Convert protonated DC/C residue to valid DNR/CH, respectively (NEFTra…

…nslator v3.6.3 relating DAOTHER-9198)
wwPDB · Jul 1, 2024 · 0c6aa6b · 0c6aa6b
1 parent d3c7de7
commit 0c6aa6b
Show file tree

Hide file tree

Showing 2 changed files with 39 additions and 2 deletions.
diff --git a/wwpdb/utils/nmr/BMRBChemShiftStat.py b/wwpdb/utils/nmr/BMRBChemShiftStat.py
@@ -818,7 +818,7 @@ def loadStatFromCsvFile(self, file_name, primary_th, secondary_th=None, comp_id_
                     _atom_id = 'MG2'
 
                 elif _atom_id.endswith('"'):
-                    _atom_id = _atom_id[:-1] + "'"
+                    _atom_id = _atom_id[:-1] + "''"
 
                 # methyl proton group
                 if _atom_id.startswith('M'):
@@ -1712,6 +1712,9 @@ def checkAtomNomenclature(self, atom_id, comp_id=None, verbose=False):
         if comp_id == 'DC' and atom_id == 'H3':  # DAOTHER-9198
             return True, 'DNR', 'HN3'
 
+        if comp_id == 'C' and atom_id == 'H3':  # DAOTHER-9198 (RNA linking)
+            return True, 'CH', 'HN3'
+
         if verbose:
             self.__lfh.write(f"+BMRBChemShiftStat.checkAtomNomenclature() ++ Warning  - {comp_id}:{atom_id} did not match with any atom in CCD, {_ref_atom_ids}\n")
 

diff --git a/wwpdb/utils/nmr/NEFTranslator/NEFTranslator.py b/wwpdb/utils/nmr/NEFTranslator/NEFTranslator.py
@@ -96,6 +96,7 @@
 # 11-Jun-2024  M. Yokochi - add support for ligand remapping in annotation process (v3.6.0, DAOTHER-9286)
 # 28-Jun-2024  M. Yokochi - do not evaluate value in case of ValueError exception (v3.6.1, DAOTHER-9520)
 # 28-Jun-2024  M. Yokochi - throw error for key items with empty value (v3.6.2, DAOTHER-9520, 2nd case)
+# 01-Jul-2024  M. Yokochi - convert protonated DC/C to DNR/CH (v3.6.3)
 ##
 """ Bi-directional translator between NEF and NMR-STAR
     @author: Kumaran Baskaran, Masashi Yokochi
@@ -148,7 +149,7 @@
 
 
 __package_name__ = 'wwpdb.utils.nmr'
-__version__ = '3.6.2'
+__version__ = '3.6.3'
 
 __pynmrstar_v3_3_1__ = version.parse(pynmrstar.__version__) >= version.parse("3.3.1")
 __pynmrstar_v3_2__ = version.parse(pynmrstar.__version__) >= version.parse("3.2.0")
@@ -1523,6 +1524,39 @@ def skip_empty_value_error(lp, idx):
                             alt_chain_id_set.add(row)
                 has_auth_asym_id = len(alt_chain_id_set) > 0
 
+                # convert protonated DC -> DNR, protonated C -> CH
+                if 'Atom_ID' in loop.tags and 'Auth_comp_ID' not in loop.tags\
+                   and set(tags) & set(loop.tags) == set(tags):
+                    pre_tag = copy.deepcopy(tags)
+                    pre_tag.append('Atom_ID')
+                    pre_seq_data = get_lp_tag(loop, pre_tag)
+                    has_dc_h3 = has_c_h3 = False
+                    dnr_set = set()
+                    ch_set = set()
+                    for row in pre_seq_data:
+                        if row[1] == 'DC' and row[4] in ('H3', 'H3+'):
+                            has_dc_h3 = True
+                            seq_key = (row[0], row[3])
+                            dnr_set.add(seq_key)
+                        if row[1] == 'C' and row[4] in ('H3', 'H3+'):
+                            has_c_h3 = True
+                            seq_key = (row[0], row[3])
+                            ch_set.add(seq_key)
+                    if has_dc_h3:
+                        comp_id_col = loop.tags.index('Comp_ID')
+                        for idx, row in enumerate(pre_seq_data):
+                            if row[1] == 'DC':
+                                seq_key = (row[0], row[3])
+                                if seq_key in dnr_set:
+                                    loop.data[idx][comp_id_col] = 'DNR'
+                    if has_c_h3:
+                        comp_id_col = loop.tags.index('Comp_ID')
+                        for idx, row in enumerate(pre_seq_data):
+                            if row[1] == 'C':
+                                seq_key = (row[0], row[3])
+                                if seq_key in ch_set:
+                                    loop.data[idx][comp_id_col] = 'CH'
+
             if lp_category == '_Atom_chem_shift' and self.__remediation_mode and has_auth_asym_id\
                and set(tags) & set(loop.tags) == set(tags) and set(tags__) & set(loop.tags) == set(tags__):
                 factor = max(len(alt_chain_id_set), 2)  # 2lnh