Add manually curated tables

Add manually curated tables (curation by @SvetlanaUP & @Vedanth-Ramji) to complete/fix automated outputs.

argnorm/data/manual_curation/abricate_argannot_both_manual_curation.tsv

@@ -0,0 +1,11 @@
+	Original ID	ARO Replacement	Gene Name in CARD
+0	argannot~~~(Bla)PBP1a~~~JN645776:1-2160	3003041	(Bla)PBP1a
+1	argannot~~~(Ntmdz)nimj_Nitroimidazole_Gene~~~KJ816753.1:19295-19792	3007112	(Ntmdz)nimj_Nitroimidazole_Gene
+2	argannot~~~(Phe)cpt_strepv~~~U09991:1412-1948	(Phe)cpt_strepv	(Phe)cpt_strepv
+3	argannot~~~(Bla)Penicillin_Binding_Protein_Ecoli~~~CP002291:664439-666340	3007423	(Bla)Penicillin_Binding_Protein_Ecoli
+4	argannot~~~(AGly)aac3-I~~~AJ877225:5293-5757	3007384	(AGly)aac3-I
+5	argannot~~~(Bla)penA~~~AB511945:1298-3049	(Bla)penA	(Bla)penA
+6	argannot~~~(Phe)catB6~~~AJ223604:3727-4341	3002678	(Phe)catB6
+7	argannot~~~(Tet)tetR~~~HF545434:53576-54226	3003479	(Tet)tetR
+8	argannot~~~(Tet)tetR(G)~~~S52438:113-745	(Tet)tetR(G)	(Tet)tetR(G)
+9	argannot~~~(Bla)PBP1b~~~AF101781:1-2466	(Bla)PBP1b	(Bla)PBP1b

argnorm/data/manual_curation/argsoap_sarg_orfs_manual_curation.tsv

@@ -0,0 +1,15 @@
+	Categories_in_database	ARO Replacement
+0	beta-lactam__fmtC	fmtC
+1	beta-lactam__PBP-1A	PBP-1A
+2	beta-lactam__PBP-1B	PBP-1B
+3	beta-lactam__PBP-2X	PBP-2X
+4	beta-lactam__penA	penA
+5	macrolide-lincosamide-streptogramin__ermC	{'ARO:3000250'}
+6	multidrug__mexT	{'ARO:3000814'}
+7	tetracycline__tetC	{'ARO:3000656', 'ARO:3000167'}
+8	tetracycline__tetD	{'ARO:3000168'}
+9	tetracycline__tetR	{'ARO:0000051'}
+10	tetracycline__tetracycline_resistance_protein	tetracycline_resistance_protein
+11	tetracycline__tetU	{'ARO:3004650'}
+12	unclassified__cob(I)alamin adenolsyltransferase	cob(I)alamin adenolsyltransferase
+13	unclassified__cystathionine beta-lyase patB	cystathionine beta-lyase patB

