ClinVarRecord Fix c.HGVS parsing with new method (if it contains gene…

… symbol)
SACGF · Sep 1, 2023 · 2e56142 · 2e56142
1 parent d6f7f35
commit 2e56142
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 1 deletion.
diff --git a/annotation/clinvar_xml_parser.py b/annotation/clinvar_xml_parser.py
@@ -57,7 +57,7 @@ class ClinVarXmlParserOutput:
 class ClinVarXmlParser(XmlParser, ABC):
 
     RE_DATE_EXTRACTOR = re.compile("([0-9]+-[0-9]+-[0-9]+).*")
-    RE_GOOD_CHGVS = re.compile("^(N._[0-9]+[.][0-9]+:c[.][0-9_a-zA-Z>]+)( .*)?$")
+    RE_GOOD_CHGVS = re.compile("^(N._(.+)?:c[.][0-9_a-zA-Z>]+)( .*)?$")
     RE_ORPHA = re.compile("ORPHA([0-9]+)")
 
     @staticmethod

diff --git a/annotation/clinvar_xml_parser_via_vcv.py b/annotation/clinvar_xml_parser_via_vcv.py
@@ -92,6 +92,17 @@ def parse_c_hgvs(self, elem):
             if hgvs := elem.text:
                 self.latest.c_hgvs = ClinVarXmlParser.parse_hgvs(hgvs)
 
+    @parser_path(
+        "SimpleAllele",
+        "HGVSlist",
+        "HGVS",
+        "Expression")
+    def parse_hgvs_2(self, elem):
+        if not self.latest.c_hgvs:
+            if text := elem.text:
+                if text.startswith("NM_"):
+                    self.latest.c_hgvs = text
+
     @parser_path(
         "SimpleAllele",
         "Location",