fix to work prefixed phasing with <*> allele (#268)

EBIvariation · Dec 18, 2024 · 80d1ad8 · 80d1ad8
1 parent 85a04a0
commit 80d1ad8
Show file tree

Hide file tree

Showing 3 changed files with 38 additions and 1 deletion.
diff --git a/src/vcf/validate_optional_policy.cpp b/src/vcf/validate_optional_policy.cpp
@@ -124,7 +124,14 @@ namespace ebi
     void ValidateOptionalPolicy::check_body_entry_alt_gvcf_gt_value(ParsingState & state, Record const & record) const
     {
         if (util::contains(record.alternate_alleles, GVCF_NON_VARIANT_ALLELE) && record.format[0] == vcf::GT) {
-            for (auto & sample : record.samples) {
+            //with v44, there can be optional leading phasing info, remove it and use
+            bool checkprefix = state.source->version < Version::v44? false : true;
+            for (auto sample : record.samples) {
+                if (checkprefix && !sample.empty()) {
+                    if (sample.at(0) == '/' || sample.at(0) == '|') {
+                        sample.erase(0,1);
+                    }
+                }
                 if (sample_has_reference_in_all_alleles(sample)) {
                     return;
                 }

diff --git a/test/input_files/v4.4/passed/passed_body_alt.vcf b/test/input_files/v4.4/passed/passed_body_alt.vcf
@@ -21,6 +21,7 @@
 1	1900	rs180734498	C	AGT[1:100[	100	PASS	AN=2184;AC=249;AF=0.11;SVLEN=.;SVCLAIM=J	GT:DS:GL	0|0:0.050:-0.13,-0.58,-3.62	0|1:1.000:-2.45,-0.00,-5.00
 1	1900	rs180734498	C	.	100	PASS	AN=2184;AC=249;AF=0.11	GT:DS:GL	0|0:0.050:-0.13,-0.58,-3.62	0|1:1.000:-2.45,-0.00,-5.00
 1	4370	.	G	<*>	.	.	END=4383;SVLEN=.	GT:DP:GQ:MIN:PL	0/0:25:60:23:0,60,900	0/0:25:45:25:0,42,630
+1	4371	.	G	<*>	.	.	END=4383;SVLEN=.	GT:DP:GQ:MIN:PL	/0/0:25:60:23:0,60,900	|0/1:25:45:25:0,42,630
 1	4389	.	T	TC,<*>	213.73	.	.	GT:DP:GQ:PL	0/1:23:99:51,0,36,93,92,86	0/0:24:52:0,52,95,66,95,97
 1	4389	.	C	<*>,<DEL>	.	.	END=4388;SVLEN=.,1;SVCLAIM=.,D	GT:DP:GQ	0/0:25:45	0/1:23:99
 1	4390	.	C	*,<*>	.	.	END=4390	GT:DP:GQ	0/0:26:0	0/0:25:60

diff --git a/test/vcf/optional_policy_test.cpp b/test/vcf/optional_policy_test.cpp
@@ -963,6 +963,20 @@ namespace ebi
                                 { "0/0" },
                                 sources[1]})) );
 
+            CHECK_NOTHROW( (optional_policy.optional_check_body_entry(parsing_state2, vcf::Record{  //gvcf with prefixed phasing
+                                1,
+                                "chr1",
+                                123456,
+                                { "id123" },
+                                "A",
+                                { vcf::GVCF_NON_VARIANT_ALLELE },
+                                1.0,
+                                { vcf::PASS },
+                                { { vcf::END, "0" } },
+                                { vcf::GT },
+                                { "|0/0" },
+                                sources[1]})) );
+
             CHECK_NOTHROW( (optional_policy.optional_check_body_entry(parsing_state2, vcf::Record{
                                 1,
                                 "chr1",
@@ -991,6 +1005,21 @@ namespace ebi
                                 { "0|1" },
                                 sources[1]})),
                             vcf::AlternateAllelesBodyError*);
+
+            CHECK_THROWS_AS( (optional_policy.optional_check_body_entry(parsing_state2, vcf::Record{    //gvcf with prefixed phasing
+                                1,
+                                "chr1",
+                                123456,
+                                { "id123" },
+                                "A",
+                                { vcf::GVCF_NON_VARIANT_ALLELE },
+                                1.0,
+                                { vcf::PASS },
+                                { { vcf::END, "0" } },
+                                { vcf::GT },
+                                { "/0|1" },
+                                sources[1]})),
+                            vcf::AlternateAllelesBodyError*);
         }
     }