argnorm/data/manual_curation/resfinder_manual_curation.tsv

@@ -0,0 +1,78 @@
+	Original ID	ARO Replacement	Gene Name in CARD
+0	qnrB11_1_EF653270	3002725	qnrB11
+1	blaOXY-1-5_5_AY077486	blaOXY-1-5	blaOXY-1-5
+2	aph(2'')-If_2_AY701528	3004191	aph(2'')-If
+3	qnrB10_2_HM439644	3002724	qnrB10
+4	blaLEN10_1_AJ635419	blaLEN10	blaLEN10
+5	qnrB6_1_EF520349	3002720	qnrB6
+6	qnrB24_1_HM192542	3002739	qnrB24
+7	blaACT-4_2_AJ311172	blaACT-4	blaACT-4
+8	blaCARB-3_1_S46063	blaCARB-3	blaCARB-3
+9	blaLEN13_1_AJ635403	blaLEN13	blaLEN13
+10	qnrB49_1_JQ582718	3002764	qnrB49
+11	blaLEN9_1_AJ635405	blaLEN9	blaLEN9
+12	blaCMY-59_1_NG_048854	blaCMY-59	blaCMY-59
+13	aac(6')-Ib3_1_X60321	3002576	aac(6')-Ib3
+14	qnrB59_1_JX259320	3002772	qnrB59
+15	qnrB6_2_EF523819	3002720	qnrB6
+16	qnrB3_1_DQ303920	3002716	qnrB3
+17	nimJ_1_NZ_JH815495	3007112	nimJ
+18	qnrB5_1_DQ303919	3002719	qnrB5
+19	qnrB23_1_FJ981622	3002738	qnrB23
+20	qnrB29_1_HM439649	3002744	qnrB29
+21	qnrB13_1_EU273755	3002727	qnrB13
+22	sul2_8_AJ877041	3000412	sul2
+23	blaVMB-1_1_MN719868	blaVMB-1	blaVMB-1
+24	qnrB17_1_JN173051	3002732	qnrB17
+25	fomB_1_AB016934	3000449	fomB
+26	fosA4_1_AB908992	3003210	fosA4
+27	blaB-6_1_AF189302	3005554	blaB-6
+28	mcr-1.26_1_NG_068217	3007280	mcr-1.26
+29	formA_1_X73835	formA	formA
+30	qnrB31_1_HQ418999	3002746	qnrB31
+31	qnrB41_1_JN166690	3002756	qnrB41
+32	qnrB14_1_EU273757	3002728	qnrB14
+33	aac(6')-29a_1_AF263519	3002583	aac(6')-29a
+34	vat(D)_1_L12033	3002843	vat(D)
+35	blaACC-4_1_KM087831	blaACC-4	blaACC-4
+36	qnrB58_1_JX259319	3002771	qnrB58
+37	penA_1_AF515059	penA	penA
+38	qnrB15_1_EU302865	3002730	qnrB15
+39	EstDL136_1_JN242251	EstDL136	EstDL136
+40	blaCTX-M-131_1_JN969893	blaCTX-M-131	blaCTX-M-131
+41	qnrB20_1_AB379831	3002735	qnrB20
+42	qnrB57_1_JX259318	3002770	qnrB57
+43	mcr-3.2_1_NMWW01000143	3004511	mcr-3.2
+44	tet(Q)_4_Z21523	3000191	tet(Q)
+45	qnrB22_1_FJ981621	3002737	qnrB22
+46	blaOXA-85_1_JANA01000064	blaOXA-85	blaOXA-85
+47	qnrB56_1_JX259317	3002769	qnrB56
+48	aac(6')-Ib-cr_2_EF636461	3005113	aac(6')-Ib-cr
+49	qnrB53_1_HQ704413	3002761	qnrB53
+50	ant(3'')-Ia_1_X02340	3000232	ant(3'')-Ia
+51	qnrB30_1_HM439650	3002745	qnrB30
+52	blaCARB-21_1_NG_048724	blaCARB-21	blaCARB-21
+53	blaOXA-289_1_APRM01000005	blaOXA-289	blaOXA-289
+54	blaBRO-1_1_Z54180	blaBRO-1	blaBRO-1
+55	erm(N)_2_MZ015744	erm(N)	erm(N)
+56	qnrB16_1_EU136183	3002731	qnrB16
+57	mcr-3.4_1_FLXA01000011	3004691	mcr-3.4
+58	qnrB26_1_HQ386846	3002741	qnrB26
+59	blaLEN12_1_AJ635406	blaLEN12	blaLEN12
+60	qnrB12_2_AM774474	3002726	qnrB12
+61	aac(3)-Ia_1_X15852	3002528	aac(3)-Ia
+62	blaCARB-23_1_NG_048726	blaCARB-23	blaCARB-23
+63	aac(3)-I_1_AJ877225	3007384	aac(3)-I
+64	mcr-3.5_1_MF489760	3004505	mcr-3.5
+65	qnrB62_1_JX987101	3002775	qnrB62
+66	mcr-3.3_1_MF495680	3004662	mcr-3.3
+67	blaACC-1_2_HG530658	blaACC-1	blaACC-1
+68	aac(6')-29b_1_AF263519	3002584	aac(6')-29b
+69	sul2_9_FJ197818	3000412	sul2
+70	qnrB34_1_JN173056	3002749	qnrB34
+71	blaLEN7_1_AJ635425	blaLEN7	blaLEN7
+72	aadA9_1_AJ420072	3002609	aadA9
+73	qnrB10_1_DQ631414	3002724	qnrB10
+74	qnrB18_1_AM919399	3002733	qnrB18
+75	blaLEN8_1_AJ635424	blaLEN8	blaLEN8
+76	qnrB17_2_AM919398	3002732	qnrB17

