology_name),
"miscellaneous hematopoietic/lymphoid|hodgkin disease") ~ "Lymphoid",
topography_name == "Ureter" ~ "Kidney",
topography_name == "Unknown site" ~ "Any",
topography_name == "Urethra" ~ "Bladder/Urinary Tract",
topography_name == "Spinal cord" ~ "Peripheral Nervous System",
topography_name == "Uterus, cervix" ~ "Cervix",
topography_name == "Large intestine" ~ "Colon/Rectum",
topography_name == "Bladder" ~ "Bladder/Urinary Tract",
topography_name == "Adrenal" ~ "Adrenal Gland",
topography_name == "Intraabdominal" ~ "Esophagus/Stomach",
topography_name == "Oesophagus" ~ "Esophagus/Stomach",
TRUE ~ topography_name
)) |>
dplyr::rename(primary_diagnosis = morphology_name,
primary_site = topography_name2) |>
dplyr::select(variant,
karyotype,
primary_diagnosis,
primary_site,
source_id,
evidence_id) |>
dplyr::distinct() |>
dplyr::mutate(variant_id = paste(
variant, karyotype, sep = " - "
))
evidence_id <- nrow(fusion_event_data) + 1
fusion_event_data$variant_alias <- paste(
fusion_event_data$variant,
paste(
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,2],
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,1],
sep = "-"
),
paste(
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,1],
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,2],
sep = "/"
),
paste(
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,1],
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,2],
sep = "-"
),
fusion_event_data$karyotype,
sep = "@@@"
)
fusion_event_data$variant_name_primary <- fusion_event_data$variant
bcr_abl_custom <-
data.frame('variant_id' = "BCR::ABL1 - t(9;22)(q34;q11)",
'variant_alias' = 'BCR::ABL',
'karyotype' = 't(9;22)(q34;q11)',
'variant_name_primary' = 'BCR::ABL1',
'evidence_id' = paste0('MITDB_',evidence_id),
'primary_diagnosis' = NA,
'primary_site' = NA,
'source_id' = NA,
'symbol' = 'BCR-ABL1',
stringsAsFactors = F)
mitelman_db <- as.data.frame(
fusion_event_data |>
dplyr::select(variant_alias,
variant_id,
karyotype,
variant_name_primary,
primary_diagnosis,
primary_site,
evidence_id,
source_id) |>
tidyr::separate_rows(variant_alias, sep = "@@@") |>
#dplyr::rename(variant_alias = alias) |>
dplyr::filter(
!stringr::str_detect(
variant_alias,"RARA::PML|ABL1::BCR|ERG::TMPRSS2")) |>
dplyr::bind_rows(
bcr_abl_custom
) |>
dplyr::mutate(alteration_type = "FUSION",
variant_origin = "somatic",
variant_consequence = "transcript_fusion",
evidence_url = "https://mitelmandatabase.isb-cgc.org/",
biomarker_source = "mitelmandb",
#biomarker_entity = T,
#alias_type = "other_gene",
biomarker_source_datestamp = db_datestamp) |>
dplyr::arrange(evidence_id) |>
dplyr::rename(citation_id = source_id) |>
dplyr::select(biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
#alias_type,
variant_name_primary,
alteration_type,
dplyr::everything())
)
mitelman_db$symbol <- stringr::str_split_fixed(
mitelman_db$variant_id, " - ", 2)[,1]
biomarker_items <- list()
biomarker_items[['clinical']] <- mitelman_db
biomarker_items[['variant']] <- biomarker_items[['clinical']] |>
#dplyr::select(variant_id, variant_alias, alteration_type, symbol) |>
dplyr::mutate(alias_type = "other_gene") |>
dplyr::rename(gene = symbol) |>
#dplyr::rename(variant_alias = variant) |>
dplyr::select(biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
variant_name_primary,
alteration_type,
alias_type,
variant_consequence,
gene) |>
dplyr::distinct()
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::select(-c("variant_alias", "variant_name_primary",
"alteration_type","variant_consequence",
"symbol")) |>
dplyr::distinct()
biomarker_items <-
get_literature_references(biomarker_items)
return(biomarker_items)
}
sigven/oncoPharmaDB documentation built on Dec. 24, 2024, 9:46 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
clean_variant_consequence <- function(variant_df){
variant_df <- variant_df |>
dplyr::mutate(
variant_consequence =
dplyr::if_else(
stringr::str_detect(variant,"Overexpression") &
variant_consequence == "N/A",
"transcript_amplification",
as.character(variant_consequence))) |>
dplyr::mutate(
variant_consequence =
dplyr::if_else(
stringr::str_detect(variant,"Underexpression") &
variant_consequence == "N/A",
"transcript_ablation",
as.character(variant_consequence))) |>
dplyr::mutate(
variant_consequence =
dplyr::if_else(
stringr::str_detect(
variant_consequence,
"^(plus_1_frameshift|minus_1_frameshift|frameshift_)"),
"frameshift_variant",
as.character(variant_consequence))) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
variant_consequence == "" &
stringr::str_detect(variant,"EXON") &
stringr::str_detect(variant,"DELETION"),
"exon_loss_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
variant_consequence == "" &
stringr::str_detect(variant,"Exon") &
stringr::str_detect(variant,"Mutation"),
"exon_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
variant_consequence == "" &
stringr::str_detect(variant,"^(Loss|Loss-of-function)$"),
"loss_of_function_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant,"^(Activating |Gain-of)") &
(is.na(variant_consequence) | nchar(variant_consequence) == 0),
"gain_of_function_variant",
as.character(variant_consequence)
)) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant,"^Wildtype") &
(is.na(variant_consequence) | nchar(variant_consequence) == 0),
"wild_type",
as.character(variant_consequence)
))
return(variant_df)
}
set_alteration_type <- function(variant_df){
variant_df <- variant_df |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(toupper(variant),"EXPRESSION"),
"EXP",as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::case_when(
alteration_type == "EXP" &
stringr::str_detect(
toupper(variant),"OVER") ~ "EXP_OVER",
alteration_type == "EXP" &
stringr::str_detect(
toupper(variant),"UNDER") ~ "EXP_UNDER",
TRUE ~ as.character(alteration_type)
)
) |>
dplyr::mutate(alteration_type = dplyr::if_else(
variant_consequence == "stop_gained",
"MUT_LOF",as.character(alteration_type)
)) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant),"LOSS-OF-FUNCTION|LOSS|TRUNCAT|INACTIVAT") |
stringr::str_detect(
variant_consequence,
"stop_gained,loss_of_function_variant|loss_of_function_variant,stop_gained") |
(stringr::str_detect(
variant_consequence, "loss_of_function_variant") &
toupper(variant) == "MUTATION") |
stringr::str_detect(variant_consequence,"truncation"),
"MUT_LOF",as.character(alteration_type))) |>
dplyr::mutate(alteration_type = dplyr::if_else(
toupper(variant) == "BIALLELIC INACTIVATION",
"MUT_LOF_BIALLELIC",
as.character(alteration_type)
)) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
alteration_type == 'MUT_LOF' &
(stringr::str_detect(
toupper(variant_consequence),"FRAMESHIFT") |
stringr::str_detect(
variant_consequence,"feature_truncation")),
"MUT_LOF_FS",as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant),
"AMPLIFICATION|DELETION|COPY") &
stringr::str_detect(
variant_consequence, "_amplification|_ablation"),
"CNA",as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant_consequence),
"TRANSCRIPT_FUSION|REGION_FUSION") |
stringr::str_detect(toupper(variant),"FUSION") |
(variant_consequence == "" &
stringr::str_detect(variant,"::")),
"FUSION",
as.character(alteration_type))) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
toupper(variant),"REARRANGEMENT") &
variant_consequence == "",
"REARRANGEMENT",
as.character(alteration_type)))
return(variant_df)
}
clean_variant_naming <- function(
variant_df, col = "molecular_profile_name"){
variant_df[,"cc_clean"] <- variant_df[,col]
variant_df <- variant_df |>
dplyr::mutate(cc_clean = stringr::str_trim(cc_clean)) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean,"RS[0-9]{5,}"),
stringr::str_replace_all(cc_clean,"RS","rs"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = stringr::str_replace_all(
cc_clean,
paste0(
"(ENST[0-9]{1,}|",
"NM_[0-9]{1,})(\\.[0-9]{1,})?:c\\."),"c."
)) |>
dplyr::mutate(cc_clean = stringr::str_replace_all(
cc_clean,
paste0(
"(ENSP[0-9]{1,}|NP_[0-9]{1,}",
")(\\.[0-9]{1,})?:p\\."),"p."
)) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "^NC_[0-9]{1,}"),
"", as.character(cc_clean)
)) |>
dplyr::mutate(cc_clean = dplyr::case_when(
cc_clean == "D835H/Y" ~ "D835H,D835Y",
cc_clean == "G12/G13" ~ "G12,G13",
cc_clean == "Q157P/R" ~ "Q157P,Q157R",
cc_clean == "S310F/Y" ~ "S310F,S310Y",
cc_clean == "S34Y/F" ~ "S34Y,S34F",
cc_clean == "S893A/T" ~ "S893A,S893T",
TRUE ~ as.character(cc_clean)
)) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "^(CD44s|p16|PRPS1|IKZF1|NUCLEAR) "),
stringr::str_replace_all(
cc_clean, "^(CD44s|p16|PRPS1|IKZF1|NUCLEAR) ", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, " NUCLEAR "),
stringr::str_replace_all(
cc_clean, " NUCLEAR ", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "IVS2\\+1G>A"),
stringr::str_replace_all(
cc_clean, "^IVS2\\+1G>A", "c.444+1G>A"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, " HOMOZYGOSITY"),
stringr::str_replace_all(
cc_clean, " HOMOZYGOSITY", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "deletion and mutation"),
stringr::str_replace_all(
cc_clean, "deletion and mutation", "Mutation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = stringr::str_replace(
cc_clean, " \\(.+\\)$","")) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "\\*$"),
stringr::str_replace_all(cc_clean, "\\*$", "X"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "^\\*"),
stringr::str_replace_all(cc_clean, "^\\*", "X"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "\\* "),
stringr::str_replace_all(cc_clean, "\\* ", "X "),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Rare "),
stringr::str_replace_all(cc_clean, "Rare ", ""),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, " expression"),
stringr::str_replace_all(
cc_clean, " expression", " Expression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "BIALLELIC"),
stringr::str_replace_all(
cc_clean, "BIALLELIC", "Biallelic"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "MUTATION|mutation"),
stringr::str_replace_all(
cc_clean, "MUTATION|mutation", "Mutation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "DOMAIN"),
stringr::str_replace_all(
cc_clean, "DOMAIN", "Domain"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "CONSERVED"),
stringr::str_replace_all(
cc_clean, "CONSERVED", "Conserved"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "DELETION"),
stringr::str_replace_all(
cc_clean, "DELETION", "Deletion"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "FRAMESHIFT|FRAME SHIFT"),
stringr::str_replace_all(
cc_clean, "FRAMESHIFT|FRAME SHIFT", "Frameshift"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "INACTIVATING"),
stringr::str_replace_all(
cc_clean, "INACTIVATING", "Inactivating"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "INACTIVATION"),
stringr::str_replace_all(
cc_clean, "INACTIVATION", "Inactivation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "OVEREXPRESSION"),
stringr::str_replace_all(
cc_clean, "OVEREXPRESSION", "Overexpression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "UNDEREXPRESSION"),
stringr::str_replace_all(
cc_clean, "UNDEREXPRESSION", "Underexpression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "EXPRESSION"),
stringr::str_replace_all(
cc_clean, "EXPRESSION", "Expression"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "EXON"),
stringr::str_replace_all(cc_clean, "EXON", "Exon"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "LOSS$"),
stringr::str_replace_all(cc_clean, "LOSS", "Loss"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "AMPLIFICATION"),
stringr::str_replace_all(
cc_clean, "AMPLIFICATION", "Amplification"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "TRUNCATING"),
stringr::str_replace_all(
cc_clean, "TRUNCATING", "Truncating"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "TRUNCATION"),
stringr::str_replace_all(
cc_clean, "TRUNCATION", "Truncation"),
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(
cc_clean, "TRANSLOCATION|rearrangement"),
stringr::str_replace_all(
cc_clean, "TRANSLOCATION|rearrangement", "Rearrangement"),
as.character(cc_clean)))
if(col != "molecular_profile_name"){
variant_df <- variant_df |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Binding Domain Mutation") &
feature_name == "TP53",
"aa_region:100-288",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "SH2 Domain") &
feature_name == "STAT3",
"aa_region:584-651",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Kinase Domain") &
feature_name == "ERBB2",
"aa_region:722-975",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "TKD ") &
feature_name == "FLT3",
"aa_region:610-942",
as.character(cc_clean))) |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "Conserved Domain") &
feature_name == "TP53",
paste0(
"aa_region:117-143,aa_region:171-181,",
"aa_region:234-258,aa_region:270-286"),
as.character(cc_clean)))
}
variant_df <- variant_df |>
dplyr::mutate(cc_clean = dplyr::if_else(
stringr::str_detect(cc_clean, "N/A"),
stringr::str_replace_all(cc_clean, "N/A", ""),
as.character(cc_clean))) |>
dplyr::distinct()
variant_df[,col] <- variant_df[,"cc_clean"]
variant_df[,"cc_clean"] <- NULL
return(variant_df)
}
get_literature_references <- function(biomarker_items){
## set logging layout
lgr::lgr$appenders$console$set_layout(
lgr::LayoutFormat$new(timestamp_fmt = "%Y-%m-%d %T"))
biomarker_items[['literature']] <- as.data.frame(
biomarker_items[['clinical']] |>
dplyr::filter(!is.na(citation_id)) |>
tidyr::separate_rows(citation_id, sep=";") |>
dplyr::mutate(
citation_id = stringr::str_trim(citation_id)) |>
dplyr::filter(
!stringr::str_detect(
tolower(citation_id),
"asco|aacr|abstr|suppl|confex|jci|blood|annonc|caris")) |>
dplyr::distinct() |>
dplyr::mutate(citation_id = dplyr::case_when(
citation_id == "PMC3638050" ~ "PMID:22038996",
citation_id == "PMC3936420" ~ "PMID:24265155",
TRUE ~ as.character(
stringr::str_replace(citation_id,"PMID: ","PMID:"))
)) |>
dplyr::filter(!stringr::str_detect(citation_id, "^[0-9]{1,5}$")) |>
dplyr::filter(citation_id != "PMID:238900088" &
citation_id != "PMID:24569783" &
citation_id != "2556735") |>
dplyr::mutate(citation_id = dplyr::case_when(
stringr::str_detect(citation_id, "^FDA") ~ "FDA",
stringr::str_detect(citation_id, "NCCN") ~ "NCCN",
TRUE ~ as.character(citation_id)
)) |>
dplyr::mutate(citation_id = stringr::str_replace(
citation_id, "PMID:",""
)) |>
dplyr::select(citation_id, evidence_id) |>
dplyr::rename(source_id = citation_id) |>
dplyr::mutate(source_id = as.character(source_id)) |>
dplyr::group_by(source_id) |>
dplyr::summarise(evidence_id = paste(
unique(sort(evidence_id)), collapse = ";"
), .groups = "drop") |>
dplyr::left_join(
literatureVault::vault,
by = "source_id",
relationship = "many-to-many",
multiple = "all"
)
)
biomarker_items[['clinical']]$citation_id <- NULL
not_found <- biomarker_items$literature |>
dplyr::filter(is.na(link)) |>
dplyr::select(source, source_id)
if(NROW(not_found) > 0){
lgr::lgr$info(
"Citations missing in sigven/literatureVault::vault - update needed")
return(not_found)
}else{
lgr::lgr$info(
"Found all literature references in sigven/literatureVault::vault"
)
}
## ignore molecular profiles/biomarkers that lack literature support
biomarkers_with_literature_support <- as.data.frame(
biomarker_items[['literature']] |>
dplyr::select(evidence_id, source_id) |>
tidyr::separate_rows(evidence_id, sep = ";") |>
dplyr::group_by(evidence_id) |>
dplyr::mutate(source_id = paste(
unique(source_id), collapse = ";"
)) |>
dplyr::distinct()
)
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::inner_join(
biomarkers_with_literature_support,
by = "evidence_id", multiple = "all",
relationship = "many-to-many")
return(biomarker_items)
}
#' #' A function that splits an array into chunks of equal size
#' chunk <- function(x,n) split(x, factor(sort(rank(x)%%n)))
#'
#' #' A function that returns a citation with first author, journal and year for a PubMed ID
#' #'
#' #' @param pmid An array of Pubmed IDs
#' #' @param chunk_size Size of PMID chunks
#' #'
#' #' @export
#' #'
#' #' @return citation PubMed citation, with first author, journal and year
#' #'
#' get_citations_pubmed <- function(
#' pmid,
#' pmid_cache_dir = NA,
#' chunk_size = 2){
#'
#' ## make chunk of maximal 400 PMIDs from input array (limit by EUtils)
#' pmid_chunks <- chunk(pmid, ceiling(length(pmid)/chunk_size))
#' j <- 0
#' all_citations <- data.frame()
#' cat('Retrieving PubMed citations for PMID list, total length', length(pmid))
#' cat('\n')
#' while(j < length(pmid_chunks)){
#' #cat(unlist(pmid_chunks),"\n")
#' pmid_chunk <- pmid_chunks[[as.character(j)]]
#' #cat(unlist(pmid_chunk),"\n")
#' #cat('Processing chunk ',j,' with ',length(pmid_chunk),'PMIDS')
#' #cat('\n')
#' pmid_string <- paste(pmid_chunk,collapse = " ")
#' res <- RISmed::EUtilsGet(
#' RISmed::EUtilsSummary(pmid_string, type="esearch", db="pubmed", retmax = 5000)
#' )
#'
#'
#' year <- RISmed::YearPubmed(res)
#' authorlist <- RISmed::Author(res)
#' pmid_list <- RISmed::PMID(res)
#' i <- 1
#' first_author <- c()
#' while(i <= length(authorlist)){
#'
#'
#' if(length(authorlist[[i]]) == 5){
#' first_author <- c(
#' first_author,
#' paste(authorlist[[i]][1,]$LastName," et al.",sep=""))
#' }else{
#' first_author <- c(
#' first_author, as.character("Unknown et al.")
#' )
#' }
#' i <- i + 1
#' }
#' journal <- RISmed::ISOAbbreviation(res)
#' citations <- data.frame('pmid' = as.integer(pmid_list),
#' 'citation' = paste(first_author,year,journal,sep=", "),
#' stringsAsFactors = F)
#' citations$link <- paste0('<a href=\'https://www.ncbi.nlm.nih.gov/pubmed/',
#' citations$pmid,'\' target=\'_blank\'>',
#' citations$citation,'</a>')
#' all_citations <- dplyr::bind_rows(
#' all_citations, citations)
#' j <- j + 1
#' }
#'
