Issue 11

q2_intervention/_intervention.py

@@ -13,16 +13,19 @@
                          _between_subject_distance_distribution, _visualize,
                          _get_paired_differences, _stats_and_visuals,
                          _add_metric_to_metadata, _linear_effects,
-                         _regplot_subplots_from_dataframe, _load_metadata)
+                         _regplot_subplots_from_dataframe, _load_metadata,
+                         _check_inputs)
 
 
 def paired_differences(output_dir: str, metadata: qiime2.Metadata,
                        group_column: str, metric: str, state_column: str,
                        state_1: str, state_2: str, individual_id_column: str,
                        parametric: bool=False, palette: str='Set1',
-                       drop_duplicates: bool=True, table: pd.DataFrame=None
+                       drop_replicates: str='error', table: pd.DataFrame=None
                        ) -> None:
 
+    _check_inputs(state_1, state_2)
+
     # find metric in metadata or derive from table and merge into metadata
     metadata = _add_metric_to_metadata(table, metadata, metric)
 
@@ -35,24 +38,26 @@ def paired_differences(output_dir: str, metadata: qiime2.Metadata,
             individual_id_column=individual_id_column,
             group_column=group_column, state_column=state_column,
             state_values=[state_1, state_2],
-            drop_duplicates=drop_duplicates)
+            drop_replicates=drop_replicates)
         pairs[group] = _get_paired_differences(metadata, group_pairs, metric)
 
     # Calculate test statistics and generate boxplots
     _stats_and_visuals(
         output_dir, pairs, metric, group_column, state_column, state_1,
         state_2, individual_id_column, parametric, palette,
-        drop_duplicates, multiple_group_test=True, pairwise_tests=True,
+        drop_replicates, multiple_group_test=True, pairwise_tests=True,
         paired_difference_tests=True, boxplot=True)
 
 
 def pairwise_distance(output_dir: str, distance_matrix: DistanceMatrix,
                       metadata: qiime2.Metadata, group_column: str,
                       state_column: str, state_1: str, state_2: str,
                       individual_id_column: str, parametric: bool=False,
-                      palette: str='Set1', drop_duplicates: bool=True,
+                      palette: str='Set1', drop_replicates: str='error',
                       between_group_distance: bool=False) -> None:
 
+    _check_inputs(state_1, state_2)
+
     metadata = _load_metadata(metadata)
 
     # calculate pairwise distance distributions
@@ -64,7 +69,7 @@ def pairwise_distance(output_dir: str, distance_matrix: DistanceMatrix,
             individual_id_column=individual_id_column,
             group_column=group_column, state_column=state_column,
             state_values=[state_1, state_2],
-            drop_duplicates=drop_duplicates)
+            drop_replicates=drop_replicates)
         pairs[group] = _extract_distance_distribution(
             distance_matrix, group_pairs)
         if between_group_distance:
@@ -76,7 +81,7 @@ def pairwise_distance(output_dir: str, distance_matrix: DistanceMatrix,
     _stats_and_visuals(
         output_dir, pairs, 'distance', group_column,
         state_column, state_1, state_2, individual_id_column,
-        parametric, palette, drop_duplicates, multiple_group_test=True,
+        parametric, palette, drop_replicates, multiple_group_test=True,
         pairwise_tests=True, paired_difference_tests=False, boxplot=True)
 
 

q2_intervention/_utilities.py

@@ -27,28 +27,40 @@
 TEMPLATES = pkg_resources.resource_filename('q2_intervention', 'assets')
 
 
+def _check_inputs(state_1, state_2):
+    if state_1 == state_2:
+        raise ValueError((
+            'You have chosen the same value for state_1 and state_2. These '
+            'parameters must be given different values.'))
+
+
 def _get_group_pairs(df, group_value, individual_id_column='SubjectID',
                      group_column='Group', state_column='time_point',
-                     state_values=['1', '2'], drop_duplicates=True):
+                     state_values=['1', '2'], drop_replicates='error'):
     results = []
     group_members = df[group_column] == group_value
     group_md = df[group_members]
     for individual_id in set(group_md[individual_id_column]):
         result = []
         for state_value in state_values:
             state_value = df[state_column].dtype.type(state_value)
-            individual_id = \
-                df[individual_id_column].dtype.type(individual_id)
+            individual_id = df[individual_id_column].dtype.type(individual_id)
             _state = df[state_column] == state_value
             _ind = df[individual_id_column] == individual_id
             individual_at_state_idx = group_md[_state & _ind].index
             if len(individual_at_state_idx) > 1:
                 print("Multiple values for {0} {1} at {2} {3} ({4})".format(
                     individual_id_column, individual_id, state_column,
                     state_value, ' '.join(map(str, individual_at_state_idx))))
-                if drop_duplicates:
+                if drop_replicates == 'error':
+                    raise ValueError((
+                        'Replicate values for individual {0} at state {1}. '
+                        'Remove replicate values from input files or set '
+                        'drop_replicates parameter to select how replicates '
+                        'are handled.'))
+                elif drop_replicates == 'drop':
                     break
-                else:
+                elif drop_replicates == 'random':
                     individual_at_state_idx = [choice(individual_at_state_idx)]
             elif len(individual_at_state_idx) == 0:
                 print("No values for {0} {1} at {2} {3}".format(
@@ -355,7 +367,7 @@ def _visualize(output_dir, multiple_group_test=False, pairwise_tests=False,
 def _stats_and_visuals(output_dir, pairs, metric, group_column,
                        state_column, state_1, state_2,
                        individual_id_column, parametric, palette,
-                       drop_duplicates,
+                       drop_replicates,
                        multiple_group_test=True, pairwise_tests=True,
                        paired_difference_tests=True, boxplot=True):
     # kruskal test or ANOVA between groups
@@ -381,10 +393,10 @@ def _stats_and_visuals(output_dir, pairs, metric, group_column,
 
