Merge pull request #127 from MannLabs/multicova

Release 0.4.3

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.4.2
+current_version = 0.4.3
 commit = True
 tag = False
 parse = (?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\-(?P<release>[a-z]+)(?P<build>\d+))?

HISTORY.md

@@ -1,5 +1,9 @@
 # Changelog
 
+# 0.4.3
+* FIX loading dataset with columns
+* ADD log2-transformation
+
 # 0.4.2
 * ADD option compare_preprocessing_modes
 * update to streamlit 1.19 with new caching functions

alphastats/DataSet.py

@@ -115,7 +115,7 @@ def create_matrix(self):
         """
 
         regex_find_intensity_columns = self.intensity_column.replace("[sample]", ".*")
-
+        
         df = self.rawinput
         df = df.set_index(self.index_column)
         df = df.filter(regex=(regex_find_intensity_columns), axis=1)
@@ -159,6 +159,7 @@ def load_metadata(self, file_path):
 
         # check whether sample labeling matches protein data
         #  warnings.warn("WARNING: Sample names do not match sample labelling in protein data")
+        df.columns = df.columns.astype(str)
         self.metadata = df
 
     def _save_dataset_info(self):
@@ -168,6 +169,7 @@ def _save_dataset_info(self):
             "Matrix: Number of ProteinIDs/ProteinGroups": self.mat.shape[1],
             "Matrix: Number of samples": self.mat.shape[0],
             "Intensity used for analysis": self.intensity_column,
+            "Log2-transformed": False,
             "Normalization": None,
             "Imputation": None,
             "Contaminations have been removed": False,

alphastats/DataSet_Preprocess.py

@@ -129,7 +129,7 @@ def _imputation(self, method):
     @ignore_warning(UserWarning)
     @ignore_warning(RuntimeWarning)
     def _normalization(self, method):
-
+    
         if method == "zscore":
             scaler = sklearn.preprocessing.StandardScaler()
             normalized_array = scaler.fit_transform(self.mat.values)
@@ -153,11 +153,10 @@ def _normalization(self, method):
                 "Choose from 'zscore', 'quantile', 'linear' normalization. or 'vst' for variance stabilization transformation"
             )
 
-        # TODO logarithimic normalization
-
         self.mat = pd.DataFrame(
             normalized_array, index=self.mat.index, columns=self.mat.columns
         )
+
         self.preprocessing_info.update({"Normalization": method})
 
     def reset_preprocessing(self):
@@ -193,11 +192,16 @@ def _compare_preprocessing_modes(self, func, params_for_func):
             results_list.append(res)
 
         return results_list
-
+
+    def _log2_transform(self):
+        self.mat = np.log2(self.mat + 0.1)
+        self.preprocessing_info.update({"Log2 Transformed": True})
+
 
     @ignore_warning(RuntimeWarning)
     def preprocess(
         self,
+        log2_transform=True,
         remove_contaminations=False,
         subset=False,
         normalization=None,
@@ -239,6 +243,7 @@ def preprocess(
 
         Args:
             remove_contaminations (bool, optional): remove ProteinGroups that are identified as contamination.
+            log2_transform (bool, optional): Log2 transform data. Default to True.
             normalization (str, optional): method to normalize data: either "zscore", "quantile", "linear". Defaults to None.
             remove_samples (list, optional): list with sample ids to remove. Defaults to None.
             imputation (str, optional):  method to impute data: either "mean", "median", "knn" or "randomforest". Defaults to None.
@@ -249,6 +254,9 @@ def preprocess(
 
         if subset:
             self.mat = self._subset()
+
+        if log2_transform:
+            self._log2_transform()
 
         if normalization is not None:
             self._normalization(method=normalization)

alphastats/__init__.py

@@ -1,5 +1,5 @@
 __project__ = "alphastats"
-__version__ = "0.4.2"
+__version__ = "0.4.3"
 __license__ = "Apache"
 __description__ = "An open-source Python package for Mass Spectrometry Analysis"
 __author__ = "Mann Labs"

alphastats/gui/pages/03_Preprocessing.py

@@ -29,6 +29,11 @@ def preprocessing():
                 options=[True, False],
             )
 