#' return(all_citations)
#'
#' }
#'
#'
#' get_literature_links <- function(literature_df) {
#'
#' stopifnot(is.data.frame(literature_df))
#' assertable::assert_colnames(
#' literature_df, c("source","source_id","evidence_id"),
#' quiet = T)
#'
#' pubmed_source_ids <- literature_df |>
#' dplyr::filter(source == "PubMed")
#'
#' other_source_ids <- literature_df |>
#' dplyr::filter(source != "PubMed")
#'
#'
#' if (NROW(pubmed_source_ids) > 0) {
#' pubmed_source_ids <- get_citations_pubmed(
#' pmid = pubmed_source_ids$source_id, chunk_size = 100) |>
#' dplyr::rename(
#' name = citation,
#' ) |>
#' dplyr::mutate(source_id = as.character(pmid)) |>
#' dplyr::select(-pmid) |>
#' dplyr::left_join(
#' pubmed_source_ids, by = "source_id")
#' }
#'
#' if (NROW(other_source_ids) > 0) {
#' other_source_ids <- other_source_ids |>
#' dplyr::filter(source != "PubMed") |>
#' dplyr::mutate(name = source_id) |>
#' dplyr::mutate(link = dplyr::case_when(
#' source == "FDA" ~
#' "<a href='https://www.fda.gov/drugs/resources-information-approved-drugs/oncology-cancer-hematologic-malignancies-approval-notifications' target='_blank'>FDA approvals</a>",
#' source == "NCCN" ~
#' "<a href='https://www.nccn.org/guidelines/category_1' target='_blank'>NCCN guidelines</a>",
#' source == "EMA" ~ "<a href='https://www.ema.europa.eu/en/medicines/field_ema_web_categories%253Aname_field/Human' target='_blank'>European Medicines Agency (EMA)</a>",
#' source == "clinicaltrials.gov" ~ paste0(
#' "<a href='https://clinicaltrials.gov/ct2/show/",
#' source_id, "' target='_blank'>",
#' source_id, "</a>")
#' ))
#' }
#'
#' all_literature_links <- pubmed_source_ids |>
#' dplyr::select(source, source_id,
#' evidence_id,
#' dplyr::everything())
#'
#' if(NROW(other_source_ids) > 0){
#' all_literature_links <-
#' other_source_ids |>
#' dplyr::bind_rows(pubmed_source_ids)
#' }
#'
#' return(all_literature_links)
#'
#'
#' }
get_amino_acid_dictionary <- function() {
aa_df <-
data.frame('one_letter' = names(Biostrings::AMINO_ACID_CODE),
'three_letter' = Biostrings::AMINO_ACID_CODE,
'three_letter_uc' = toupper(Biostrings::AMINO_ACID_CODE),
stringsAsFactors = F) |>
dplyr::filter(one_letter != "B" &
one_letter != "J" &
one_letter != "O" &
one_letter != "U" &
one_letter != "X" &
one_letter != "Z")
rownames(aa_df) <- NULL
return(aa_df)
}
expand_hgvs_terms <- function(var, aa_dict, add_codon_markers = FALSE) {
hits <- c()
if (stringr::str_detect(toupper(var),"([A-Z]{3}[0-9]{1,}(([A-Z]{3}) {0,1})([0-9]{1,})?)") &
!stringr::str_detect(toupper(var),"(DEL|INS|DUP)|=|-|\\*")) {
aa <- stringr::str_to_title(strsplit(var, "[0-9]{1,}")[[1]])
codon <- stringr::str_match(var, "[0-9]{1,}")[[1]]
if (length(aa) == 1) {
alt1 <- paste0(aa[1],codon)
alt2 <- paste0('p.', alt1)
if (aa[1] != "Ter") {
if (aa[1] %in% aa_dict$three_letter) {
aa1 <- seqinr::a(aa[1])
if (!is.na(aa1)) {
alt3 <- paste0(seqinr::a(aa[1]), codon)
hits <- c(alt1, alt2, alt3)
}
}
}else{
alt3 <- paste0("X", codon)
hits <- c(alt1, alt2, alt3)
}
}else{
if (length(aa) == 2) {
alt1 <- stringr::str_replace(
stringr::str_replace(paste0(aa[1],codon, aa[2]),"(FS|Fs)$","fs"),
"ter","Ter")
alt2 <- stringr::str_replace(
stringr::str_replace(paste0('p.', alt1),"(FS|Fs)$","fs"),
"ter","Ter")
if (aa[1] != "Ter" &
aa[2] != "Ter" &
aa[2] != "Fs" &
nchar(aa[1]) == 3 &
nchar(aa[2]) == 3 &
aa[1] %in% aa_dict$three_letter &
aa[2] %in% aa_dict$three_letter) {
alt3 <- paste0(seqinr::a(aa[1]), codon, seqinr::a(aa[2]))
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
else{
if (aa[1] == "Ter") {
alt3 <- paste0("X", codon, seqinr::a(aa[2]))
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}else{
if (aa[2] == "Ter") {
if (aa[1] %in% aa_dict$three_letter) {
alt3 <- paste0(seqinr::a(aa[1]), codon, "X")
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
}
if (aa[2] == "Fs") {
if (aa[1] %in% aa_dict$three_letter) {
alt3 <- paste0(seqinr::a(aa[1]), codon, "fs")
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
}
}
}
}
}
}
if (stringr::str_detect(var,"^([A-Z]{1}[0-9]{1,}[A-Z]{0,1})$") &
!stringr::str_detect(var, "FS|DUP|DEL|INS|fs|dup|del|ins|\\*")) {
aa <- stringr::str_to_title(strsplit(var, "[0-9]{1,}")[[1]])
codon <- stringr::str_match(var, "[0-9]{1,}")[[1]]
if (length(aa) == 1) {
alt1 <- paste0(aa[1],codon)
alt2 <- paste0('p.', alt1)
if (aa[1] != "X" & aa[1] %in% aa_dict$one_letter) {
alt3 <- paste0(seqinr::aaa(aa[1]), codon)
hits <- c(alt1, alt2, alt3)
}else{
alt3 <- paste0("Ter", codon)
hits <- c(alt1, alt2, alt3)
}
}else{
if (length(aa) == 2) {
alt1 <- paste0(aa[1],codon, aa[2])
alt2 <- paste0('p.', alt1)
if (aa[1] != "X" &
aa[2] != "X" &
aa[1] %in% aa_dict$one_letter &
aa[2] %in% aa_dict$one_letter) {
alt3 <- paste0(seqinr::aaa(aa[1]),
codon, seqinr::aaa(aa[2]))
alt4 <- paste0('p.', alt3)
if(aa[2] == aa[1]){
alt5 <- paste0("p.", seqinr::aaa(aa[1]),
codon, "=")
alt6 <- paste0("p.", aa[1],
codon, "=")
hits <- c(alt1, alt2, alt3, alt4, alt5, alt6)
}else{
hits <- c(alt1, alt2, alt3, alt4)
}
}else{
if (aa[1] == "X" & aa[2] %in% aa_dict$one_letter) {
alt3 <- paste0("Ter",
codon, seqinr::aaa(aa[2]))
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}else{
if (aa[2] == "X" & aa[1] %in% aa_dict$one_letter) {
alt3 <- paste0(seqinr::aaa(aa[1]),
codon, "Ter")
alt4 <- paste0('p.', alt3)
hits <- c(alt1, alt2, alt3, alt4)
}
}
}
}
}
}
if (stringr::str_detect(var,"^C\\.")) {
var <- stringr::str_replace(var,"^C\\.","c.")
#hits <- res
}
if (stringr::str_detect(var,"^P\\.")) {
var <- stringr::str_replace(var,"^P\\.","p.")
#hits <- res
}
if (stringr::str_detect(var,"FS$")) {
var <- stringr::str_replace(var,"FS$","fs")
}
if (stringr::str_detect(var,";") & stringr::str_detect(var,"\\(")) {
var <- tolower(var)
}
if (stringr::str_detect(var,"^([A-Z]{1}[0-9]{1,}[A-Z]{1})$")) {
res1 <- paste0('p.', var)
if(add_codon_markers == TRUE){
res2 <- stringr::str_match(var,"^([A-Z]{1}[0-9]{1,})")[[1]]
hits <- c(hits,res1,res2)
}else{
hits <- c(hits, res1)
}
}
if (stringr::str_detect(var,"FS$")) {
res <- stringr::str_replace(var,"FS$","fs")
hits <- res
}
if (stringr::str_detect(var,"(DEL|Del|INS|Ins|Dup|DUP)") &
!stringr::str_detect(var, "INSERTION|DUPLICATION|NM_|NP_|DELETERIOUS|DELETION|MUTATION|Mutation|Deletion|")) {
var <- stringr::str_replace(var,"(DEL|Del)","del")
var <- stringr::str_replace(var,"(INS|Ins)","ins")
var <- stringr::str_replace(var,"(DUP|Dup)","dup")
if (!(stringr::str_detect(var,"^(c|p|rs|;|[0-9])"))) {
res2 <- paste0('p.', var)
hits <- c(var, res2)
}
}
if (length(hits) == 0 & var != "") {
if (!(stringr::str_detect(var,"^(c|rs|p|[0-9])")) &
stringr::str_detect(var, "[0-9]{1,}") &
!stringr::str_detect(
var,"EXON|;|MUTATION|Mutation|-|NM_|NP_|DELETERIOUS|Deletion|INTRON|DELETION|DUPLICATION")) {
if (stringr::str_detect(var,"FS$|DUP|DELINS|FS\\*")) {
var <- stringr::str_replace(var, "FS","fs")
var <- stringr::str_replace(var, "DUP","dup")
var <- stringr::str_replace(var, "DELINS","delins")
var <- stringr::str_replace(var, "DEL","del")
}
res1 <- paste0('p.',var)
hits <- c(var, res1)
}else{
hits <- c(var)
}
}
hits <- unique(hits)
return(hits)
}
load_civic_biomarkers <- function(
datestamp = '20240709',
compound_synonyms = NULL,
hg38_fasta =
"/Users/sigven/research/DB/hg38/hg38.fa",
crossmap_chainfile_dir =
"/Users/sigven/research/DB/chainFiles",
cache_dir = NA) {
gene_alias <- geneOncoX::get_alias(
cache_dir = cache_dir
)
civic_local_fnames <- list()
for(e in c("VariantSummaries",
"civic_accepted",
"ClinicalEvidenceSummaries",
"MolecularProfileSummaries")){
civic_local_fnames[[e]] <- file.path(
cache_dir,
paste0("civic_", e, "_", datestamp,".tsv")
)
remote_url <- paste0(
"https://civicdb.org/downloads/nightly/nightly-", e, ".tsv")
if(e == "civic_accepted"){
remote_url <- paste0(
"https://civicdb.org/downloads/nightly/nightly-", e, ".vcf")
civic_local_fnames[['vcf_grch37']] <- file.path(
cache_dir,
paste0("civic_", datestamp,"_grch37.vcf")
)
civic_local_fnames[['vcf_grch38']] <- file.path(
cache_dir,
paste0("civic_", datestamp,"_grch38.vcf")
)
}
if(e == "civic_accepted"){
if(!file.exists(civic_local_fnames[['vcf_grch37']])){
download.file(
url = remote_url,
destfile = civic_local_fnames[['vcf_grch37']])
}
}else{
if(!file.exists(civic_local_fnames[[e]])){
download.file(
url = remote_url,
destfile = civic_local_fnames[[e]])
}
}
if(e == 'civic_accepted'){
vcfhelpR::crossmap_vcf(
direction = "hg19Tohg38",
chain_file_directory = crossmap_chainfile_dir,
target_genome_file = hg38_fasta,
source_vcf = civic_local_fnames[['vcf_grch37']],
target_vcf = civic_local_fnames[['vcf_grch38']])
system(paste0('bgzip -dc ',civic_local_fnames[['vcf_grch38']],
'.gz > ',
civic_local_fnames[['vcf_grch38']]))
}
}
aa_dict <- get_amino_acid_dictionary()
variant_coordinates <- data.frame()
for(assembly in c('grch37','grch38')){
skip_lines <- 8
if(assembly == "grch38"){
skip_lines <- 7
}
variant_coordinates <- as.data.frame(
dplyr::bind_rows(
variant_coordinates,
readr::read_tsv(
file = civic_local_fnames[[paste0('vcf_',assembly)]],
skip = skip_lines,
show_col_types = F,
col_names = T,
guess_max = 1000) |>
janitor::clean_names() |>
dplyr::rename(variant_id = id) |>
dplyr::mutate(
vcf_coord = paste0(
assembly, ":",
paste(number_chrom, pos, ref, alt, sep="_"))
) |>
dplyr::select(variant_id, vcf_coord)
)
)
}
## ignore these biomarkers
biomarker_names_skip_regex <-
paste0(
"methylat|phosphoryl|serum lev|mislocali|peri-therap|",
"alternative|cytoplasmic|p-loop|reg_|alternative|polymorphism|",
"loss of heterozygosity|philadelphia|internal|n-terminal")
molecularProfileSummary <- as.data.frame(
data.table::fread(
civic_local_fnames[['MolecularProfileSummaries']],
select = c(1:13), fill = T)) |>
dplyr::rename(variant_id = variant_ids,
molecular_profile_name = name,
molecular_profile_summary = summary) |>