argnorm/drug_categorization.py

@@ -4,7 +4,7 @@
 import pronto
 from typing import List, Tuple
 
-# Load the ArgNorm ontology from the 'aro.obo' file
+# Load the ArgNorm ontology from internet
 ARO = pronto.Ontology.from_obo_library('aro.obo')
 
 def get_immediate_drug_classes(aro_num: str) -> List[Tuple]:

argnorm/normalizers.py

@@ -16,21 +16,40 @@
 
 _ROOT = os.path.abspath(os.path.dirname(__file__))
 
+def is_number(num):
+    """
+    Required for checking aro mappings to discern between numbers and other
+    string identifiers.
+    """
+    try:
+        float(num)
+    except ValueError:
+        return False
 
-def get_data_path(path):
-    return os.path.join(_ROOT, 'data', path)
+    return True
 
+def get_data_path(path, getting_manual_curation):
+    """
+    Gets mapping tables and manual curation tables.
+    Giving 'True' as argument after 'path' will get manual curation table.
+    Else mapping table will be returned.
+    """
+    if getting_manual_curation:
+        return os.path.join(_ROOT, 'data/manual_curation', path)
+
+    return os.path.join(_ROOT, 'data', path)
 
 class BaseNormalizer:
     """
     Inherit this class and customize subclass methods to implement the normalization of tools.
     """
 
-    def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
+    def __init__(self, database=None, is_hamronized=False, mode=None, uses_manual_curation=True) -> None:
         self.tool = ''
         self.database = database
         self.mode = mode
         self.is_hamronized = is_hamronized
+        self.uses_manual_curation = uses_manual_curation
         self._set_input_gene_col()
         self._set_ref_gene_and_aro_cols()
 
@@ -103,9 +122,34 @@ def get_aro_mapping_table(self):
         """
         Don't customize this unless you're using your own (not package built-in) reference data.
         """
-        df = pd.read_csv(get_data_path(f'{self.tool}_{self.database}_{self.mode}_ARO_mapping.tsv'), sep='\t', index_col=0)
-        if self.tool != 'argsoap' or self.mode != 'orfs':
-            df[TARGET_ARO_COL] = df[TARGET_ARO_COL].map(lambda a: f'ARO:{int(a) if a == a else "nan"}') # a == a checks that a is not nan
+        df = pd.read_csv(get_data_path(f'{self.tool}_{self.database}_{self.mode}_ARO_mapping.tsv', False), sep='\t', index_col=0)
+
+        if self.uses_manual_curation:
+            if self.database == 'sarg' and self.mode == 'orfs':
+                gene_identifier = 'Categories_in_database'
+            else:
+                gene_identifier = 'Original ID'
+
+            if self.database == 'ncbi':
+                manual_curation = pd.read_csv(get_data_path('ncbi_manual_curation.tsv', True), sep='\t')
+            elif self.database == 'resfinder':
+                manual_curation = pd.read_csv(get_data_path('resfinder_manual_curation.tsv', True), sep='\t')
+            else:
+                manual_curation = pd.read_csv(get_data_path(f'{self.tool}_{self.database}_{self.mode}_manual_curation.tsv', True), sep='\t')
+
+            aro_nan_indices = [(list(df[gene_identifier]).index(manual_curation.loc[i, gene_identifier])) for i in range(manual_curation.shape[0])]
+
+            for i in range(len(aro_nan_indices)):
+                df.loc[aro_nan_indices[i], 'ARO'] = manual_curation.loc[i, 'ARO Replacement']
+
+                if self.tool != 'argsoap' and self.mode != 'orfs':
+                    df.loc[aro_nan_indices[i], 'Gene Name in CARD'] = manual_curation.loc[i, 'Gene Name in CARD']
+            if self.tool != 'argsoap' or self.mode != 'orfs':
+                df[TARGET_ARO_COL] = df[TARGET_ARO_COL].map(lambda a: f'ARO:{int(float(a)) if is_number(a) == True else a}')
+        else:
+            if self.tool != 'argsoap' or self.mode != 'orfs':
+                df[TARGET_ARO_COL] = df[TARGET_ARO_COL].map(lambda a: f'ARO:{int(a) if a == a else "nan"}') # a == a checks that a is not nan
+
         return df
 
