Merge pull request #496 from at7/bugfix/resume_after_homref

Bugfix/resume after homref
Ensembl · Jun 17, 2019 · dad1066 · dad1066
2 parents 0f37f44 + 60c752e
commit dad1066
Show file tree

Hide file tree

Showing 6 changed files with 49 additions and 2 deletions.
diff --git a/modules/Bio/EnsEMBL/VEP/Constants.pm b/modules/Bio/EnsEMBL/VEP/Constants.pm
@@ -53,7 +53,7 @@ use warnings;
 use base qw(Exporter);
 
 our $VEP_VERSION     = 96;
-our $VEP_SUB_VERSION = 2;
+our $VEP_SUB_VERSION = 3;
 
 our @EXPORT_OK = qw(
   @FLAG_FIELDS

diff --git a/modules/Bio/EnsEMBL/VEP/Runner.pm b/modules/Bio/EnsEMBL/VEP/Runner.pm
@@ -415,6 +415,8 @@ sub _buffer_to_output {
   else {
     push @output, @{$output_factory->get_all_lines_by_InputBuffer($input_buffer)};
   }
+  # if output is empty here rerun because there could still be VFs left coming from the parser 
+  return $self->_buffer_to_output($input_buffer, $output_as_hash) unless scalar @output  || $self->param('fork');
 
   return \@output;
 }

diff --git a/t/Runner.t b/t/Runner.t
@@ -14,7 +14,6 @@
 
 use strict;
 use warnings;
-
 use Test::More;
 use Test::Exception;
 use FindBin qw($Bin);
@@ -1023,6 +1022,41 @@ STACK S3
   'warnings - with multiple'
 );
 
+# Test to check that after variant is filtered out because all individuals are homozygous for reference
+# vep doesn't stop annotating because of empty output array in Runner.pm
+$runner = Bio::EnsEMBL::VEP::Runner->new({
+  %$cfg_hash,
+  input_data => undef,
+  input_file => $test_cfg->{test_vcf5},
+  dir => $test_cfg->{cache_root_dir},
+  output_file => $test_cfg->{user_file}.'.out',
+  individual => 'all',
+});
+ok($runner->run, 'run - ok');
+open IN, $test_cfg->{user_file}.'.out';
+@tmp_lines = <IN>;
+close IN;
+
+is(scalar (grep {/^var5_chr22/} @tmp_lines), 5, 'run - individual all');
+
+unlink($test_cfg->{user_file}.'.out');
+
+$runner = Bio::EnsEMBL::VEP::Runner->new({
+  %$cfg_hash,
+  input_data => undef,
+  input_file => $test_cfg->{test_vcf6},
+  dir => $test_cfg->{cache_root_dir},
+  output_file => $test_cfg->{user_file}.'.out',
+});
+ok($runner->run, 'run - ok');
+open IN, $test_cfg->{user_file}.'.out';
+@tmp_lines = <IN>;
+close IN;
+
+is(scalar (grep {/^var5_chr22/} @tmp_lines), 5, 'run - ref and alt are the same');
+
+unlink($test_cfg->{user_file}.'.out');
+
 ## run_rest
 ###########
 

diff --git a/t/VEPTestingConfig.pm b/t/VEPTestingConfig.pm
@@ -55,6 +55,8 @@ our %DEFAULTS = (
   test_vcf2      => $Bin.'/testdata/input/test2.vcf',
   test_vcf3      => $Bin.'/testdata/input/test3.vcf',
   test_vcf4      => $Bin.'/testdata/input/test4.vcf',
+  test_vcf5      => $Bin.'/testdata/input/test5.vcf',
+  test_vcf6      => $Bin.'/testdata/input/test6.vcf',
   vr_vcf         => $Bin.'/testdata/input/idt_test.vcf',
   windows_vcf    => $Bin.'/testdata/input/windows.vcf',
   test_gzvcf     => $Bin.'/testdata/input/test.vcf.gz',

diff --git a/t/testdata/input/test5.vcf b/t/testdata/input/test5.vcf
@@ -0,0 +1,5 @@
+##fileformat=VCFv4.1
+##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	P1
+21	25585733	rs142513484	C	T	.	.	.	GT	0/0
+22	25769065	var5_chr22	G	T	.	.	.	GT	0/1
diff --git a/t/testdata/input/test6.vcf b/t/testdata/input/test6.vcf
@@ -0,0 +1,4 @@
+##fileformat=VCFv4.1
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
+21	25585733	rs142513484	C	C	.	.	.
+22	25769065	var5_chr22	G	T	.	.	.