dplyr::select(variant_id,
molecular_profile_id,
molecular_profile_name,
molecular_profile_summary) |>
## ignore profiles with aberration combinations for now
tidyr::separate_rows(variant_id, sep = ",") |>
##dplyr::filter(!stringr::str_detect(variant_id, ",")) |>
dplyr::mutate(variant_id = as.integer(variant_id)) |>
clean_variant_naming(col = "molecular_profile_name") |>
dplyr::filter(!stringr::str_detect(
tolower(molecular_profile_name),
biomarker_names_skip_regex))
clinicalEvidenceSummary <- as.data.frame(
readr::read_tsv(
civic_local_fnames[['ClinicalEvidenceSummaries']],
show_col_types = F, guess_max = 5000)) |>
dplyr::filter(source_type == "PubMed") |>
dplyr::mutate(citation_id = as.character(citation_id)) |>
dplyr::select(molecular_profile_id,
evidence_id,
evidence_civic_url,
disease,
doid,
therapies,
citation_id,
evidence_statement,
evidence_type,
evidence_level,
evidence_direction,
significance,
variant_origin,
rating) |>
dplyr::mutate(citation_id = dplyr::case_when(
citation_id == "27577079" ~ "28947956",
citation_id == "27556863" ~ "28978004",
TRUE ~ as.character(citation_id)
)) |>
dplyr::mutate(disease = dplyr::if_else(
is.na(disease),
"Cancer",
as.character(disease)
)) |>
dplyr::mutate(doid = dplyr::if_else(
is.na(doid),
"162",
as.character(doid)
)) |>
dplyr::rename(disease_ontology_id = doid,
clinical_significance = significance,
evidence_description = evidence_statement,
evidence_url = evidence_civic_url,
cancer_type = disease) |>
dplyr::mutate(
clinical_significance =
dplyr::case_when(
clinical_significance == "Resistance" |
clinical_significance == "Reduced Sensitivity" ~ "Resistance/Non-response",
TRUE ~ as.character(clinical_significance))) |>
dplyr::mutate(disease_ontology_id = paste0(
"DOID:", disease_ontology_id)) |>
dplyr::mutate(
evidence_level =
dplyr::case_when(
evidence_level == 'A' ~ 'A: Validated',
evidence_level == 'B' ~ 'B: Clinical evidence',
evidence_level == 'C' ~ 'C: Case study',
evidence_level == 'D' ~ 'D: Preclinical evidence',
evidence_level == 'E' ~ 'E: Indirect evidence',
TRUE ~ as.character(evidence_level)
)
) |>
dplyr::mutate(
variant_origin =
dplyr::if_else(
variant_origin == "N/A" |
variant_origin == "Unknown",
as.character(NA),
as.character(
stringr::str_replace(
variant_origin,"Rare |Common ",""))))
therapeutic_contexts <- as.data.frame(
clinicalEvidenceSummary |>
dplyr::select(evidence_id, therapies) |>
dplyr::filter(!is.na(therapies) &
nchar(therapies) > 0) |>
tidyr::separate_rows(therapies, sep = ",") |>
dplyr::mutate(therapies = tolower(therapies)) |>
dplyr::left_join(
dplyr::select(compound_synonyms,
alias_lc,
drug_name,
molecule_chembl_id),
by = c("therapies" = "alias_lc"),
multiple = "all",
relationship = "many-to-many"
) |>
dplyr::filter(!is.na(drug_name)) |>
dplyr::select(evidence_id, molecule_chembl_id) |>
dplyr::distinct() |>
dplyr::group_by(evidence_id) |>
dplyr::summarise(molecule_chembl_id = paste(
unique(sort(molecule_chembl_id)), collapse = "|"),
.groups = "drop"
) |>
dplyr::ungroup() |>
dplyr::mutate(molecule_chembl_id = stringr::str_replace(
molecule_chembl_id,"^NA\\||\\|NA$",""
))
)
clinicalEvidenceSummary <- clinicalEvidenceSummary |>
dplyr::left_join(
therapeutic_contexts,
by = "evidence_id", multiple = "all",
relationship = "many-to-many")
variantSummary <- as.data.frame(
data.table::fread(civic_local_fnames[['VariantSummaries']],
select = c(1:25), fill = T)) |>
## Ignore the "Factor" feature type (Kataegis, CK, methylation)
dplyr::filter(feature_type != "Factor") |>
#dplyr::rename(molecular_profile_id = variant_id) |>
dplyr::mutate(alteration_type = "MUT") |>
dplyr::mutate(variant = stringr::str_trim(variant)) |>
dplyr::mutate(variant = dplyr::if_else(
!is.na(variant) &
(variant == "Fusion" | '::' %in% variant),
as.character(feature_name),
as.character(variant)
)) |>
dplyr::mutate(variant_aliases = dplyr::if_else(
gene == "BRAF" & variant == "V600",
paste(variant_aliases, "V640", sep = ","),
as.character(variant_aliases)
)) |>
dplyr::mutate(variant_aliases = dplyr::if_else(
!is.na(hgvs_descriptions) &
hgvs_descriptions != "NA" &
nchar(hgvs_descriptions) > 0,
paste(variant_aliases, hgvs_descriptions, sep=","),
as.character(variant_aliases))) |>
dplyr::select(-c("hgvs_descriptions")) |>
dplyr::mutate(variant_aliases = stringr::str_replace(
variant_aliases, "^,", ""
)) |>
tidyr::separate_rows(variant_aliases, sep = ",") |>
tidyr::separate_rows(variant, sep=",") |>
clean_variant_naming(col = "variant") |>
clean_variant_naming(col = "variant_aliases") |>
dplyr::rename(variant_consequence = variant_types,
molecular_profile_id =
single_variant_molecular_profile_id,
variant_alias = variant_aliases) |>
clean_variant_consequence() |>
#dplyr::filter(variant != "Alu insertion") |>
dplyr::filter(
!stringr::str_detect(
tolower(variant),
biomarker_names_skip_regex)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
feature_type == "Fusion" &
nchar(variant_alias) == 0 |
stringr::str_detect(
variant_alias,"multiple partners"),
paste0(
stringr::str_replace(
feature_name,"v::|::v","")," Fusion"),
as.character(variant_alias)
)) |>
dplyr::mutate(variant = dplyr::if_else(
nchar(variant_alias) > 0 &
feature_type == "Fusion" &
stringr::str_detect(variant,"^e[0-9]{1,}::"),
stringr::str_replace(variant_alias," E[0-9]{1,}-E[0-9]{1,}$",""),
as.character(variant))) |>
dplyr::mutate(gene = dplyr::if_else(
feature_type == "Fusion",
as.character(variant),
as.character(gene)
)) |>
set_alteration_type() |>
dplyr::select(variant_id,
molecular_profile_id,
variant_consequence,
variant,
alteration_type,
gene,
variant_alias) |>
dplyr::mutate(
variant = stringr::str_replace_all(
variant,
paste0(
"((Inactivating|Activating|Deleterious|",
"Promoter|Truncating|TKD) Mutation)|Loss-of-function"),
"Mutation"
)
) |>
dplyr::mutate(variant_alias = paste(
variant_alias, variant, sep=","
)) |>
tidyr::separate_rows(
variant_alias, sep=","
) |>
dplyr::filter(
!stringr::str_detect(
tolower(variant_alias),
biomarker_names_skip_regex)) |>
dplyr::filter(
(!(stringr::str_detect(variant,"H$") &
stringr::str_detect(toupper(variant_alias),"TYR$"))) &
(!(stringr::str_detect(variant,"Y$") &
stringr::str_detect(toupper(variant_alias),"HIS$")))
) |>
dplyr::distinct() |>
dplyr::mutate(
variant_name_primary = dplyr::case_when(
alteration_type == "CNA" |
(alteration_type == "MUT_LOF" & !endsWith(variant,"X")) |
stringr::str_detect(
variant,
"Mutation|Wildtype|Splice|Copy Number|Deletion|Amplification") |
stringr::str_detect(alteration_type,"^EXP") ~
paste0(gene," ",stringr::str_to_title(variant)),
variant_consequence == "transcript_fusion" |
(nchar(variant_consequence) == 0 & variant == "FUSION") ~
paste0(gene, " Fusion"),
variant == "REARRANGEMENT" ~
paste0(gene, " Rearrangement"),
TRUE ~ as.character(
paste0(gene, " ", variant)
))
) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant,"^aa_region:"),
as.character("missense_variant"),
as.character(variant_consequence))) |>
dplyr::filter(!nchar(variant_alias) == 0) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "FS","fs")
) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "DELINS","delins")
) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "DEL","del")
) |>
dplyr::mutate(variant_name_primary = stringr::str_replace(
variant_name_primary, "INS$","ins")
)
variants_expanded <- list()
variants_expanded[['mut']] <- data.frame()
for (i in unique(variantSummary$variant_id)) {
vrows <- variantSummary[variantSummary$variant_id == i,]
variant_entries <- unique(c(vrows$variant,
vrows$variant_alias))
gene <- unique(vrows$gene)
variant_consequence <- unique(vrows$variant_consequence)
alteration_type <- unique(vrows$alteration_type)
variant_name_primary <- unique(vrows$variant_name_primary)
all_aliases <- c()
if (stringr::str_detect(alteration_type,"MUT")) {
for (e in variant_entries) {
#cat(e,'\n')
clean_aliases <- expand_hgvs_terms(e, aa_dict = aa_dict)
for (c in clean_aliases) {
all_aliases <- unique(c(all_aliases, c))
}
}
for (variant_alias in all_aliases) {
df <- data.frame(
'variant_id' = i,
'symbol' = gene,
'variant_name_primary' = variant_name_primary,
'variant_alias' = variant_alias,
'alteration_type' = alteration_type,
'variant_consequence' = variant_consequence,
stringsAsFactors = F)
variants_expanded[['mut']] <- dplyr::bind_rows(
variants_expanded[['mut']], df
)
}
}
}
variants_expanded[['mut']] <- variants_expanded[['mut']] |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
stringr::str_detect(
variant_alias,
"^(p\\.)?(([A-Z]{1}[0-9]{1,})|([A-Z]{1}[a-z]{2}[0-9]{1,}))$"),
as.character("CODON"),
as.character(alteration_type)
)) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"FS"),
stringr::str_replace(variant_alias,"FS","fs"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"DEL") &
!stringr::str_detect(variant_alias, "DELETION|DELETERIOUS"),
stringr::str_replace(variant_alias,"DEL","del"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"INS") &
!stringr::str_detect(variant_alias,"INSERTION"),
stringr::str_replace(variant_alias,"INS","ins"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"TER|ter") &
!stringr::str_detect(
tolower(variant_alias),
"alteration"),
stringr::str_replace(variant_alias,"ter|TER","Ter"),
as.character(variant_alias)
)
) |>
dplyr::mutate(
variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias,"DUP") &
!stringr::str_detect(variant_alias,"DUPLICATION"),
stringr::str_replace(variant_alias,"DUP","dup"),
as.character(variant_alias)
)
) |>
dplyr::filter(
variant_alias != "p.=" &
!stringr::str_detect(variant_alias, "^p\\.\\(") &
!stringr::str_detect(variant_alias, "^p\\.Exon") &
!stringr::str_detect(variant_alias, "ENSP|ENST") &
!stringr::str_detect(variant_alias, "SPLICE"))
aa_dict <- get_amino_acid_dictionary()
for (i in 1:nrow(aa_dict)) {
