🤿towards v0.3.3 (#16)

* 💦opt for pisa

* loose time

* 🚋add export-interaction-mapping command

* pdb_start -> pdb_beg

pdb_profiling/__init__.py

@@ -4,7 +4,7 @@
 # @Author: ZeFeng Zhu
 # @Last Modified: 2020-05-13 08:54:09 pm
 # @Copyright (c) 2020 MinghuiGroup, Soochow University
-__version__ = '0.3.2'
+__version__ = '0.3.3'
 
 
 def default_config(folder='./'):

pdb_profiling/commands/__init__.py

@@ -108,7 +108,7 @@ class ResidueAnnotation(orm.Model):
             chain_id = orm.String(max_length=10)
             resource = orm.String(max_length=100, primary_key=True)
             resource_id = orm.String(max_length=200, primary_key=True)
-            pdb_start = orm.Integer()
+            pdb_beg = orm.Integer()
             pdb_end = orm.Integer()
 
         class UniProtAnnotation(orm.Model):
@@ -118,7 +118,7 @@ class UniProtAnnotation(orm.Model):
             UniProt = orm.String(max_length=50, primary_key=True)
             resource = orm.String(max_length=100, primary_key=True)
             resource_id = orm.String(max_length=200, primary_key=True)
-            unp_start = orm.Integer()
+            unp_beg = orm.Integer()
             unp_end = orm.Integer()
 
         class SMRModel(orm.Model):
@@ -145,7 +145,29 @@ class MappedMutation(orm.Model):
             Ref = orm.String(max_length=3, primary_key=True)
             Pos = orm.Integer(primary_key=True)
             Alt = orm.String(max_length=3, primary_key=True)
-
+
+
+        class PI(orm.Model):
+            __tablename__ = 'PI'
+            __metadata__ = self.metadata
+            __database__ = self.database
+
+            UniProt = orm.String(max_length=50, primary_key=True)
+            pdb_id = orm.String(max_length=4, primary_key=True)
+            entity_id = orm.Integer(primary_key=True)
+            struct_asym_id = orm.String(max_length=10, primary_key=True)
+            chain_id = orm.String(max_length=10)
+            assembly_id = orm.Integer(primary_key=True)
+            model_id = orm.Integer()
+            struct_asym_id_in_assembly = orm.String(max_length=10, primary_key=True)
+            interface_id = orm.Integer(primary_key=True)
+            css = orm.Float()
+            i_select_tag = orm.Boolean()
+            i_select_rank = orm.Integer()
+            pdb_beg = orm.Integer()
+            pdb_end = orm.Integer()
+
+
         self.AAThree2one = AAThree2one
         self.UniProtSeq = UniProtSeq
         self.Mutation = Mutation
@@ -156,3 +178,4 @@ class MappedMutation(orm.Model):
         self.ResidueAnnotation = ResidueAnnotation
         self.SMRModel = SMRModel
         self.MappedMutation = MappedMutation
+        self.PI = PI

pdb_profiling/commands/command.py

@@ -145,6 +145,7 @@ def id_mapping(ctx, input, column, sep, chunksize, auto_assign, sleep):
             if sleep:
                 tsleep(uniform(1, 10))
 
+
 @Interface.command('check-muta-conflict')
 @click.option('--chunksize', type=int, default=100000)
 @click.pass_context
@@ -198,7 +199,7 @@ def get_unp_id(args):
         Entry, isoform, is_canonical = args
         return Entry if is_canonical else isoform
 
-    kwargs = dict(sub.split('=') for item in kwargs for sub in item.split(','))
+    kwargs = dict(sub.split('=') for item in kwargs for sub in item.split(';'))
     if len(kwargs) > 0:
         for key,value in kwargs.items():
             kwargs[key] = eval(value)
@@ -305,7 +306,7 @@ def residue_mapping(ctx, input, chunksize, output, sleep):
             sqlite_api.sync_insert(sqlite_api.ResidueMappingRange, res_mapping_df.to_dict('records'))
         console.log(f'Done: {done}')
         if sleep:
-            tsleep(uniform(0, 3))
+            tsleep(uniform(0, 2))
 
 
 @Interface.command('insert-sele-mapping')
@@ -383,6 +384,52 @@ def expand_iso_range(res):
         console.log(f'Done: {len(res)+chunksize*i}')
 