+            log2_transform = st.selectbox(
+                "Log2-transform dataset",
+                options=[True, False],
+            )
+
             normalization = st.selectbox(
                 "Normalization", options=[None, "zscore", "quantile", "vst", "linear"]
             )
@@ -42,6 +47,7 @@ def preprocessing():
         if submitted:
             st.session_state.dataset.preprocess(
                 remove_contaminations=remove_contaminations,
+                log2_transform=log2_transform,
                 subset=subset,
                 normalization=normalization,
                 imputation=imputation,

alphastats/loader/AlphaPeptLoader.py

@@ -41,6 +41,7 @@ def __init__(
         # add contamination column "Reverse"
         self._add_contamination_reverse_column()
         self._add_contamination_column()
+        self._read_all_columns_as_string()
         #  make ProteinGroup column
         self.rawinput["ProteinGroup"] = self.rawinput[self.index_column].map(
             self._standardize_protein_group_column

alphastats/loader/BaseLoader.py

@@ -32,6 +32,7 @@ def __init__(self, file, intensity_column, index_column, sep):
         self.ptm_df = None
         self._add_contamination_column()
         self._check_if_columns_are_present()
+        self._read_all_columns_as_string()
 
     def _check_if_columns_are_present(self):
         """check if given columns present in rawinput"""
@@ -46,9 +47,11 @@ def _check_if_columns_are_present(self):
                 "FragPipe Format: https://fragpipe.nesvilab.org/docs/tutorial_fragpipe_outputs.html#combined_proteintsv"
                 "MaxQuant Format: http://www.coxdocs.org/doku.php?id=maxquant:table:proteingrouptable"
             )
+
+    def _read_all_columns_as_string(self):
+        self.rawinput.columns = self.rawinput.columns.astype(str)
 
     def _check_if_indexcolumn_is_unique(self):
-        # TODO make own duplicates functions to have less dependencies
         duplicated_values = list(duplicates(self.rawinput[self.index_column].to_list()))
         if len(duplicated_values) > 0:
             # error or warning, duplicates could be resolved with preprocessing/filtering

alphastats/loader/DIANNLoader.py

@@ -42,6 +42,7 @@ def __init__(
         self._remove_filepath_from_name()
         self._add_tag_to_sample_columns()
         self._add_contamination_column()
+        self._read_all_columns_as_string()
 
     def _add_tag_to_sample_columns(self):
         """

alphastats/loader/MaxQuantLoader.py

@@ -34,6 +34,7 @@ def __init__(
         self.confidence_column = confidence_column
         self.software = "MaxQuant"
         self._set_filter_columns_to_true_false()
+        self._read_all_columns_as_string()
 
         if gene_names_column in self.rawinput.columns.to_list():
             self.gene_names = gene_names_column

alphastats/loader/SpectronautLoader.py

@@ -48,6 +48,7 @@ def __init__(
 
         self._reshape_spectronaut(sample_column=sample_column, gene_names_column=gene_names_column)
         self._add_contamination_column()
+        self._read_all_columns_as_string()
 
 
     def _reshape_spectronaut(self, sample_column, gene_names_column):

docs/conf.py

@@ -23,7 +23,7 @@
 author = "Elena Krismer"
 
 # The full version, including alpha/beta/rc tags
-release = "0.4.2"
+release = "0.4.3"
 
 
 # -- General configuration ---------------------------------------------------

release/one_click_linux_gui/control

@@ -1,5 +1,5 @@
 Package: alphastats
-Version: 0.4.2
+Version: 0.4.3
 Architecture: all
 Maintainer: MannLabs
 Description: alphastats

release/one_click_linux_gui/create_installer_linux.sh

@@ -17,7 +17,7 @@ python setup.py sdist bdist_wheel
 # Setting up the local package
 cd release/one_click_linux_gui
 # Make sure you include the required extra packages and always use the stable or very-stable options!
-pip install "../../dist/alphastats-0.4.2-py3-none-any.whl"
+pip install "../../dist/alphastats-0.4.3-py3-none-any.whl"
 
 # Creating the stand-alone pyinstaller folder
 pip install pyinstaller==5.8

release/one_click_macos_gui/Info.plist

@@ -9,9 +9,9 @@
 	<key>CFBundleIconFile</key>
 	<string>alphapeptstats_logo.icns</string>
 	<key>CFBundleIdentifier</key>
-	<string>alphastats.0.4.2</string>
+	<string>alphastats.0.4.3</string>
 	<key>CFBundleShortVersionString</key>
-	<string>0.4.2</string>
+	<string>0.4.3</string>
 	<key>CFBundleInfoDictionaryVersion</key>
 	<string>6.0</string>
 	<key>CFBundleName</key>

release/one_click_macos_gui/create_installer_macos.sh

@@ -20,7 +20,7 @@ python setup.py sdist bdist_wheel
 
 # Setting up the local package
 cd release/one_click_macos_gui
-pip install "../../dist/alphastats-0.4.2-py3-none-any.whl"
+pip install "../../dist/alphastats-0.4.3-py3-none-any.whl"
 