aa <- aa_dict[i,"three_letter_uc"]
aa_three_lc <- aa_dict[i, "three_letter"]
if (nrow(variants_expanded[['mut']][!is.na(variants_expanded[['mut']]$variant_alias) &
stringr::str_detect(variants_expanded[['mut']]$variant_alias,aa),]) > 0) {
variants_expanded[['mut']][!is.na(variants_expanded[['mut']]$variant_alias) &
stringr::str_detect(variants_expanded[['mut']]$variant_alias, aa),]$variant_alias <-
stringr::str_replace_all(variants_expanded[['mut']][!is.na(variants_expanded[['mut']]$variant_alias) &
stringr::str_detect(variants_expanded[['mut']]$variant_alias, aa),]$variant_alias,
aa,
aa_three_lc)
}
}
variants_expanded[['mut']] <- variants_expanded[['mut']] |>
dplyr::mutate(alias_type = "hgvsp") |>
dplyr::mutate(variant_alias = stringr::str_trim(variant_alias)) |>
dplyr::mutate(alias_type = dplyr::case_when(
stringr::str_detect(variant_alias, "^rs[0-9]{1,}$") ~ "dbsnp",
stringr::str_detect(variant_alias, "^aa_region") ~ "aa_region",
stringr::str_detect(variant_alias, "^c\\.[0-9]{1,}") ~ "hgvsc",
stringr::str_detect(variant_alias, "^Exon ") ~ "exon",
stringr::str_detect(
variant_alias,
"Mutation|Copy|Wildtype|Loss|Truncation|Alteration|Frameshift") ~ "other_gene",
TRUE ~ as.character(alias_type)
)) |>
dplyr::left_join(variant_coordinates,
by = "variant_id",
relationship = "many-to-many") |>
dplyr::mutate(variant_alias = dplyr::if_else(
!is.na(vcf_coord),
paste(variant_alias, vcf_coord, sep=","),
as.character(variant_alias)
)) |>
tidyr::separate_rows(
variant_alias, sep = ","
) |>
dplyr::select(-vcf_coord) |>
dplyr::distinct() |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch37"),
"genomic_grch37",
as.character(alias_type)
)) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch38"),
"genomic_grch38",
as.character(alias_type)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch37"),
stringr::str_replace(variant_alias, "grch37:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch38"),
stringr::str_replace(variant_alias, "grch38:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(alteration_type = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic"),
"MUT",
as.character(alteration_type)
)) |>
dplyr::filter(
symbol != "TERT" |
(symbol == "TERT" &
!stringr::str_detect(variant_alias, "Cys|^p\\.") &
alteration_type != "CODON")) |>
dplyr::mutate(variant_alias = dplyr::if_else(
symbol == "TERT" & variant_alias == "-146C>T",
paste(variant_alias,
"grch37:5_1295250_G_A,grch38:5_1295135_G_A",
sep=","),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
symbol == "MET" &
variant_alias == "Exon 14 Skipping Mutation",
paste(
variant_alias,
"c.3028+1G>T",
"c.3028+1G>C",
"c.3028+1G>A",
"c.3028+2T>C",
"c.3028+2T>G",
"c.3082+1G>T",
"c.3082+1G>C",
"c.3082+1G>A",
"c.3082+2T>C",
"c.3082+2T>G",
sep=","),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
symbol == "TERT" &
variant_alias == "-124C>T",
paste(variant_alias,
"grch37:5_1295228_G_A,grch38:5_1295113_G_A",
sep=","),
as.character(variant_alias)
)) |>
tidyr::separate_rows(
variant_alias, sep = ","
) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch37"),
"genomic_grch37",
as.character(alias_type)
)) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^grch38"),
"genomic_grch38",
as.character(alias_type)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch37"),
stringr::str_replace(variant_alias, "grch37:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic_grch38"),
stringr::str_replace(variant_alias, "grch38:",""),
as.character(variant_alias)
)) |>
dplyr::mutate(alteration_type = dplyr::if_else(
stringr::str_detect(alias_type, "^genomic"),
"MUT",
as.character(alteration_type)
)) |>
dplyr::mutate(alias_type = dplyr::if_else(
stringr::str_detect(variant_alias, "^c\\."),
"hgvsc",
as.character(alias_type)
)) |>
dplyr::mutate(variant_exon = dplyr::if_else(
stringr::str_detect(variant_alias, "Exon"),
stringr::str_replace_all(
variant_alias,
"Exon |( (Frameshift|((Skipping )?Mutation)|Deletion|Insertion))",""),
as.character(NA)
)) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,"18-21",
"18|19|20|21"),
as.character(variant_exon))
) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,
"1-2","1|2"),
as.character(variant_exon))
) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,
"10 OR 21","10|21"),
as.character(variant_exon))
) |>
dplyr::mutate(variant_exon = dplyr::if_else(
!is.na(variant_exon) &
variant_consequence == "exon_variant",
stringr::str_replace_all(
variant_exon,
"10 AND 21","10&21"),
as.character(variant_exon))
) |>
dplyr::filter(
is.na(variant_exon)
| (variant_exon != "9 AND 20" &
!stringr::str_detect(
variant_exon,"Deletion"))) |>
dplyr::distinct()
## Make sure alteration types for genomic coordinates
## are correct
variants_nongenomic <- variants_expanded[['mut']] |>
dplyr::filter(
!stringr::str_detect(
alias_type, "genomic_"
)
) |>
dplyr::group_by(
variant_id
) |>
dplyr::summarise(
alteration_type =
paste(unique(alteration_type), collapse=";")
) |>
dplyr::mutate(
alteration_type = dplyr::if_else(
stringr::str_detect(alteration_type, ";"),
stringr::str_replace(
alteration_type, ";?CODON;?",""
),
as.character(alteration_type)
)
)
variants_expanded_mut1 <- variants_expanded[['mut']] |>
dplyr::filter(!stringr::str_detect(
alias_type, "genomic"))
variants_expanded_mut2 <- variants_expanded[['mut']] |>
dplyr::filter(stringr::str_detect(
alias_type, "genomic"))
variants_expanded_mut2$alteration_type <- NULL
variants_expanded_mut2 <- variants_expanded_mut2 |>
dplyr::left_join(variants_nongenomic, by = "variant_id",
relationship = "many-to-many")
variants_expanded[['mut']] <- variants_expanded_mut1 |>
dplyr::bind_rows(variants_expanded_mut2)
## Combine the two dataframes
variants_expanded[['other']] <- as.data.frame(
variantSummary |>
dplyr::filter(!stringr::str_detect(alteration_type,"MUT")) |>
dplyr::select(variant_id, variant, variant_consequence,
variant_alias, variant_name_primary, alteration_type,
gene) |>
dplyr::rename(symbol = gene) |>
dplyr::mutate(
alias_type = "other_gene"
) |>
dplyr::distinct()
)
attr(variants_expanded[['other']],"groups") <- NULL
gene_aliases <- gene_alias$records |>
dplyr::filter(.data$n_primary_map == 1 &
source == "NCBI") |>
dplyr::select(alias, symbol, entrezgene) |>
dplyr::distinct()
biomarker_items <- list()
biomarker_items[['variant']] <-
dplyr::bind_rows(
variants_expanded[['mut']],
variants_expanded[['other']]) |>
dplyr::select(variant_id, variant_alias,
alias_type, dplyr::everything()) |>
dplyr::rename(gene = symbol) |>
dplyr::left_join(
gene_aliases,
by = c("gene" = "alias"),
relationship = "many-to-one"
) |>
dplyr::mutate(variant_consequence = dplyr::case_when(
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "delins([A-Z]{1,4})+$") ~ "inframe_insertion",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "ins([A-Z]{1,4})+$") ~ "inframe_insertion",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "[0-9]{1,}?del([A-Z]{1,4})+$") ~ "inframe_deletion",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "fs") ~ "frameshift_variant",
nchar(variant_consequence) == 0 & stringr::str_detect(
variant_alias, "^Mutation$") ~ "protein_altering_variant",
nchar(variant_consequence) == 0 &
alias_type == "hgvsp" ~ "missense_variant",
TRUE ~ as.character(variant_consequence)
)) |>
dplyr::mutate(variant_alias = dplyr::case_when(
(is.na(variant_alias) | nchar(variant_alias) == 0) &
!is.na(gene) &
alias_type == "other_gene" &
stringr::str_detect(gene,"-") ~ gene,
(is.na(variant_alias) | nchar(variant_alias) == 0) &
!is.na(gene) &
alias_type == "other_gene" &
!stringr::str_detect(gene,"-") ~
paste0(gene, " Fusion"),
TRUE ~ as.character(variant_alias)
)) |>
dplyr::distinct()
biomarker_items[['clinical']] <-
clinicalEvidenceSummary |>
dplyr::full_join(
molecularProfileSummary, by = "molecular_profile_id",
multiple = "all", relationship = "many-to-many") |>
dplyr::select(molecular_profile_id, molecular_profile_name,
molecular_profile_summary,
variant_id, dplyr::everything()) |>
dplyr::rename(therapeutic_context = therapies) |>
dplyr::mutate(evidence_id = paste0("EID",evidence_id)) |>
dplyr::filter(evidence_id != "EIDNA") |>
dplyr::mutate(biomarker_source = "civic",
biomarker_source_datestamp = datestamp,
biomarker_entity = T) |>
dplyr::inner_join(dplyr::select(
biomarker_items[['variant']], variant_id),
by = "variant_id",
multiple = "all", relationship = "many-to-many") |>
dplyr::distinct() |>
dplyr::arrange(evidence_level, molecular_profile_id)
biomarker_items[['clinical']] <- map_biomarker_phenotypes(
biomarker_items[['clinical']],
cache_dir = cache_dir)
biomarker_items <- get_literature_references(biomarker_items)
if(is.data.frame(biomarker_items)){
if(unique(colnames(biomarker_items) == c("source","source_id")) == T){
lgr::lgr$error("literature missing ")
return(biomarker_items)
}
}
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::mutate(molecular_profile_type = "Single") |>
dplyr::mutate(molecular_profile_type = dplyr::case_when(
stringr::str_detect(molecular_profile_name, " AND|and|And ") ~
paste0(
"Combo_", stringr::str_count(variant_id, " AND|and|And ")[1] + 1),
stringr::str_detect(molecular_profile_name, " OR|or|Or ") ~
paste0(
"Combo_Opt_", stringr::str_count(variant_id, ",")[1] + 1),
TRUE ~ molecular_profile_type
)) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
biomarker_entity,
molecular_profile_id,
molecular_profile_name,
molecular_profile_summary,
molecular_profile_type,
variant_id,
source_id,
evidence_id,
variant_origin,
primary_site,
cancer_type,
disease_ontology_id,
do_name,
efo_id,
efo_name,
cui,
cui_name,
dplyr::everything()
)
biomarker_items[['variant']] <- as.data.frame(
biomarker_items[['variant']] |>
dplyr::mutate(biomarker_source = "civic",
biomarker_source_datestamp = datestamp) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