 
+def pi2records(dfrm: DataFrame, usecols: list, pair_cols: list):
+    yield from yield_interact_records(dfrm[usecols[:13]].rename(columns=dict(zip(usecols[6:13], pair_cols))))
+    yield from yield_interact_records(dfrm[usecols[:6]+usecols[13:]].rename(columns=dict(zip(usecols[13:], pair_cols))))
+
+def yield_interact_records(dfrm: DataFrame):
+    if 'UniProt' in dfrm.columns:
+        for row in dfrm.itertuples(index=False):
+            for beg, end in eval(row.interface_range):
+                yield dict(UniProt=row.UniProt, pdb_id=row.pdb_id, entity_id=row.entity_id, 
+                           struct_asym_id=row.struct_asym_id, chain_id=row.chain_id, 
+                           assembly_id=row.assembly_id, model_id=row.model_id,
+                           struct_asym_id_in_assembly=row.struct_asym_id_in_assembly,
+                           interface_id=row.interface_id, css=row.css,
+                           i_select_tag=row.i_select_tag, i_select_rank=row.i_select_rank,
+                           pdb_beg=beg, pdb_end=end)
+    else:
+        for row in dfrm.itertuples(index=False):
+            for beg, end in eval(row.interface_range):
+                yield dict(UniProt='NaN', pdb_id=row.pdb_id, entity_id=row.entity_id,
+                           struct_asym_id=row.struct_asym_id, chain_id=row.chain_id,
+                           assembly_id=row.assembly_id, model_id=row.model_id,
+                           struct_asym_id_in_assembly=row.struct_asym_id_in_assembly,
+                           interface_id=row.interface_id, css=row.css,
+                           i_select_tag=row.i_select_tag, i_select_rank=row.i_select_rank,
+                           pdb_beg=beg, pdb_end=end)
+
+@Interface.command('insert-interaction')
+@click.option('-i', '--input', type=click.Path())
+@click.option('--chunksize', type=int, help="the chunksize parameter", default=5000)
+@click.option('--ppi/--no-ppi', is_flag=True, default=True)
+@click.pass_context
+def insert_interaction(ctx, input, chunksize, ppi):
+    custom_db = ctx.obj['custom_db']
+    common_cols = ['pdb_id', 'assembly_id', 'interface_id', 'css', 'i_select_tag', 'i_select_rank']
+    pair_cols = ['entity_id', 'struct_asym_id', 'chain_id', 'model_id', 'struct_asym_id_in_assembly', 'interface_range', 'UniProt']
+    usecols = common_cols + [col+'_1' for col in pair_cols] + [col+'_2' for col in pair_cols]
+    df_usecols = usecols if ppi else usecols[:-1]
+    dfs = read_csv(input, sep='\t', keep_default_na=False, na_values=[''], chunksize=chunksize, usecols=df_usecols)
+    done: int = 0
+    with console.status("[bold green]inserting..."):
+        for df in dfs:
+            custom_db.sync_insert(custom_db.PI, list(pi2records(df[df.i_select_rank.ne(-1)], usecols, pair_cols)))
+            done += df.shape[0]
+    console.log(f'Done: {done}')
+
+
 @Interface.command('export-mutation-mapping')
 @click.option('--with_id/--no-with_id', is_flag=True, default=False)
 @click.option('--sele/--no-sele', is_flag=True, default=True)
@@ -446,6 +493,67 @@ def export_residue_remapping(ctx, with_id, sele, output):
     console.log(f'result saved in {output_path}')
 
