FIX: correctly label 'Unclassfied' MAGs in the respective taxonomy (#183

)
bokulich-lab · Jun 21, 2024 · 3e7b75c · 3e7b75c
1 parent 396c161
commit 3e7b75c
Show file tree

Hide file tree

Showing 3 changed files with 33 additions and 0 deletions.
diff --git a/q2_moshpit/kraken2/select.py b/q2_moshpit/kraken2/select.py
@@ -23,6 +23,12 @@
 RANKS = 'dkpcofgs'
 
 
+def _fill_unclassified(row):
+    if row.isnull().all():
+        row[0] = 'Unclassified'
+    return row
+
+
 def _find_lcas(taxa_list: List[pd.DataFrame], mode: str):
     """Find the least common ancestor in every DataFrame of taxa.
 
@@ -54,6 +60,7 @@ def _find_lcas(taxa_list: List[pd.DataFrame], mode: str):
         results[mag_id] = result
 
     results = pd.DataFrame.from_dict(results, orient='index')
+    results = results.apply(_fill_unclassified, axis=1)
     results = results.apply(lambda x: x.tolist(), axis=1).to_frame()
     results.columns = ['Taxon']
 

diff --git a/q2_moshpit/kraken2/tests/data/mag-taxa-5.csv b/q2_moshpit/kraken2/tests/data/mag-taxa-5.csv
@@ -0,0 +1,5 @@
+Feature ID,Taxon,mag_id
+k141_1566,d__Bacteria;k__Bacteria;p__Actinomycetota;c__Actinomycetes;o__Mycobacteriales;f__Mycobacteriaceae;g__Mycobacterium;s__Mycobacterium florentinum,fed92059-3222-4573-b0ec-726c49fbfabb
+k141_1045,d__Bacteria;k__Bacteria;p__Actinomycetota;c__Actinomycetes;o__Mycobacteriales;f__Mycobacteriaceae;g__Mycobacterium;s__Mycobacterium florentinum,fed92059-3222-4573-b0ec-726c49fbfabb
+k141_1566,d__Bacteria;k__Bacteria;p__Actinomycetota;c__Actinomycetes;o__Mycobacteriales;f__Mycobacteriaceae;g__Mycobacterium;s__Mycobacterium florentinum,fed92059-3222-4573-b0ec-726c49fbfabb
+k141_1045,d__Eukaryota;k__Metazoa;p__Chordata;c__Mammalia;o__Primates;f__Hominidae;g__Homo;s__Homo sapiens,fed92059-3222-4573-b0ec-726c49fbfabb
diff --git a/q2_moshpit/kraken2/tests/test_selection.py b/q2_moshpit/kraken2/tests/test_selection.py
@@ -77,6 +77,9 @@ def setUp(self):
         self.taxa_mag4 = pd.read_csv(
             self.get_data_path('mag-taxa-4.csv'), index_col=0
         )
+        self.taxa_mag5 = pd.read_csv(
+            self.get_data_path('mag-taxa-5.csv'), index_col=0
+        )
 
     def tearDown(self):
         shutil.rmtree(self.temp_dir)
@@ -160,6 +163,24 @@ def test_find_lcas_mode_lca(self):
         exp.index.name = 'Feature ID'
         pandas.testing.assert_frame_equal(obs, exp)
 
+    def test_find_lcas_mode_lca_unclassified(self):
+        taxa = [self.taxa_mag1, self.taxa_mag2, self.taxa_mag5]
+        obs = _find_lcas(taxa, mode='lca')
+        exp = pd.DataFrame.from_dict({
+            '0e514d88-16c4-4273-a1df-1a360eb2c823': [
+                'd__Bacteria;k__Bacteria;p__Actinomycetota;c__Actinomycetes;'
+                'o__Mycobacteriales;f__Mycobacteriaceae;g__Mycobacterium'
+            ],
+            '3acec411-b0d0-4441-b936-5b8b571fa328': [
+                'd__Bacteria;k__Bacteria;p__Actinomycetota;c__Actinomycetes;'
+                'o__Mycobacteriales;f__Mycobacteriaceae;g__Mycobacterium'
+            ],
+            'fed92059-3222-4573-b0ec-726c49fbfabb': ['d__Unclassified']
+        }, orient='index')
+        exp.columns = ['Taxon']
+        exp.index.name = 'Feature ID'
+        pandas.testing.assert_frame_equal(obs, exp)
+
     # def test_find_lcas_mode_majority(self):
     #     taxa = [
     #       self.taxa_mag1, self.taxa_mag2, self.taxa_mag3, self.taxa_mag4