Cleaned up intermediate files in gCNV WDL and fixed miscellaneous typ…

…os. (#5382)
broadinstitute · Mar 4, 2019 · c65b2c2 · c65b2c2
1 parent fd1b1c2
commit c65b2c2
Show file tree

Hide file tree

Showing 5 changed files with 28 additions and 19 deletions.
diff --git a/scripts/cnv_wdl/cnv_common_tasks.wdl b/scripts/cnv_wdl/cnv_common_tasks.wdl
@@ -313,11 +313,12 @@ task ScatterIntervals {
                 --SUBDIVISION_MODE INTERVAL_COUNT \
                 --SCATTER_CONTENT ${num_intervals_per_scatter} \
                 --OUTPUT ${output_dir_} &&
-            # output files are named output_dir_/temp_0001_of_N/scattered.interval_list, etc. (N = num_intervals_per_scatter);
+            # output files are named output_dir_/temp_0001_of_N/scattered.interval_list, etc. (N = number of scatters);
             # we rename them as output_dir_/base_filename.scattered.0000.interval_list, etc.
             ls ${output_dir_}/*/scattered.interval_list | \
                 cat -n | \
                 while read n filename; do mv $filename ${output_dir_}/${base_filename}.scattered.$(printf "%04d" $n).interval_list; done
+            rm -rf ${output_dir_}/temp_*_of_*
         } || {
             # if only a single shard is required, then we can just rename the original interval list
             >&2 echo "IntervalListTools failed because only a single shard is required. Copying original interval list..."
@@ -405,21 +406,22 @@ task PostprocessGermlineCNVCalls {
             model_args="$model_args --model-shard-path MODEL_$index"
         done
 
-        mkdir extracted-contig-ploidy-calls
-        tar xzf ${contig_ploidy_calls_tar} -C extracted-contig-ploidy-calls
+        mkdir contig-ploidy-calls
+        tar xzf ${contig_ploidy_calls_tar} -C contig-ploidy-calls
 
         gatk --java-options "-Xmx${command_mem_mb}m" PostprocessGermlineCNVCalls \
             $calls_args \
             $model_args \
             ${sep=" " allosomal_contigs_args} \
             --autosomal-ref-copy-number ${ref_copy_number_autosomal_contigs} \
-            --contig-ploidy-calls extracted-contig-ploidy-calls \
+            --contig-ploidy-calls contig-ploidy-calls \
             --sample-index ${sample_index} \
             --output-genotyped-intervals ${genotyped_intervals_vcf_filename} \
             --output-genotyped-segments ${genotyped_segments_vcf_filename}
 
-        rm -r CALLS_*
-        rm -r MODEL_*
+        rm -rf CALLS_*
+        rm -rf MODEL_*
+        rm -rf contig-ploidy-calls
     >>>
 
     runtime {

diff --git a/scripts/cnv_wdl/germline/cnv_germline_case_workflow.wdl b/scripts/cnv_wdl/germline/cnv_germline_case_workflow.wdl
@@ -278,19 +278,21 @@ task DetermineGermlineContigPloidyCaseMode {
         export MKL_NUM_THREADS=${default=8 cpu}
         export OMP_NUM_THREADS=${default=8 cpu}
 
-        mkdir input-contig-ploidy-model
-        tar xzf ${contig_ploidy_model_tar} -C input-contig-ploidy-model
+        mkdir contig-ploidy-model
+        tar xzf ${contig_ploidy_model_tar} -C contig-ploidy-model
 
         gatk --java-options "-Xmx${command_mem_mb}m" DetermineGermlineContigPloidy \
             --input ${sep=" --input " read_count_files} \
-            --model input-contig-ploidy-model \
+            --model contig-ploidy-model \
             --output ${output_dir_} \
             --output-prefix case \
             --verbosity DEBUG \
             --mapping-error-rate ${default="0.01" mapping_error_rate} \
             --sample-psi-scale ${default="0.0001" sample_psi_scale}
 
         tar czf case-contig-ploidy-calls.tar.gz -C ${output_dir_}/case-calls .
+
+        rm -rf contig-ploidy-model
     >>>
 
     runtime {
@@ -372,16 +374,16 @@ task GermlineCNVCallerCaseMode {
         export MKL_NUM_THREADS=${default=8 cpu}
         export OMP_NUM_THREADS=${default=8 cpu}
 
