SigProfilerAssignment 0.0.29__tar.gz → 0.0.31__tar.gz

{SigProfilerAssignment-0.0.29/SigProfilerAssignment.egg-info → SigProfilerAssignment-0.0.31}/PKG-INFO

@@ -1,12 +1,11 @@
 Metadata-Version: 2.1
 Name: SigProfilerAssignment
-Version: 0.0.29
+Version: 0.0.31
 Summary: Mutational signatures attribution and decomposition tool
 Home-page: https://github.com/AlexandrovLab/SigProfilerAssignment.git
 Author: Raviteja Vangara
 Author-email: rvangara@health.ucsd.edu
 License: UCSD
-Platform: UNKNOWN
 Description-Content-Type: text/markdown
 License-File: LICENSE.txt
 
@@ -175,5 +174,3 @@ This software and its documentation are copyright 2022 as a part of the SigProfi
 
 ## <a name="contact"></a> Contact Information
 Please address any queries or bug reports to Raviteja Vangara at rvangara@health.ucsd.edu or Marcos Díaz-Gay at mdiazgay@health.ucsd.edu.
-
-

{SigProfilerAssignment-0.0.29 → SigProfilerAssignment-0.0.31}/SigProfilerAssignment/decompose_subroutines.py

@@ -480,7 +480,7 @@ def signature_decomposition(signatures, mtype, directory, genome_build="GRCh37",
 def make_final_solution(processAvg, allgenomes, allsigids, layer_directory, m, index, allcolnames, process_std_error = "none", signature_stabilities = " ", \
                         signature_total_mutations= " ", signature_stats = "none",  cosmic_sigs=False, attribution= 0, denovo_exposureAvg  = "none", add_penalty=0.05, \
                         remove_penalty=0.01, initial_remove_penalty=0.05, de_novo_fit_penalty=0.02, background_sigs=0, genome_build="GRCh37", sequence="genome", export_probabilities=True, export_probabilities_per_mutation=False, \
-                        refit_denovo_signatures=True, collapse_to_SBS96=True, connected_sigs=True, pcawg_rule=False, verbose=False,make_plots = True, samples='./', input_type='matrix', denovo_refit_option=True):
+                        refit_denovo_signatures=True, collapse_to_SBS96=True, connected_sigs=True, pcawg_rule=False, verbose=False,make_plots = True, samples='./', input_type='matrix', denovo_refit_option=True, exome=False):
 
     if processAvg.shape[0]==allgenomes.shape[0] and processAvg.shape[0] != 96:
         collapse_to_SBS96=False
@@ -794,7 +794,7 @@ def make_final_solution(processAvg, allgenomes, allsigids, layer_directory, m, i
         if export_probabilities==True:
             if input_type=='vcf':
                 if m=='96' or m=='78' or m=='83':
-                    probability_per_mutation, samples_prob_per_mut = probabilities_per_mutation(probability, samples, m)
+                    probability_per_mutation, samples_prob_per_mut = probabilities_per_mutation(probability, samples, m, exome)
 
                     if denovo_refit_option==True:
                         if refit_denovo_signatures==True:
@@ -941,7 +941,7 @@ def probabilities(W, H, index, allsigids, allcolnames):
 
 
 ################################################### Generation of probabilities for each processes given to A mutation ############################################
-def probabilities_per_mutation(probability_matrix, samples_path, m):  
+def probabilities_per_mutation(probability_matrix, samples_path, m, exome=False):  
 #
     probability_matrix=probability_matrix.reset_index()
 #
@@ -963,11 +963,15 @@ def probabilities_per_mutation(probability_matrix, samples_path, m):
 #
     all_mutations = pd.DataFrame()
     for file in seqinfo_files:
-        try:
-            new = pd.read_csv(seqinfo_path + file, sep='\t',header=None)
-            all_mutations = pd.concat([all_mutations, new])
-        except (pd.errors.EmptyDataError):
-            pass
+        if 'exome' in file:
+            exome_df = pd.read_csv(seqinfo_path + file, sep='\t',header=None)
+        else:
+            try:
+                new = pd.read_csv(seqinfo_path + file, sep='\t',header=None)
+                all_mutations = pd.concat([all_mutations, new])
+            except (pd.errors.EmptyDataError):
+                pass
+
     all_mutations[3] = all_mutations[3].str[interval_low:interval_high]
     if m=='96' or m=='78':
         del all_mutations[4]
@@ -977,6 +981,14 @@ def probabilities_per_mutation(probability_matrix, samples_path, m):
         del all_mutations[4]
 
     all_mutations.columns = ['Sample Names', 'Chr', 'Pos', 'MutationType']
+    if exome==True:
+        del exome_df[2]
+        del exome_df[3]
+        del exome_df[4]
+        exome_df.columns = ['Chr', 'Pos']
+        exome_df['Chr'] = [str(x) for x in (exome_df['Chr']).to_list()]
+        all_mutations['Chr'] = [str(x) for x in (all_mutations['Chr']).to_list()]
+        all_mutations = pd.merge(all_mutations, exome_df)
 #
     all_samples_mutations = [y for x, y in all_mutations.groupby('Sample Names')]
 #