 
+@Interface.command('export-interaction-mapping')
+@click.option('--with_id/--no-with_id', is_flag=True, default=False)
+@click.option('-o', '--output', type=str, help='filename of output file')
+@click.pass_context
+def export_interface_mapping(ctx, with_id, output):
+    output_path = ctx.obj['folder']/output
+    query = """
+        SELECT DISTINCT 
+                    %s
+                    CASE IDMapping.is_canonical
+                        WHEN 1
+                        THEN IDMapping.Entry
+                        ELSE IDMapping.isoform
+                    END edUniProt, Mutation.Ref, Mutation.Pos, Mutation.Alt,
+                    Mutation.Pos - ResidueMappingRange.unp_beg + ResidueMappingRange.pdb_beg AS residue_number,
+                    Mutation.Pos - ResidueMappingRange.unp_beg + ResidueMappingRange.auth_pdb_beg AS author_residue_number,
+                    ResidueMappingRange.author_insertion_code,
+                    ResidueMappingRange.observed_ratio,
+                    ResidueMappingRange.pdb_id,
+                    ResidueMappingRange.entity_id,
+                    ResidueMappingRange.struct_asym_id,
+                    ResidueMappingRange.chain_id,
+                    PI.assembly_id,
+                    PI.model_id,
+                    PI.struct_asym_id_in_assembly,
+                    PI.interface_id,
+                    PI.css,
+                    PI.i_select_tag,
+                    PI.i_select_rank,
+                    (Mutation.Pos - ResidueMappingRange.unp_beg + ResidueMappingRange.pdb_beg >= PI.pdb_beg AND Mutation.Pos - ResidueMappingRange.unp_beg + ResidueMappingRange.pdb_beg <= PI.pdb_end) AS is_interface_residue
+        FROM Mutation,ResidueMappingRange
+            INNER JOIN IDMapping ON Mutation.ftId = IDMapping.ftId
+            INNER JOIN UniProtSeq ON UniProtSeq.isoform = IDMapping.isoform 
+                                AND UniProtSeq.Pos = Mutation.Pos 
+                                AND UniProtSeq.Ref = Mutation.Ref
+            INNER JOIN SelectedMappingMeta ON SelectedMappingMeta.UniProt = ResidueMappingRange.UniProt
+                                        AND SelectedMappingMeta.pdb_id = ResidueMappingRange.pdb_id
+                                        AND SelectedMappingMeta.struct_asym_id = ResidueMappingRange.struct_asym_id
+            INNER JOIN PI ON PI.UniProt = ResidueMappingRange.UniProt
+                         AND PI.pdb_id = ResidueMappingRange.pdb_id
+                         AND PI.struct_asym_id = ResidueMappingRange.struct_asym_id
+        WHERE ResidueMappingRange.UniProt = edUniProt
+        AND Mutation.Pos >= ResidueMappingRange.unp_beg
+        AND Mutation.Pos <= ResidueMappingRange.unp_end
+        AND ResidueMappingRange.conflict_code IS NULL
+        AND ResidueMappingRange.observed_ratio > 0
+        AND (ResidueMappingRange.residue_name = '' OR ResidueMappingRange.residue_name IN (SELECT three_letter_code FROM AAThree2one))
+        AND SelectedMappingMeta.select_rank != -1
+        ;
+    """
+    query = query % ('Mutation.ftId,' if with_id else '')
+    with console.status("[bold green]query..."):
+        dfs = read_sql_query(query, ctx.obj['custom_db'].engine, chunksize=10000)
+        for df in dfs:
+            if df.shape[0] == 0:
+                continue
+            df.rename(columns={'edUniProt': 'UniProt'}).to_csv(
+                output_path, index=False, mode='a+', sep='\t', header=not output_path.exists())
+    console.log(f'result saved in {output_path}')
+
+
 @Interface.command('insert-sele-mutation-mapping')
 @click.option('-i', '--input', type=click.Path())
 @click.option('--chunksize', type=int, help="the chunksize parameter", default=10000)

pdb_profiling/processors/pdbe/__init__.py

@@ -101,7 +101,7 @@ class profile_id(orm.Model):
 
 
         class PISAInterfaceDict(orm.Model):
-            __tablename__ = 'PISAInterface'
+            __tablename__ = 'PISAInterfaceDict'
             __metadata__ = self.metadata
             __database__ = self.database
             entity_id_1 = orm.Integer()
@@ -127,6 +127,8 @@ class PISAInterfaceDict(orm.Model):
             interface_id = orm.Integer(primary_key=True)
             use_au = orm.Boolean()
             css = orm.Float()
+            is_polymer_1 = orm.Boolean()
+            is_polymer_2 = orm.Boolean()
 
         self.ResidueMapping = ResidueMapping
         self.StatsProteinEntitySeq = StatsProteinEntitySeq

pdb_profiling/processors/pdbe/api.py

@@ -401,10 +401,7 @@ def yieldPISAInterfaceList(data: Dict):
             for record in records:
                 flatten_dict(record, 'structure_1')
                 flatten_dict(record, 'structure_2')
-            flatten_dict(data[pdb], 'page_title', False)
-            cols = sorted(i for i in data[pdb].keys()
-                          if i != 'interfaceentries')
-            yield records, cols, tuple(data[pdb][col] for col in cols)
+            yield records, ('pdb_id', 'assembly_id'), (pdb, data[pdb]['page_title']['assemble_code'])
 
     @staticmethod
     @dispatch_on_set('api/pisa/interfacedetail/')