-        mkdir contig-ploidy-calls-dir
-        tar xzf ${contig_ploidy_calls_tar} -C contig-ploidy-calls-dir
+        mkdir contig-ploidy-calls
+        tar xzf ${contig_ploidy_calls_tar} -C contig-ploidy-calls
 
         mkdir gcnv-model
         tar xzf ${gcnv_model_tar} -C gcnv-model
 
         gatk --java-options "-Xmx${command_mem_mb}m"  GermlineCNVCaller \
             --run-mode CASE \
             --input ${sep=" --input " read_count_files} \
-            --contig-ploidy-calls contig-ploidy-calls-dir \
+            --contig-ploidy-calls contig-ploidy-calls \
             --model gcnv-model \
             --output ${output_dir_} \
             --output-prefix case \
@@ -425,6 +427,9 @@ task GermlineCNVCallerCaseMode {
             tar czf case-gcnv-calls-shard-${scatter_index}-sample-$CURRENT_SAMPLE_WITH_LEADING_ZEROS.tar.gz -C ${output_dir_}/case-calls/SAMPLE_$CURRENT_SAMPLE .
             let CURRENT_SAMPLE=CURRENT_SAMPLE+1
         done
+
+        rm -rf contig-ploidy-calls
+        rm -rf gcnv-model
     >>>
 
     runtime {

diff --git a/scripts/cnv_wdl/germline/cnv_germline_cohort_workflow.wdl b/scripts/cnv_wdl/germline/cnv_germline_cohort_workflow.wdl
@@ -487,14 +487,14 @@ task GermlineCNVCallerCohortMode {
         export MKL_NUM_THREADS=${default=8 cpu}
         export OMP_NUM_THREADS=${default=8 cpu}
 
-        mkdir contig-ploidy-calls-dir
-        tar xzf ${contig_ploidy_calls_tar} -C contig-ploidy-calls-dir
+        mkdir contig-ploidy-calls
+        tar xzf ${contig_ploidy_calls_tar} -C contig-ploidy-calls
 
         gatk --java-options "-Xmx${command_mem_mb}m"  GermlineCNVCaller \
             --run-mode COHORT \
             -L ${intervals} \
             --input ${sep=" --input " read_count_files} \
-            --contig-ploidy-calls contig-ploidy-calls-dir \
+            --contig-ploidy-calls contig-ploidy-calls \
             ${"--annotated-intervals " + annotated_intervals} \
             --interval-merging-rule OVERLAPPING_ONLY \
             --output ${output_dir_} \
@@ -549,6 +549,8 @@ task GermlineCNVCallerCohortMode {
             tar czf ${cohort_entity_id}-gcnv-calls-shard-${scatter_index}-sample-$CURRENT_SAMPLE_WITH_LEADING_ZEROS.tar.gz -C ${output_dir_}/${cohort_entity_id}-calls/SAMPLE_$CURRENT_SAMPLE .
             let CURRENT_SAMPLE=CURRENT_SAMPLE+1
         done
+
+        rm -rf contig-ploidy-calls
     >>>
 
     runtime {

diff --git a/src/main/java/org/broadinstitute/hellbender/tools/copynumber/CollectAllelicCountsSpark.java b/src/main/java/org/broadinstitute/hellbender/tools/copynumber/CollectAllelicCountsSpark.java
@@ -30,8 +30,8 @@
  * See {@link CollectAllelicCounts}.  This behaves the same, except that it supports spark.
  */
 @CommandLineProgramProperties(
-        summary = "Collects ref/alt counts at sites.",
-        oneLineSummary = "Collects ref/alt counts at sites.",
+        summary = "Collects ref/alt counts at sites",
+        oneLineSummary = "Collects ref/alt counts at sites",
         programGroup = CopyNumberProgramGroup.class
 )
 public class CollectAllelicCountsSpark extends LocusWalkerSpark {

diff --git a/...main/java/org/broadinstitute/hellbender/tools/copynumber/PostprocessGermlineCNVCalls.java b/...main/java/org/broadinstitute/hellbender/tools/copynumber/PostprocessGermlineCNVCalls.java
@@ -91,8 +91,8 @@
  * @author Andrey Smirnov &lt;asmirnov@broadinstitute.org&gt;
  */
 @CommandLineProgramProperties(
-        summary = "Postprocesses the output of GermlineCNVCaller and generates VCF files.",
-        oneLineSummary = "Postprocesses the output of GermlineCNVCaller and generates VCF files.",
+        summary = "Postprocesses the output of GermlineCNVCaller and generates VCF files",
+        oneLineSummary = "Postprocesses the output of GermlineCNVCaller and generates VCF files",
         programGroup = CopyNumberProgramGroup.class
 )
 @DocumentedFeature