variant_name_primary,
alteration_type,
alias_type,
variant_consequence,
variant_exon,
gene,
symbol,
entrezgene,
dplyr::everything()
) |>
dplyr::inner_join(
dplyr::select(biomarker_items[['clinical']],
variant_id, molecular_profile_id),
by = "variant_id",
relationship = "many-to-many"
) |>
dplyr::distinct() |>
dplyr::select(-c("variant")) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
variant_id,
molecular_profile_id,
dplyr::everything()
)
)
return(biomarker_items)
}
load_cgi_biomarkers <- function(compound_synonyms = NULL,
datestamp = '20221027',
cache_dir = NA) {
aa_dict <- get_amino_acid_dictionary()
gene_alias <- geneOncoX::get_alias(
cache_dir = cache_dir
)
cancer_type_abbreviations <-
read.table(
file = file.path(cache_dir, "cgi_cancer_subtypes.tsv"),
sep = "\t", header = T,
stringsAsFactors = F, quote = "",
na.strings = "",fill = T) |>
dplyr::mutate(
disease_ontology_id = paste0(
"DOID:",disease_ontology_id))
## SNVs/InDels
## SKIP/IGNORE mutations that are
# 1) not explicitly mapped to protein alterations (MUT)
# 2) marked with uncertainty, rechecking etc (Comments)
# 3) involved in toxicity
# 4) curated from abstracts etc (not PMIDs or guidelines)
# 5) Non-malignant/Non-cancer conditions
## NEW
cgi_biomarkers <- as.data.frame(
readr::read_tsv(
file.path(cache_dir, "cgi_biomarkers_20221017.tsv"),
col_names = T, show_col_types = F, guess_max = 1000) |>
janitor::clean_names() |>
dplyr::mutate(
biomarker =
stringr::str_replace_all(
biomarker,
"\\.",""
)) |>
dplyr::select(-c(curator, curation_date, targeting,
drug_family, drug_status, drug, tcgi_included,
metastatic_tumor_type,
primary_tumor_type_full_name, assay_type)) |>
dplyr::distinct() |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "12,13,59,61,117,146",
"G12,G13,A59,Q61,K117,A146")) |>
dplyr::mutate(alteration_type = stringr::str_replace(
alteration_type, "EXPR", "EXP"
)) |>
dplyr::mutate(alteration_type = stringr::str_replace(
alteration_type, "FUS", "FUSION"
)) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "\\*","X")) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "618,620,634,768,791,891,918",
"C618,C620,C634,E768,Y791,S891,M918")) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, "12,13","G12,G13")) |>
dplyr::mutate(biomarker = stringr::str_replace(
biomarker, "undexpression","Underexpression")) |>
dplyr::mutate(biomarker = stringr::str_replace_all(
biomarker, ",537,538,",",Y537,D538,")) |>
dplyr::rename(
molecular_profile = biomarker
) |>
dplyr::mutate(molecular_profile = stringr::str_replace_all(
molecular_profile, ";|(, )", ","
)) |>
dplyr::mutate(molecular_profile = stringr::str_replace_all(
molecular_profile, " (\\+|-)$", ""
)) |>
dplyr::filter(
!stringr::str_detect(source, "^(ENA|ESMO|ASCO|AACR)")
) |>
tidyr::separate_rows(
primary_tumor_type, sep = ";") |>
dplyr::left_join(
cancer_type_abbreviations,
by = c("primary_tumor_type" = "cgi_abbreviation"),
relationship = "many-to-many") |>
dplyr::mutate(molecular_profile_id = dplyr::row_number()) |>
dplyr::mutate(molecular_profile_type = dplyr::if_else(
stringr::str_detect(molecular_profile, " \\+ "),
"Combo",
"Single"
)) |>
dplyr::mutate(molecular_profile_name = molecular_profile) |>
tidyr::separate_longer_delim(molecular_profile, delim = " + ")
)
combo_index <- cgi_biomarkers |>
dplyr::group_by(molecular_profile_id) |>
dplyr::reframe(combo_index = dplyr::row_number()) |>
dplyr::select(-molecular_profile_id)
cgi_biomarkers <- as.data.frame(
cgi_biomarkers |>
dplyr::bind_cols(combo_index) |>
dplyr::mutate(molecular_profile_type = dplyr::if_else(
molecular_profile_type == "Combo",
paste(
molecular_profile_type, combo_index, sep="_"),
as.character(molecular_profile_type)
)) |>
dplyr::select(-combo_index) |>
dplyr::mutate(alteration2 =
stringr::str_replace_all(
stringr::str_match(
molecular_profile, "\\([A-Z0-9]+(,[A-Z0-9]+){0,}\\)")[,1],
"\\)|\\(","")
) |>
dplyr::mutate(alteration_type = dplyr::case_when(
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " amplification") ~ "CNA",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " deletion") &
!stringr::str_detect(molecular_profile, "inframe") ~ "CNA",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " fusion") ~ "FUSION",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(tolower(molecular_profile), "overexpression") ~ "EXP_OVER",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(tolower(molecular_profile), "underexpression") ~ "EXP_OVER",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(molecular_profile, " expression") ~ "EXP",
stringr::str_detect(molecular_profile_type, "Combo") &
stringr::str_detect(alteration_type,"MUT") &
stringr::str_detect(tolower(molecular_profile), "oncogenic|\\(") ~ "MUT_ONC",
TRUE ~ as.character(alteration_type)
)) |>
tidyr::separate(gene, c("gene1","gene2"), sep = ";") |>
dplyr::mutate(gene = dplyr::if_else(
!is.na(gene1) & !is.na(gene2) &
startsWith(molecular_profile, gene2),
as.character(gene2),
as.character(gene1)
)) |>
dplyr::select(-c(gene1,gene2)) |>
dplyr::select(molecular_profile, gene, dplyr::everything()) |>
dplyr::mutate(alteration_type = dplyr::case_when(
alteration_type == "BIA" ~ "MUT_LOF_BIALLELIC",
stringr::str_detect(
molecular_profile, " wildtype") ~ "WT",
stringr::str_detect(
molecular_profile, " oncogenic") ~ "MUT_ONC",
stringr::str_detect(
tolower(molecular_profile), "overexpression") ~ "EXP_OVER",
stringr::str_detect(
tolower(molecular_profile), "underexpression") ~ "EXP_UNDER",
stringr::str_detect(
molecular_profile, " fusion$") ~ "FUSION",
TRUE ~ as.character(alteration_type)
)) |>
dplyr::mutate(variant_consequence = dplyr::case_when(
alteration_type == "WT" ~ as.character(NA),
alteration_type == "CNA" &
stringr::str_detect(
alteration,":amp") ~ "transcript_amplification",
alteration_type == "CNA" &
stringr::str_detect(
alteration,":del") ~ "transcript_ablation",
alteration_type == "EXP_OVER" ~ "transcript_amplification",
alteration_type == "FUSION" ~ "transcript_fusion",
alteration_type == "EXP_UNDER" ~ "transcript_ablation",
alteration_type == "MUT_LOF_BIALLELIC" ~ "loss_of_function_variant",
alteration_type == "MUT_ONC" ~ "protein_altering_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
!stringr::str_detect(molecular_profile,"inframe|splice|insertion|deletion") ~ "missense_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"splice donor") ~ "splice_donor_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"splice acceptor") ~ "splice_acceptor_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"inframe") &
stringr::str_detect(molecular_profile,"insertion") &
stringr::str_detect(molecular_profile, "deletion") ~ "inframe_deletion,inframe_insertion",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"inframe") &
stringr::str_detect(molecular_profile, "deletion") ~ "inframe_deletion",
molecular_profile == "FLT3-ITD" ~ "inframe_insertion",
molecular_profile == "PDGFRA exon 18 mutations" ~ "inframe_deletion,inframe_insertion,missense_variant",
(alteration_type == "MUT" | alteration_type == "MUT_ONC") &
stringr::str_detect(molecular_profile,"inframe") &
stringr::str_detect(molecular_profile, "insertion") ~ "inframe_insertion",
(alteration_type == "MUT") &
stringr::str_detect(molecular_profile, " exon ") &
stringr::str_detect(alteration, "inframe_insertion") ~ "inframe_insertion",
(alteration_type == "MUT") &
stringr::str_detect(molecular_profile, " exon ") &
stringr::str_detect(alteration, "inframe_deletion") ~ "inframe_deletion",
TRUE ~ as.character(NA)
)) |>
dplyr::rename(
variant_alias = alteration2
) |>
dplyr::mutate(variant_alias = dplyr::case_when(
alteration_type == "EXP" ~ "Expression",
stringr::str_detect(molecular_profile, "wildtype") ~ "Wildtype",
is.na(variant_alias) &
(alteration_type == "MUT" |
alteration_type == "MUT_ONC" |
alteration_type == "MUT_LOF_BIALLELIC") ~ "Mutation",
alteration_type == "CNA" &
variant_consequence == "transcript_amplification" ~ "Amplification",
alteration_type == "CNA" &
variant_consequence == "transcript_ablation" ~ "Deletion",
alteration_type == "FUSION" ~ "Fusion",
alteration_type == "EXP_UNDER" &
variant_consequence == "transcript_ablation" ~ "Underexpression",
alteration_type == "EXP_OVER" &
variant_consequence == "transcript_amplification" ~ "Overexpression",
TRUE ~ as.character(variant_alias)
)) |>
dplyr::mutate(gene = dplyr::if_else(
alteration_type == "FUSION" &
stringr::str_detect(molecular_profile, "-"),
stringr::str_replace(
stringr::str_replace(molecular_profile, " fusion",""),
"ABL1-BCR","BCR-ABL1"),
as.character(gene))
) |>
dplyr::mutate(gene = dplyr::case_when(
gene == "MLL2" ~ "KMT2D",
TRUE ~ as.character(gene)
)) |>
dplyr::mutate(variant_origin = "Somatic") |>
dplyr::mutate(variant_origin = dplyr::if_else(
!is.na(comments) &
stringr::str_detect(tolower(comments),"germline"),
"Germline",
as.character(variant_origin)
)) |>
dplyr::mutate(evidence_type = 'Predictive',
therapeutic_context = drug_full_name,
clinical_significance = NA) |>
dplyr::mutate(
clinical_significance =
dplyr::case_when(
association == "Resistant" ~ "Resistance/Non-response",
association == "No Responsive" ~ "Resistance/Non-response",
association == "Responsive" ~ "Sensitivity/Response",
stringr::str_detect(association, "Toxicity") ~ "Toxicity",
TRUE ~ as.character("Other"))) |>
dplyr::rename(evidence_level_raw = evidence_level) |>
dplyr::mutate(mapping_rank = 1) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"guidelines"),
"A: FDA/NCCN/ELN guidelines",
as.character(NA))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Late trials"),
"B1: Clinical evidence: late trials",
as.character(evidence_level))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Early trials"),
"B2: Clinical evidence: early trials",
