Adding tandem duplications as an SV fusion type and automatically fil…

…tering out all events where the genes being fused are the same gene.
KCCG · Mar 9, 2018 · f8c2f40 · f8c2f40
1 parent b7e27c7
commit f8c2f40
Show file tree

Hide file tree

Showing 2 changed files with 23 additions and 5 deletions.
diff --git a/functions/gbs_analysis_functions.php b/functions/gbs_analysis_functions.php
@@ -518,7 +518,7 @@ function analysis_type_svfusions_gbs(array $samples_to_query, array $gene_list_t
 	# WRITE ALL POTENTIALLY FUSING BLOCKS PER SAMPLE OUT TO BED FILES
 	#############################################
 
-	$samples_bed_paths = write_event_types_per_sample_to_beds($samples_to_query, array("inversion", "BND", "deletion"), "split_deletions");
+	$samples_bed_paths = write_event_types_per_sample_to_beds($samples_to_query, array("inversion", "BND"), array("deletion", "tandem duplication"));
 	// Note: this will write block types with 2 breakpoints (e.g. INV/BND) on 2 lines with their separate block IDs but for blocks where it's one event (e.g. deletions) it will write the single block ID followed by -1 and -2 for the 2 breakpoints at the start and end
 
 	if ($samples_bed_paths === false) {
@@ -620,6 +620,24 @@ function analysis_type_svfusions_gbs(array $samples_to_query, array $gene_list_t
 		}
 	}
 
+	// Go through each parsed block id
+	foreach (array_keys($intersection_results) as $block_id) {
+		// If the block id is the first one of one of the ones that were split into their breakpoints
+		if (preg_match("/\-1$/", $block_id)) {
+			$paired_split_block_id = preg_replace("/-1$/", "-2", $block_id);
+
+			// If the paired breakpoint also overlapped with one or more genes
+			if (isset($intersection_results[$paired_split_block_id])) {
+				// If both only overlapped with 1 gene and that gene is the same
+				if (count($intersection_results[$block_id]["gene"]) == 1 && count($intersection_results[$paired_split_block_id]["gene"]) == 1 && $intersection_results[$block_id]["gene"][0] == $intersection_results[$paired_split_block_id]["gene"][0]) {
+					// Delete both
+					unset($intersection_results[$block_id], $intersection_results[$paired_split_block_id]);
+				}
+
+			}
+		}
+	}
+
 	// Array to store block IDs that were successfully intersected with the gene list (i.e. overlap a gene)
 	$intersected_block_ids = array_keys($intersection_results);
 

diff --git a/functions/gbs_query_functions.php b/functions/gbs_query_functions.php
@@ -495,7 +495,7 @@ function write_samples_methods_to_beds(array $sample_names, array $method_names)
 # WRITE EVENT TYPE BLOCKS TO BED FILES PER SAMPLE
 #############################################
 
-function write_event_types_per_sample_to_beds(array $sample_names, array $event_types, $deletions_split_flag = NULL) {
+function write_event_types_per_sample_to_beds(array $sample_names, array $event_types, array $split_event_types) {
 	// At least one sample and event type must be supplied
 	if (count($sample_names) == 0 || count($event_types) == 0) {
 		return false;
@@ -505,7 +505,7 @@ function write_event_types_per_sample_to_beds(array $sample_names, array $event_
 	# FETCH ALL BLOCKS FOR SAMPLES AND EVENT TYPES
 	#############################################
 
-	$GBS_results = fetch_blocks_for_samples_methods_events_gbs($sample_names, array(), $event_types);
+	$GBS_results = fetch_blocks_for_samples_methods_events_gbs($sample_names, array(), array_merge($event_types, $split_event_types));
 
 	if ($GBS_results === false) {
 		return false;
@@ -533,8 +533,8 @@ function write_event_types_per_sample_to_beds(array $sample_names, array $event_
 		foreach (array_keys($GBS_results[$sample_name]) as $method_name) {
 			// Go through each block
 			foreach (array_keys($GBS_results[$sample_name][$method_name]) as $block_id) {
-				// If the flag to split deletions into multiple blocks has been set and the current event is a deletion
-				if (isset($deletions_split_flag) && $deletions_split_flag == "split_deletions" && $GBS_results[$sample_name][$method_name][$block_id]["event_type"] == "deletion") {
+				// If the current event is one of the ones for which to split one block into 2 breakpoints because the block is not stored as 2 breakpoints (e.g. deletions)
+				if (in_array($GBS_results[$sample_name][$method_name][$block_id]["event_type"], $split_event_types)) {
 					// Create 2 lines in the BED file for the sample + split block into start and end and add a -1/-2 to the end of the block id to enable teasing apart what genes they overlapped with later
 					array_push($bed_output, $GBS_results[$sample_name][$method_name][$block_id]["chromosome"]."\t".$GBS_results[$sample_name][$method_name][$block_id]["start"]."\t".($GBS_results[$sample_name][$method_name][$block_id]["start"] + 1)."\t".$block_id."-1\n");
 					array_push($bed_output, $GBS_results[$sample_name][$method_name][$block_id]["chromosome"]."\t".$GBS_results[$sample_name][$method_name][$block_id]["end"]."\t".($GBS_results[$sample_name][$method_name][$block_id]["end"] + 1)."\t".$block_id."-2\n");