     summary = pd.Series(
         [metric, group_column, state_column, state_1, state_2,
-         individual_id_column, parametric, drop_duplicates],
+         individual_id_column, parametric, drop_replicates],
         index=['Metric', 'Group column', 'State column', 'State 1',
                'State 2', 'Individual ID column', 'Parametric',
-               'Drop duplicates'],
+               'Drop replicates'],
         name='Paired difference tests')
 
     _visualize(output_dir, multiple_group_test, pairwise_tests,

q2_intervention/plugin_setup.py

@@ -44,7 +44,8 @@
     'state_1': Str,
     'state_2': Str,
     'parametric': Bool,
-    'drop_duplicates': Bool,
+    'drop_replicates': Str % Choices(
+        ['error', 'random', 'drop', 'mean', 'median']),
 }
 
 base_parameter_descriptions = {
@@ -55,9 +56,9 @@
                          'across which samples are paired.'),
         'individual_id_column': (
             'Metadata column containing subject IDs  to use for pairing '
-            'samples. WARNING: if duplicates exist for an individual ID at '
+            'samples. WARNING: if replicates exist for an individual ID at '
             'either state_1 or state_2, that subject will be dropped and '
-            'reported in standard output by default. Set duplicates="ignore" '
+            'reported in standard output by default. Set replicates="ignore" '
             'to instead randomly select one member, and use --verbose to list '
             'conflicts.'),
         'palette': 'Color palette to use for generating boxplots.',
@@ -72,10 +73,13 @@
         'parametric': ('Perform parametric (ANOVA and t-tests) or non-'
                        'parametric (Kruskal-Wallis, Wilcoxon, and Mann-'
                        'Whitney U tests) statistical tests.'),
-        'drop_duplicates': (
-            'If True, will discard all subject IDs with duplicate samples '
-            'at either state_1 or state_2. If False, will instead '
-            'choose one representative at random from among duplicates.')
+        'drop_replicates': (
+            'Choose how replicate samples are handled. If replicates are '
+            'detected, "error" causes method to fail; "drop" will discard all '
+            'subject IDs with replicate samples at either state_1 or state_2; '
+            '"random" chooses one representative at random from among '
+            'replicates; "mean" and "median" compute average values across '
+            'replicates.')
 }
 
 

q2_intervention/tests/test_intervention.py

@@ -45,15 +45,15 @@ class UtilitiesTests(InterventionTestPluginBase):
     def test_get_group_pairs(self):
         res = _get_group_pairs(
             md, 'a', individual_id_column='ind', group_column='Group',
-            state_column='Time', state_values=[1, 2])
+            state_column='Time', state_values=[1, 2], drop_replicates='drop')
         self.assertEqual(res, [('0', '3'), ('1', '4'), ('2', '5')])
         res = _get_group_pairs(
             md_dup, 'a', individual_id_column='ind', group_column='Group',
-            state_column='Time', state_values=[1, 2])
+            state_column='Time', state_values=[1, 2], drop_replicates='drop')
         self.assertEqual(res, [('0', '3')])
         res = _get_group_pairs(
             md_dup, 'a', individual_id_column='ind', group_column='Group',
-            state_column='Time', state_values=[1, 2], drop_duplicates=False)
+            state_column='Time', state_values=[1, 2], drop_replicates='random')
         self.assertEqual(res[0], ('0', '3'))
         self.assertIn(res[1], [('1', '4'), ('2', '4')])
 
@@ -146,22 +146,23 @@ def test_paired_differences(self):
             output_dir=self.temp_dir.name, table=None,
             metadata=self.md_ecam_fp, group_column='delivery',
             state_column='month', state_1=0, state_2=3,
-            individual_id_column='studyid', metric='observed_otus')
+            individual_id_column='studyid', metric='observed_otus',
+            drop_replicates='drop')
 
     def test_paired_differences_taxa(self):
         paired_differences(
             output_dir=self.temp_dir.name, table=self.table_ecam_fp,
             metadata=self.md_ecam_fp, group_column='delivery',
             state_column='month', state_1=0, state_2=3,
             individual_id_column='studyid',
-            metric='e2c3ff4f647112723741aa72087f1bfa')
+            metric='e2c3ff4f647112723741aa72087f1bfa', drop_replicates='drop')
 
     def test_pairwise_distance(self):
         pairwise_distance(
             output_dir=self.temp_dir.name, distance_matrix=self.md_ecam_dm,
             metadata=self.md_ecam_fp, group_column='delivery',
             state_column='month', state_1=0, state_2=3,
-            individual_id_column='studyid')
+            individual_id_column='studyid', drop_replicates='drop')
 
     def test_linear_mixed_effects(self):
         linear_mixed_effects(