as.character(evidence_level))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Case report"),
"C: Case study",
as.character(evidence_level))) |>
dplyr::mutate(
evidence_level =
dplyr::if_else(
stringr::str_detect(evidence_level_raw,"Pre-clinical"),
"D: Preclinical evidence",
as.character(evidence_level))) |>
dplyr::filter(!is.na(evidence_level_raw)) |>
dplyr::select(-comments) |>
dplyr::mutate(alteration = dplyr::if_else(
molecular_profile == "KIT mutation in exon 9,11,13,14 or 17",
"KIT:449-514,550-592,627-664,664-714,788-828",
as.character(alteration)
)) |>
dplyr::mutate(alteration = dplyr::if_else(
stringr::str_detect(molecular_profile, " exon |-ITD$"),
stringr::str_replace_all(
alteration,
"KIT:|([A-Z0-9]{3,}::consequence::(skipping_mutation|(inframe_(insertion|variant|deletion))):)",
""
),
as.character(alteration)
)) |>
dplyr::mutate(alteration = dplyr::if_else(
stringr::str_detect(molecular_profile, " exon |-ITD$"),
stringr::str_replace_all(
alteration, "::(inframe_insertion|missense_variant):", ""),
as.character(alteration)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(molecular_profile, " exon |-ITD$") &
!is.na(alteration),
as.character(alteration),
as.character(variant_alias)
)) |>
tidyr::separate_longer_delim(variant_alias, delim = ",") |>
dplyr::distinct() |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(variant_alias, "-"),
paste0("aa_region:", variant_alias),
as.character(variant_alias)
)) |>
dplyr::mutate(variant_alias = dplyr::if_else(
stringr::str_detect(
molecular_profile, "^(IL7R inframe insertion \\(237)"),
"aa_region:237-255",
as.character(variant_alias)
)) |>
dplyr::left_join(
dplyr::select(gene_alias$records,
alias, symbol, entrezgene),
by = c("gene" = "alias"), multiple = "all",
relationship = "many-to-many"
) |>
dplyr::mutate(alteration = stringr::str_replace_all(
alteration,"__\\.","::v")
) |>
dplyr::mutate(variant_consequence = dplyr::if_else(
stringr::str_detect(variant_alias, "X$"),
"stop_gained",
as.character(variant_consequence)
)) |>
dplyr::mutate(alteration_type = dplyr::if_else(
stringr::str_detect(
variant_consequence,
"stop_gained|splice_donor|splice_acceptor"),
"MUT_LOF",
as.character(alteration_type)
))
)
unique_variants <- as.data.frame(
cgi_biomarkers |>
dplyr::filter(!is.na(variant_alias) &
!is.na(gene) &
!is.na(alteration_type)) |>
dplyr::select(gene, symbol, entrezgene, variant_alias,
alteration_type, variant_consequence) |>
dplyr:: arrange(symbol, alteration_type, variant_alias) |>
dplyr::distinct() |>
dplyr::mutate(variant_id = dplyr::row_number()) |>
dplyr::mutate(variant_name_primary = dplyr::case_when(
!is.na(symbol) ~ paste(symbol, variant_alias),
is.na(symbol) &
!is.na(gene) ~ paste(gene, variant_alias),
TRUE ~ as.character(NA)
))
)
cgi_variants <- data.frame()
for (i in 1:nrow(unique_variants)) {
alt <- unique_variants[i,"variant_alias"]
terms <- expand_hgvs_terms(alt, aa_dict = aa_dict)
for (t in terms) {
df <- data.frame(
'variant_id' =
unique_variants[i,"variant_id"],
'variant_alias' = t,
'variant_name_primary' = unique_variants[i,"variant_name_primary"],
'gene' = unique_variants[i,"gene"],
'entrezgene' = unique_variants[i,"entrezgene"],
'symbol' = unique_variants[i,"symbol"],
'variant_consequence' = unique_variants[i,"variant_consequence"],
'alteration_type' = unique_variants[i,"alteration_type"],
stringsAsFactors = F)
cgi_variants <- cgi_variants |>
dplyr::bind_rows(df)
}
}
cgi_variants <- cgi_variants |>
dplyr::mutate(alias_type = "hgvsp") |>
dplyr::mutate(alias_type = dplyr::case_when(
stringr::str_detect(variant_alias, "aa") ~ "aa_region",
stringr::str_detect(variant_alias, "^(Mut|Wild|Exp|Overexp|Underexp|Fus|Ampli|Delet)") ~ "other_gene",
TRUE ~ as.character(alias_type)
)) |>
dplyr::mutate(variant_exon = NA) |>
dplyr::mutate(biomarker_source = "cgi") |>
dplyr::mutate(biomarker_source_datestamp = datestamp) |>
dplyr::mutate(
alteration_type =
dplyr::if_else(
!is.na(alteration_type) & !is.na(variant_alias) &
stringr::str_detect(
variant_alias,
"^(([A-Z]{1}[0-9]{1,})|([A-Z]{1}[a-z]{2}[0-9]{1,}))$"),
as.character("CODON"),
as.character(alteration_type)
)) |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
variant_name_primary,
alteration_type,
alias_type,
variant_consequence,
variant_exon,
gene,
symbol,
entrezgene,
) |>
dplyr::distinct()
cgi_clinical <- cgi_biomarkers |>
## get variant id's per molecular profile
dplyr::inner_join(
dplyr::select(
cgi_variants, variant_id,
gene, alias_type, variant_alias),
by = c("gene", "variant_alias"),
relationship = "many-to-many"
) |>
dplyr::select(-c(entrezgene, variant_consequence, variant_alias,
alias_type, gene, symbol, alteration_type,
alteration)) |>
dplyr::distinct() |>
dplyr::mutate(therapeutic_context = dplyr::if_else(
stringr::str_detect(
therapeutic_context,
"(ib|ide|ole|ant|stat|ine|us|one|mab|in|el|[0-9])( | )\\("),
stringr::str_replace(
therapeutic_context,
"( | )\\(.*\\)( )?$",""),
as.character(therapeutic_context)
))
therapeutic_contexts <- as.data.frame(
cgi_clinical |>
dplyr::select(molecular_profile_id, therapeutic_context) |>
dplyr::filter(!is.na(therapeutic_context) &
nchar(therapeutic_context) > 0) |>
tidyr::separate_rows(therapeutic_context, sep = " \\+ ") |>
dplyr::mutate(therapeutic_context = tolower(therapeutic_context)) |>
dplyr::left_join(
dplyr::select(compound_synonyms,
alias_lc,
drug_name,
molecule_chembl_id),
by = c("therapeutic_context" = "alias_lc"),
multiple = "all", relationship = "many-to-many"
) |>
dplyr::filter(!is.na(drug_name)) |>
dplyr::select(molecular_profile_id, molecule_chembl_id) |>
dplyr::distinct() |>
dplyr::group_by(molecular_profile_id) |>
dplyr::summarise(molecule_chembl_id = paste(
unique(sort(molecule_chembl_id)), collapse="|"),
.groups = "drop"
) |>
dplyr::ungroup() |>
dplyr::mutate(molecule_chembl_id = stringr::str_replace(
molecule_chembl_id,"^NA\\||\\|NA$",""
))
)
cgi_clinical <- cgi_clinical |>
dplyr::left_join(
therapeutic_contexts, by = "molecular_profile_id",
multiple = "all", relationship = "many-to-many") |>
dplyr::mutate(evidence_url = "https://www.cancergenomeinterpreter.org/biomarkers") |>
dplyr::mutate(biomarker_source = "cgi") |>
dplyr::mutate(biomarker_source_datestamp = datestamp) |>
dplyr::mutate(evidence_description = NA,
evidence_id = paste0("CGI", molecular_profile_id),
molecular_profile_summary = NA,
rating = NA,
biomarker_entity = T) |>
dplyr::select(-c(association, evidence_level_raw,
primary_tumor_type, drug_full_name)) |>
dplyr::rename(cancer_type = disease_name,
citation_id = source)
cgi_clinical <- map_biomarker_phenotypes(
biomarkers_clinical = cgi_clinical,
cache_dir = cache_dir) |>
dplyr::distinct()
biomarker_items <- list()
biomarker_items[['clinical']] <- cgi_clinical
biomarker_items[['clinical']]$molecular_profile <- NULL
biomarker_items <- get_literature_references(biomarker_items)
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::select(
biomarker_source,
biomarker_source_datestamp,
biomarker_entity,
molecular_profile_id,
molecular_profile_name,
molecular_profile_summary,
molecular_profile_type,
variant_id,
source_id,
evidence_id,
variant_origin,
primary_site,
cancer_type,
disease_ontology_id,
do_name,
efo_id,
efo_name,
cui,
cui_name,
dplyr::everything()
) |>
dplyr::distinct()
biomarker_items[['variant']] <- as.data.frame(
cgi_variants |>
dplyr::inner_join(
dplyr::select(
biomarker_items$clinical,
variant_id
), by = "variant_id",
relationship = "many-to-many"
) |>
dplyr::distinct()
)
return(biomarker_items)
}
map_biomarker_phenotypes <- function(biomarkers_clinical = NULL,
cache_dir = NA) {
cancer_term_map <-
phenOncoX::get_terms(cache_dir = cache_dir)$records |>
dplyr::select(do_id, efo_id, efo_name, do_name,
cui, cui_name, primary_site) |>
dplyr::filter(!is.na(do_id) & !is.na(cui)) |>
dplyr::rename(disease_ontology_id = do_id) |>
dplyr::distinct()
cancer_aux_pheno_maps <-
phenOncoX::get_aux_maps(cache_dir = cache_dir)
umls_terms_all <-
cancer_aux_pheno_maps$records$umls$concept |>
dplyr::filter(main_term == T) |>
dplyr::select(cui, cui_name)
do_terms_all <-
cancer_aux_pheno_maps$records$do |>
dplyr::select(
cui, do_id, do_name
) |>
dplyr::distinct()
efo_terms_all <-
cancer_aux_pheno_maps$records$efo$efo2xref |>
dplyr::select(
cui, efo_id, efo_name, primary_site) |>
dplyr::distinct()
biomarkers_all_phenotypes <- biomarkers_clinical |>
dplyr::left_join(
cancer_term_map, by = "disease_ontology_id",
multiple = "all", relationship = "many-to-many")
## if conditions in biomarkers is not found in phenOncoX terms
## match against individual dictionaries
missing_do_phenotypes <- biomarkers_all_phenotypes |>
dplyr::filter(is.na(cui)) |>
dplyr::select(-c(cui, cui_name, primary_site,
efo_id, do_name, efo_name)) |>
dplyr::inner_join(
do_terms_all,
by = c("disease_ontology_id" = "do_id"),
multiple = "all", relationship = "many-to-many") |>
dplyr::left_join(
umls_terms_all, by = "cui",
multiple = "all", relationship = "many-to-many") |>
dplyr::left_join(
efo_terms_all,
by = "cui", multiple = "all",
relationship = "many-to-many") |>
dplyr::distinct() |>
dplyr::mutate(primary_site = dplyr::case_when(
do_name == "cancer" ~ as.character(NA),
is.na(primary_site) & stringr::str_detect(do_name,"brain") ~ "CNS/Brain",
is.na(primary_site) & stringr::str_detect(do_name,"breast") ~ "Breast",
is.na(primary_site) & stringr::str_detect(do_name,"colon|rectum") ~ "Colon/Rectum",
is.na(primary_site) & stringr::str_detect(do_name,"gastric|stomach|esophag") ~ "Esophagus/Stomach",
is.na(primary_site) & stringr::str_detect(do_name,"prostate") ~ "Prostate",
is.na(primary_site) & stringr::str_detect(do_name,"pancrea") ~ "Pancreas",
is.na(primary_site) & stringr::str_detect(do_name,"lung") ~ "Lung",
is.na(primary_site) & stringr::str_detect(do_name,"myeloid") ~ "Myeloid",
