Fixing uploads with OR instead of BETA

pipeline_steps/common_extraction_functions.R

@@ -333,3 +333,24 @@ split_into_regions <- function(gwas, ld_blocks, study_metadata, p_value_threshol
     return(data.frame())
   }
 }
+
+#' convert_or_to_beta: Given an OR and lower and upper bounds,
+#'   calculates the BETA, and SE.
+#'   based on this answer: https://stats.stackexchange.com/a/327684
+#'
+#' @param gwas: dataframe with the following columns: OR, LB (lower bound), UB (upper bound)
+#' @return gwas with new columns BETA and SE
+#' @import stats
+#' @export
+convert_or_to_beta <- function(gwas) {
+  gwas <- get_file_or_dataframe(gwas)
+  if (!all(c("OR", "OR_LB", "OR_UB") %in% colnames(gwas))) {
+    stop("Need OR, OR_LB + OR_UB to complete conversion")
+  }
+
+  z_score <- stats::qnorm(.975, mean = 0, sd = 1) #1.96
+  gwas$BETA <- log(gwas$OR)
+  gwas$SE <- (log(gwas$OR_LB) - gwas$BETA) / -z_score
+
+  return(gwas)
+}

pipeline_steps/constants.R

@@ -114,6 +114,8 @@ coverage_types <- list(dense = "dense", sparse = "sparse")
 
 standardised_gwas_columns <- c("CHR", "BP", "EA", "OA", "EAF", "BETA", "SE", "P", "SNP", "Z", "GENE")
 required_columns <- c("CHR", "BP", "EA", "OA", "EAF", "BETA", "SE", "P")
+beta_columns <- c("BETA", "SE")
+or_columns <- c("OR", "OR_LB", "OR_UB")
 
 standardised_column_types <- vroom::cols(
   chr = vroom::col_character(),

tests/testing_complete.txt

@@ -1 +1 @@
-SUCCESS: All tests passed on branch: bug/bad-import-compile-results
+SUCCESS: All tests passed on branch: bug/pipeline-gwas-with-or

worker/pipeline_worker.R

@@ -182,6 +182,9 @@ process_message <- function(original_gwas_info, original_payload = NULL) {
         vroom::vroom_write(gwas, gwas_info$metadata$file_location)
         flog.info(paste(gwas_info$metadata$guid, "Added EAF column (NA) for LD panel fill-in during standardisation"))
       }
+      if (all(or_columns %in% colnames(updated_gwas)) && !all(beta_columns %in% colnames(updated_gwas))) {
+        updated_gwas <- convert_or_to_beta(updated_gwas)
+      }
 
       flog.info(paste(gwas_info$metadata$guid, "Extracting regions"))
       extract_regions <- glue::glue(
@@ -294,8 +297,6 @@ verify_gwas_data <- function(gwas_info, gwas) {
 
   gwas <- change_column_names(gwas, gwas_info$metadata$column_names)
   mandatory_columns <- c("CHR", "BP", "P", "EA", "OA")
-  beta_columns <- c("BETA", "SE")
-  or_columns <- c("OR", "OR_LB", "OR_UB")
 
   has_beta <- all(beta_columns %in% colnames(gwas))
   has_or <- all(or_columns %in% colnames(gwas))