     def load_input(self, input_file):
@@ -116,14 +160,14 @@ def load_input(self, input_file):
 
 
 class ARGSOAPNormalizer(BaseNormalizer):
-    def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
+    def __init__(self, database=None, is_hamronized=False, mode=None, uses_manual_curation=True) -> None:
         if not database:
             warnings.warn('No `database` specified. Will try using SARG.')
             database = 'sarg'
         elif database != 'sarg':
             warnings.warn('The `database` is not supported. Will try using SARG instead.')
             database = 'sarg'
-        super().__init__(database, is_hamronized, mode)
+        super().__init__(database, is_hamronized, mode, uses_manual_curation)
         self.tool = 'argsoap'
 
     def _set_ref_gene_and_aro_cols(self):
@@ -182,7 +226,7 @@ def _raise_incorrect_mode_error(self):
 
 class DeepARGNormalizer(BaseNormalizer):
 
-    def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
+    def __init__(self, database=None, is_hamronized=False, mode=None, uses_manual_curation=True) -> None:
         if mode:
             warnings.warn('`mode` is not relavant for DeepARG and will be ignored.')
             mode = 'both'
@@ -195,7 +239,7 @@ def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
         elif database != 'deeparg':
             warnings.warn('The `database` is not supported. Will try using DeepARG instead.')
             database = 'deeparg'
-        super().__init__(database, is_hamronized, mode)
+        super().__init__(database, is_hamronized, mode, uses_manual_curation)
         self.tool = 'deeparg'
 
     def _set_input_gene_col(self):
@@ -210,7 +254,7 @@ def _set_input_gene_col(self):
 
 class ResFinderNormalizer(BaseNormalizer):
 
-    def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
+    def __init__(self, database=None, is_hamronized=False, mode=None, uses_manual_curation=True) -> None:
         if mode:
             warnings.warn('`mode` is not relavant for ResFinder and will be ignored.')
             mode = 'both'
@@ -223,7 +267,7 @@ def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
         elif database != 'resfinder':
             warnings.warn('The `database` is not supported. Will try using ResFinder instead.')
             database = 'resfinder'
-        super().__init__(database, is_hamronized, mode)
+        super().__init__(database, is_hamronized, mode, uses_manual_curation)
         self.tool = 'resfinder'
 
     def _set_input_gene_col(self):
@@ -241,7 +285,7 @@ def preprocess_ref_genes(self, ref_genes):
 
 class AMRFinderPlusNormalizer(BaseNormalizer):
 
-    def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
+    def __init__(self, database=None, is_hamronized=False, mode=None, uses_manual_curation=True) -> None:
         if mode:
             warnings.warn('`mode` is not relavant for AMRFinderPlus and will be ignored.')
             mode = 'both'
@@ -254,7 +298,7 @@ def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
         elif database != 'ncbi':
             warnings.warn('The `database` is not supported. Will try using NCBI instead.')
             database = 'ncbi'
-        super().__init__(database, is_hamronized, mode)
+        super().__init__(database, is_hamronized, mode, uses_manual_curation)
         self.tool = 'amrfinderplus'
 
     def _set_input_gene_col(self):
@@ -272,14 +316,14 @@ def preprocess_ref_genes(self, ref_genes):
 
 class AbricateNormalizer(BaseNormalizer):
 
-    def __init__(self, database=None, is_hamronized=False, mode=None) -> None:
+    def __init__(self, database=None, is_hamronized=False, mode=None, uses_manual_curation=True) -> None:
         if mode:
             warnings.warn('`mode` is not relavant for Abricate and will be ignored.')
             mode = 'both'
         else:
             warnings.warn('`mode` is not specified. Will use default setting "both".')
             mode = 'both'
-        super().__init__(database, is_hamronized, mode)
+        super().__init__(database, is_hamronized, mode, uses_manual_curation)
         self.tool = 'abricate'
 
     def _set_input_gene_col(self):