is.na(primary_site) & stringr::str_detect(do_name,"lymphoma") ~ "Lymphoid",
is.na(primary_site) & stringr::str_detect(do_name,"ovary|ovarian") ~ "Ovary",
TRUE ~ as.character(primary_site)
))
biomarkers_phenotype_mapped <-
dplyr::filter(biomarkers_all_phenotypes, !is.na(cui)) |>
dplyr::bind_rows(missing_do_phenotypes)
return(biomarkers_phenotype_mapped)
}
load_custom_fusion_db <- function() {
other_translocations <- as.data.frame(read.table(
file = "data-raw/biomarkers/translocation_patterns.tsv", header = T,
sep = "\t", stringsAsFactors = F) |>
dplyr::select(ngram_pattern, biomarker) |>
dplyr::rename(alias = ngram_pattern,
variant = biomarker) |>
dplyr::mutate(biomarker_entity = T,
symbol = variant,
alias = toupper(alias)) |>
tidyr::separate_rows(alias, sep = "\\|") |>
dplyr::mutate(biomarker_source_db = "fusion_customdb") |>
dplyr::distinct()
)
other_translocations_lc <- as.data.frame(read.table(
file = "data-raw/biomarkers/translocation_patterns.tsv", header = T,
sep = "\t", stringsAsFactors = F) |>
dplyr::select(ngram_pattern, biomarker) |>
dplyr::rename(alias = ngram_pattern,
variant = biomarker) |>
dplyr::mutate(biomarker_entity = T,
symbol = variant) |>
tidyr::separate_rows(alias, sep = "\\|") |>
dplyr::mutate(biomarker_source_db = "fusion_customdb") |>
dplyr::distinct()
)
biomarker_items <- list()
biomarker_items[['variant']] <- other_translocations |>
dplyr::bind_rows(other_translocations_lc) |>
dplyr::distinct() |>
dplyr::rename(variant_alias = alias,
gene = symbol) |>
dplyr::mutate(alteration_type = "FUSION",
variant_consequence = "transcript_fusion",
variant_id = variant,
variant_name_primary = variant) |>
dplyr::mutate(alias_type = "other_gene") |>
dplyr::select(variant_id, variant_name_primary,
variant_alias, alias_type,
gene, alteration_type, variant_consequence)
return(biomarker_items)
}
load_depmap_fusions <- function(db_datestamp = "24Q4"){
# Load DepMap fusions
depmap_data <- list()
depmap_data[['fusions']] <- as.data.frame(read.csv(
file = "data-raw/depmap/OmicsFusionFiltered.csv", header = T))
depmap_data[['models']] <- as.data.frame(read.csv(
file = "data-raw/depmap/Model.csv", header = T)) |>
dplyr::select(
ModelID, CellLineName, OncotreeLineage,
OncotreePrimaryDisease, OncotreeCode,
Age, Sex, PrimaryOrMetastasis, SampleCollectionSite,
SourceType
)
return(depmap_data)
}
load_mitelman_db <- function(cache_dir = NA,
db_datestamp = "20241015") {
# Load Mitelman database
# dos2unix -q -n MBCA.TXT.DATA MBCA.TXT
morphology <- read.table(
file = file.path(
cache_dir, "mitelmandb", "KODER.TXT.DATA"),
sep = "\t", stringsAsFactors = F,
header = T, fill = T) |>
dplyr::filter(KodTyp == "MORPH" & nchar(Kod) > 0) |>
dplyr::mutate('morphology_code' = as.character(Kod),
'morphology_name' = Benamning) |>
dplyr::select(morphology_code, morphology_name) |>
dplyr::distinct()
topography <- read.table(
file = file.path(
cache_dir, "mitelmandb", "KODER.TXT.DATA"),
sep = "\t", stringsAsFactors = F,
header = T, fill = T) |>
dplyr::filter(KodTyp == "TOP" & nchar(Kod) > 0) |>
dplyr::mutate('topography_code' = as.character(Kod),
'topography_name' = Benamning) |>
dplyr::select(topography_code, topography_name) |>
dplyr::distinct()
pmid_data <- readr::read_tsv(
file = file.path(
cache_dir, "mitelmandb", "REF.TXT.DATA"),
quote = "none", show_col_types = F, guess_max = 1000000) |>
dplyr::select(RefNo, Pubmed) |>
dplyr::filter(nchar(Pubmed) > 0) |>
dplyr::mutate(reference_id = as.character(RefNo)) |>
dplyr::distinct() |>
dplyr::rename(source_id = Pubmed) |>
dplyr::select(reference_id, source_id) |>
dplyr::distinct()
fusion_event_data <- as.data.frame(readr::read_tsv(
file = file.path(
cache_dir, "mitelmandb", "MBCA.TXT.DATA"),
show_col_types = F, guess_max = 100000)) |>
dplyr::filter(stringr::str_detect(GeneShort,"::")) |>
dplyr::rename(variant = GeneShort,
karyotype = KaryShort,
reference_id = RefNo,
morphology_code = Morph,
topography_code = Topo) |>
dplyr::mutate(reference_id = as.character(reference_id),
morphology_code = as.character(morphology_code)) |>
dplyr::select(MolClin,
reference_id,
morphology_code,
topography_code,
variant, karyotype) |>
tidyr::separate_rows(variant, sep = ",") |>
dplyr::left_join(pmid_data, by = "reference_id",
multiple = "all",
relationship = "many-to-many") |>
dplyr::left_join(morphology, by = "morphology_code",
multiple = "all",
relationship = "many-to-many") |>
dplyr::left_join(topography, by = "topography_code",
multiple = "all",
relationship = "many-to-many") |>
dplyr::filter(stringr::str_detect(variant,"::")) |>
dplyr::mutate(variant = stringr::str_replace(
variant, "\\+", "")) |>
dplyr::mutate(evidence_id = dplyr::row_number()) |>
dplyr::mutate(evidence_id = paste0("MITDB_",evidence_id)) |>
dplyr::filter(stringr::str_count(variant,",") ==
stringr::str_count(karyotype,",")) |>
tidyr::separate_rows(c(variant,karyotype),
sep = ",") |>
dplyr::mutate(topography_name2 = dplyr::case_when(
is.na(topography_name) &
stringr::str_detect(
tolower(morphology_name), "leukemia|myeloma") ~ "Myeloid",
is.na(topography_name) &
stringr::str_detect(
tolower(morphology_name), "lymphoma") ~ "Lymphoid",
topography_name == "Ovary" ~ "Ovary/Fallopian Tube",
topography_name == "Fallopian tube" ~ "Ovary/Fallopian Tube",
topography_name == "Uterus, corpus" ~ "Uterus",
stringr::str_detect(morphology_name,
paste0(
"Nasal cavity/Paranasal sinuses|Nasopharynx|",
"LarynxOral cavity|Oro- and hypopharynx|",
"Larynx|Teeth")) ~ "Head and Neck",
topography_name == "Gallbladder/Biliary system" ~ "Biliary Tract",
topography_name == "Soft tissue" ~ "Soft Tissue",
topography_name == "Brain" ~ "CNS/Brain",
morphology_name == "Mature B-cell neoplasm, NOS" ~ "Lymphoid",
topography_name == "Brain stem" ~ "CNS/Brain",
topography_name == "Cerebellum" ~ "CNS/Brain",
topography_name == "Vagina" ~ "Vulva/Vagina",
topography_name == "Skeleton" ~ "Bone",
stringr::str_detect(
tolower(morphology_name),
"myelodysplastic|myeloproliferative") ~ "Myeloid",
stringr::str_detect(
tolower(morphology_name),
"miscellaneous hematopoietic/lymphoid|hodgkin disease") ~ "Lymphoid",
topography_name == "Ureter" ~ "Kidney",
topography_name == "Unknown site" ~ "Any",
topography_name == "Urethra" ~ "Bladder/Urinary Tract",
topography_name == "Spinal cord" ~ "Peripheral Nervous System",
topography_name == "Uterus, cervix" ~ "Cervix",
topography_name == "Large intestine" ~ "Colon/Rectum",
topography_name == "Bladder" ~ "Bladder/Urinary Tract",
topography_name == "Adrenal" ~ "Adrenal Gland",
topography_name == "Intraabdominal" ~ "Esophagus/Stomach",
topography_name == "Oesophagus" ~ "Esophagus/Stomach",
TRUE ~ topography_name
)) |>
dplyr::rename(primary_diagnosis = morphology_name,
primary_site = topography_name2) |>
dplyr::select(variant,
karyotype,
primary_diagnosis,
primary_site,
source_id,
evidence_id) |>
dplyr::distinct() |>
dplyr::mutate(variant_id = paste(
variant, karyotype, sep = " - "
))
evidence_id <- nrow(fusion_event_data) + 1
fusion_event_data$variant_alias <- paste(
fusion_event_data$variant,
paste(
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,2],
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,1],
sep = "-"
),
paste(
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,1],
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,2],
sep = "/"
),
paste(
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,1],
stringr::str_split_fixed(fusion_event_data$variant,"::",2)[,2],
sep = "-"
),
fusion_event_data$karyotype,
sep = "@@@"
)
fusion_event_data$variant_name_primary <- fusion_event_data$variant
bcr_abl_custom <-
data.frame('variant_id' = "BCR::ABL1 - t(9;22)(q34;q11)",
'variant_alias' = 'BCR::ABL',
'karyotype' = 't(9;22)(q34;q11)',
'variant_name_primary' = 'BCR::ABL1',
'evidence_id' = paste0('MITDB_',evidence_id),
'primary_diagnosis' = NA,
'primary_site' = NA,
'source_id' = NA,
'symbol' = 'BCR-ABL1',
stringsAsFactors = F)
mitelman_db <- as.data.frame(
fusion_event_data |>
dplyr::select(variant_alias,
variant_id,
karyotype,
variant_name_primary,
primary_diagnosis,
primary_site,
evidence_id,
source_id) |>
tidyr::separate_rows(variant_alias, sep = "@@@") |>
#dplyr::rename(variant_alias = alias) |>
dplyr::filter(
!stringr::str_detect(
variant_alias,"RARA::PML|ABL1::BCR|ERG::TMPRSS2")) |>
dplyr::bind_rows(
bcr_abl_custom
) |>
dplyr::mutate(alteration_type = "FUSION",
variant_origin = "somatic",
variant_consequence = "transcript_fusion",
evidence_url = "https://mitelmandatabase.isb-cgc.org/",
biomarker_source = "mitelmandb",
#biomarker_entity = T,
#alias_type = "other_gene",
biomarker_source_datestamp = db_datestamp) |>
dplyr::arrange(evidence_id) |>
dplyr::rename(citation_id = source_id) |>
dplyr::select(biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
#alias_type,
variant_name_primary,
alteration_type,
dplyr::everything())
)
mitelman_db$symbol <- stringr::str_split_fixed(
mitelman_db$variant_id, " - ", 2)[,1]
biomarker_items <- list()
biomarker_items[['clinical']] <- mitelman_db
biomarker_items[['variant']] <- biomarker_items[['clinical']] |>
#dplyr::select(variant_id, variant_alias, alteration_type, symbol) |>
dplyr::mutate(alias_type = "other_gene") |>
dplyr::rename(gene = symbol) |>
#dplyr::rename(variant_alias = variant) |>
dplyr::select(biomarker_source,
biomarker_source_datestamp,
variant_id,
variant_alias,
variant_name_primary,
alteration_type,
alias_type,
variant_consequence,
gene) |>
dplyr::distinct()
biomarker_items[['clinical']] <- biomarker_items[['clinical']] |>
dplyr::select(-c("variant_alias", "variant_name_primary",
"alteration_type","variant_consequence",
"symbol")) |>
dplyr::distinct()
biomarker_items <-
get_literature_references(biomarker_items)
return(biomarker_items)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.