Merge pull request #17 from christopher-mohr/development

Add functionality for parsing raw CNV data

NAMESPACE

@@ -17,6 +17,7 @@ export(add_amplification_data)
 export(add_annotation_data)
 export(add_common_theme_elements)
 export(add_tmb_variant_data)
+export(cnv)
 export(cvo)
 export(filter_consequences)
 export(filter_depth)
@@ -53,6 +54,7 @@ export(process_and_filter_small_variant_data)
 export(qualitymetrics)
 export(read_analysis_status)
 export(read_annotation_data)
+export(read_cnv_data)
 export(read_cvo_data)
 export(read_dna_expanded_metrics)
 export(read_dna_qc_metrics)
@@ -70,7 +72,13 @@ export(read_splice_variants)
 export(read_tmb_details_data)
 export(read_tmb_details_data_csv)
 export(read_tmb_trace_data)
+export(summarize_cnv_data)
 export(tmb)
 export(write_multiqc_data)
 export(write_rdata_file)
 export(write_workbook)
+importFrom(dplyr,left_join)
+importFrom(dplyr,mutate)
+importFrom(stringr,str_split_i)
+importFrom(vcfR,read.vcfR)
+importFrom(vcfR,vcfR2tidy)

R/cnv.R

@@ -0,0 +1,75 @@
+#' Read in a *CopyNumberVariants.vcf file and store as an object
+#'
+#' @description Read in a *CopyNumberVariants.vcf file
+#'
+#' @param cnv_file_path a file path to a *CopyNumberVariants.vcf file
+#'
+#' @return A cnv.output object
+#' 
+#' @export
+cnv <- function(cnv_file_path, local_app=FALSE){
+  new_cnv_output(cnv_file_path)
+}
+
+#' Constructor function for combined.cnv.output objects
+#' Not to be called directly
+#'
+#' @param cnv_file_path a file path to a *CopyNumberVariants.vcf file
+#' @param local_app specifies whether quality metrics are coming from local app
+#'
+#' @return A combined.cnv.output object
+new_cnv_output <- function(cnv_file_path, local_app=FALSE) {
+
+  cnv_data = tibble(file = cnv_file_path) %>%
+    mutate(data = lapply(file, parse_vcf_to_df)) %>%
+    unnest(data) %>%
+    mutate(sample_id = str_replace(basename(file), "_CopyNumberVariants.vcf", "")) %>%
+    select(-file) %>%
+    relocate(sample_id)
+
+  return(structure(cnv_data, class = "combined.cnv.output"))
+}
+
+#' Read in a batch of *CopyNumberVariants.vcf files into a list of CNV objects
+#'
+#' @param cnv_directory a file path to a directory containing one of more *CopyNumberVariants.vcf files
+#' @param local_app specifies whether quality metrics are coming from local app
+#'
+#' @return A named list of combined.cnv.output objects
+#'
+#' @export
+read_cnv_data <- function(cnv_directory, local_app=FALSE){
+  cnv_files <- list.files(
+    path = cnv_directory,
+    pattern = "*CopyNumberVariants.vcf",
+    full.names = TRUE
+  )
+  cnv_data <- map(cnv_files, cnv, local_app)  %>%
+    set_names(str_remove(basename(cnv_files), "\\.vcf$")) 
+  cnv_data
+}
+
+#' Read in a batch of *CopyNumberVariants.vcf files into one dataframe
+#'
+#' @param tmb_directory a file path to a directory containing one of more
+#' *tmb.json files
+#'
+#' @return A dataframe with the read CNV data
+#' 
+#' @export
+summarize_cnv_data <- function(cnv_directory){
+  cnv_files <- list.files(
+    path = cnv_directory,
+    pattern = "*cnv.vcf|*CopyNumberVariants.vcf",
+    full.names = TRUE
+  )
+
+  cnv_data = tibble(file = cnv_files) %>%
+    mutate(data = lapply(file, parse_vcf_to_df)) %>%
+    unnest(data) %>%
+    mutate(sample_id = str_replace(basename(file), "_CopyNumberVariants.vcf", "")) %>%
+    select(-file) %>%
+    relocate(sample_id)
+
+  cnv_data
+}

R/cvo.R

@@ -221,7 +221,8 @@ parse_cvo_record <- function(record_string){
     stringr::str_split("\n") %>%
     unlist() %>%
     stringr::str_remove("\\t$") %>%
-    stringr::str_split("\\t")
+    stringr::str_split("\\t") %>%
+    rapply(., function(x) ifelse(x=="NA",NA,x), how = "replace") # replace all string NAs with NA to avoid warnings from as.numeric
 
   if(stringr::str_detect(record_string, "TMB|MSI")){
     record <- purrr::map(intermediate, ~ as.numeric(.x[2]))

R/wrangle.R

@@ -219,6 +219,34 @@ read_rna_expanded_metrics <- function(qmo_list){
   return(rna_expanded_metrics)
 }
 
+#' Parse VCF files for a provided path and construct data frame.
+#'
+#' @param path path to VCF file in `*.vcf` or `*.vcf.gz` format
+#' @return {tibble} new data frame with all variants (fixed field and genotype information)
+#' @importFrom dplyr mutate left_join
+#' @importFrom vcfR read.vcfR vcfR2tidy
+#' @importFrom stringr str_split_i
+parse_vcf_to_df <- function(path) {
+# parse VCF file
+  vcf_content <- read.vcfR(path)
+
+  # fixed field content to data frame
+  fixed_df <- vcfR2tidy(vcf_content)$fix
+
+  # GT content to data frame
+  gt_df <- vcfR2tidy(vcf_content)$gt
+
+  # create addition column with observed nucleotides in order to avoid collisions when we do the left_join
+  #gt_df <- gt_df %>%
+  #  dplyr::mutate(ALT = str_split_i(gt_GT_alleles, "/", 2))
+
+  # next use ChromKey, POS and ALT for joining vcf content data frames
+  joined_vcf_df <- fixed_df %>%
+    dplyr::left_join(gt_df, by = c("ChromKey", "POS"))
+
+  as_tibble(joined_vcf_df)
+}
+
 #' Process and filter small variant data-frame to requirements
 #'
 #' @description Processes small-variant data to comply with requirements for