 # Creating the stand-alone pyinstaller folder
 pip install pyinstaller==5.8

release/one_click_macos_gui/distribution.xml

@@ -1,6 +1,6 @@
 <?xml version="1.0" encoding="utf-8" standalone="no"?>
 <installer-script minSpecVersion="1.000000">
-    <title>AlphaPeptStats 0.4.0</title>
+    <title>AlphaPeptStats 0.4.3</title>
     <background mime-type="image/png" file="alphapeptstats_logo.png" scaling="proportional"/>
     <welcome file="welcome.html" mime-type="text/html" />
     <conclusion file="conclusion.html" mime-type="text/html" />

release/one_click_windows_gui/alphastats_innoinstaller.iss

@@ -2,7 +2,7 @@
 ; SEE THE DOCUMENTATION FOR DETAILS ON CREATING INNO SETUP SCRIPT FILES!
 
 #define MyAppName "AlphaPeptStats"
-#define MyAppVersion "0.4.2"
+#define MyAppVersion "0.4.3"
 #define MyAppPublisher "MannLabs"
 #define MyAppURL "https://github.com/MannLabs/alphapeptstats"
 #define MyAppExeName "alphastats_gui.exe"

release/one_click_windows_gui/create_installer_windows.sh

@@ -17,7 +17,7 @@ python setup.py sdist bdist_wheel
 # Setting up the local package
 cd release/one_click_windows_gui
 # Make sure you include the required extra packages and always use the stable or very-stable options!
-pip install "../../dist/alphastats-0.4.2-py3-none-any.whl"
+pip install "../../dist/alphastats-0.4.3-py3-none-any.whl"
 
 # Creating the stand-alone pyinstaller folder
 pip install pyinstaller==5.8

setup.py

@@ -25,7 +25,7 @@ def create_pip_wheel():
     requirements = get_requirements()
     setuptools.setup(
         name="alphastats",
-        version="0.4.2",
+        version="0.4.3",
         license="Apache",
         description="An open-source Python package for Mass Spectrometry Analysis",
         long_description=get_long_description(),

tests/test_DataSet.py

@@ -257,7 +257,7 @@ def test_preprocess_remove_samples(self):
     def test_preprocess_normalize_zscore(self):
         self.obj.mat = pd.DataFrame({"a": [2, 5, 4], "b": [5, 4, 4], "c": [0, 10, 8]})
         # zscore Normalization
-        self.obj.preprocess(normalization="zscore")
+        self.obj.preprocess(log2_transform=False,normalization="zscore")
         expected_mat = pd.DataFrame(
             {
                 "a": [-1.33630621, 1.06904497, 0.26726124],
@@ -270,7 +270,7 @@ def test_preprocess_normalize_zscore(self):
     def test_preprocess_normalize_quantile(self):
         self.obj.mat = pd.DataFrame({"a": [2, 5, 4], "b": [5, 4, 4], "c": [0, 10, 8]})
         # Quantile Normalization
-        self.obj.preprocess(normalization="quantile")
+        self.obj.preprocess(log2_transform=False,normalization="quantile")
         expected_mat = pd.DataFrame(
             {"a": [0.0, 1.0, 0.5], "b": [1.0, 0.0, 0.0], "c": [0.0, 1.0, 0.5]}
         )
@@ -279,7 +279,7 @@ def test_preprocess_normalize_quantile(self):
     def test_preprocess_normalize_linear(self):
         self.obj.mat = pd.DataFrame({"a": [2, 5, 4], "b": [5, 4, 4], "c": [0, 10, 8]})
         # Linear Normalization
-        self.obj.preprocess(normalization="linear")
+        self.obj.preprocess(log2_transform=False,normalization="linear")
         expected_mat = pd.DataFrame(
             {
                 "a": [0.37139068, 0.42107596, 0.40824829],
@@ -292,7 +292,7 @@ def test_preprocess_normalize_linear(self):
     def test_preprocess_normalize_vst(self):
         self.obj.mat = pd.DataFrame({"a": [2, 5, 4], "b": [5, 4, 4], "c": [0, 10, 8]})
         # Linear Normalization
-        self.obj.preprocess(normalization="vst")
+        self.obj.preprocess(log2_transform=False,normalization="vst")
         expected_mat = pd.DataFrame(
             {
                 "a": [-1.30773413, 1.12010046, 0.18763367],
@@ -306,7 +306,7 @@ def test_preprocess_imputation_mean_values(self):
         self.obj.mat = pd.DataFrame(
             {"a": [2, np.nan, 4], "b": [5, 4, 4], "c": [np.nan, 10, np.nan]}
         )
-        self.obj.preprocess(imputation="mean")
+        self.obj.preprocess(log2_transform=False,imputation="mean")
         expected_mat = pd.DataFrame(
             {"a": [2.0, 3.0, 4.0], "b": [5.0, 4.0, 4.0], "c": [10.0, 10.0, 10.0]}
         )
@@ -316,7 +316,7 @@ def test_preprocess_imputation_median_values(self):
         self.obj.mat = pd.DataFrame(
             {"a": [2, np.nan, 4], "b": [5, 4, 4], "c": [np.nan, 10, np.nan]}
         )
-        self.obj.preprocess(imputation="median")
+        self.obj.preprocess(log2_transform=False,imputation="median")
         expected_mat = pd.DataFrame(
             {"a": [2.0, 3.0, 4.0], "b": [5.0, 4.0, 4.0], "c": [10.0, 10.0, 10.0]}
         )
@@ -326,7 +326,7 @@ def test_preprocess_imputation_knn_values(self):
         self.obj.mat = pd.DataFrame(
             {"a": [2, np.nan, 4], "b": [5, 4, 4], "c": [np.nan, 10, np.nan]}
         )
-        self.obj.preprocess(imputation="knn")
+        self.obj.preprocess(log2_transform=False,imputation="knn")
         expected_mat = pd.DataFrame(
             {"a": [2.0, 3.0, 4.0], "b": [5.0, 4.0, 4.0], "c": [10.0, 10.0, 10.0]}
         )
@@ -336,7 +336,7 @@ def test_preprocess_imputation_randomforest_values(self):
         self.obj.mat = pd.DataFrame(
             {"a": [2, np.nan, 4], "b": [5, 4, 4], "c": [np.nan, 10, np.nan]}
         )
-        self.obj.preprocess(imputation="randomforest")
+        self.obj.preprocess(log2_transform=False,imputation="randomforest")
         expected_mat = pd.DataFrame(
             {
                 "a": [2.00000000e00, -9.22337204e12, 4.00000000e00],
@@ -372,14 +372,14 @@ def test_plot_correlation_matrix(self):
         )
 
     def test_plot_clustermap(self):
-        self.obj.preprocess(imputation="knn")
+        self.obj.preprocess(log2_transform=False, imputation="knn")
         plot = self.obj.plot_clustermap()
         first_row = plot.data2d.iloc[0].to_list()
         expected = [487618.5371077078, 1293013.103298046]
         self.assertEqual(first_row, expected)
 
     def test_plot_clustermap_with_label_bar(self):
-        self.obj.preprocess(imputation="knn")
+        self.obj.preprocess(log2_transform=False, imputation="knn")
         plot = self.obj.plot_clustermap(label_bar=self.comparison_column)
         first_row = plot.data2d.iloc[0].to_list()
         expected = [487618.5371077078, 1293013.103298046]
@@ -456,17 +456,10 @@ def test_plot_volcano_compare_preprocessing_modes(self):
             group2=["1_71_F10", "1_73_F12"],
             compare_preprocessing_modes=True
         )
-        self.assertEqual(len(result_list), 9)
-        # check if results are different
-        # for idx, res in enumerate(result_list):
-        #     for idx2, res2 in enumerate(result_list):
-        #         if idx != idx2:
-        #             difference = dictdiffer.diff(res.to_plotly_json(), res2.to_plotly_json())
-        #             self.assertNotEqual(len(list(difference)), 0)
-
+        self.assertEqual(len(result_list), 9)               
 
     def test_preprocess_subset(self):
-        self.obj.preprocess(subset=True)
+        self.obj.preprocess(subset=True, log2_transform=False)
         self.assertEqual(self.obj.mat.shape, (48, 1364))
 
     @patch.object(Statistics, "tukey_test")
@@ -552,7 +545,7 @@ def test_plot_volcano_wald(self):
         self.assertTrue(column_added)  
 
     def test_plot_volcano_sam(self):
-        self.obj.preprocess(imputation="knn", normalization="zscore")
+        self.obj.preprocess(log2_transform=False, imputation="knn", normalization="zscore")
         plot = self.obj.plot_volcano(
             column = "disease",
             group1="type 2 diabetes mellitus", 
@@ -758,7 +751,7 @@ def test_volcano_plot_anova(self):
         plot = self.obj.plot_volcano(
             column="grouping1", group1="Healthy", group2="Disease", method="anova"
         )
-        expected_y_value = 0.09437708068494619
+        expected_y_value = 0.040890177695653236
         y_value = plot.to_plotly_json().get("data")[0].get("y")[1]
         self.assertAlmostEqual(y_value, expected_y_value)