{SigProfilerAssignment-0.0.29 → SigProfilerAssignment-0.0.31}/SigProfilerAssignment/decomposition.py

@@ -58,7 +58,7 @@ def generate_sample_reconstruction(cosmic_sigs, samples_input, activities, outpu
     samples.reset_index(inplace=True)
     for sample_name in samples.columns[1:]:
         # basis names and their corresponding weights
-        subset = activities[activities["Samples"].str.contains(sample_name)]
+        subset = activities[activities["Samples"] == sample_name]
         subset = subset.loc[:, (subset != 0).any(axis=0)]
         basis_names = subset[subset["Samples"].str.contains(sample_name)].columns[1:].tolist()
         recon_tmb = subset.sum(axis=1)
@@ -412,7 +412,7 @@ def spa_analyze(samples, output, input_type='matrix', context_type="96", signatu
                                     add_penalty=nnls_add_penalty, remove_penalty=nnls_remove_penalty,
                                     initial_remove_penalty=init_rem_denovo,connected_sigs=connected_sigs,refit_denovo_signatures=False,
                                     make_plots=make_plots,export_probabilities=export_probabilities, export_probabilities_per_mutation=export_probabilities_per_mutation,
-                                    samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option)
+                                    samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option, exome=exome)
 
         else:
             signature_stabilities=devopts['signature_stabilities']
@@ -426,7 +426,7 @@ def spa_analyze(samples, output, input_type='matrix', context_type="96", signatu
                                     background_sigs=background_sigs, verbose=verbose, genome_build=genome_build, signature_total_mutations = signature_total_mutations,
                                     add_penalty=nnls_add_penalty, remove_penalty=nnls_remove_penalty, process_std_error = processSTE, signature_stabilities = signature_stabilities,
                                     initial_remove_penalty=init_rem_denovo,connected_sigs=connected_sigs,refit_denovo_signatures=True,export_probabilities=export_probabilities,
-                                    export_probabilities_per_mutation=export_probabilities_per_mutation, samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option)
+                                    export_probabilities_per_mutation=export_probabilities_per_mutation, samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option, exome=exome)
         
         if make_metadata:
             with open(os.path.join(output,"JOB_METADATA_SPA.txt"),"a") as sysdata:
@@ -541,7 +541,7 @@ def spa_analyze(samples, output, input_type='matrix', context_type="96", signatu
                                 make_plots=make_plots,
                                 export_probabilities=export_probabilities,
                                 export_probabilities_per_mutation=export_probabilities_per_mutation,
-                                samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option)
+                                samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option, exome=exome)
 
         if make_metadata:
             with open(os.path.join(output,"JOB_METADATA_SPA.txt"),"a") as sysdata:
@@ -659,7 +659,7 @@ def spa_analyze(samples, output, input_type='matrix', context_type="96", signatu
                             make_plots =make_plots,
                             export_probabilities=export_probabilities,
                             export_probabilities_per_mutation=export_probabilities_per_mutation,
-                            samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option)
+                            samples=samples, input_type=input_type, denovo_refit_option=denovo_refit_option, exome=exome)
         if make_metadata:
             with open(os.path.join(output,"JOB_METADATA_SPA.txt"),"a") as sysdata:
                 current_time_end = datetime.datetime.now()

SigProfilerAssignment-0.0.31/SigProfilerAssignment/version.py

@@ -0,0 +1,8 @@
+
+# THIS FILE IS GENERATED FROM SigProfilerAssignment SETUP.PY
+short_version = '0.0.31'
+version = '0.0.31'
+Update = 'v0.0.31: Resolve issue with export probability per mutation for exome data'
+
+    
+    

{SigProfilerAssignment-0.0.29 → SigProfilerAssignment-0.0.31/SigProfilerAssignment.egg-info}/PKG-INFO

@@ -1,12 +1,11 @@
 Metadata-Version: 2.1
 Name: SigProfilerAssignment
-Version: 0.0.29
+Version: 0.0.31
 Summary: Mutational signatures attribution and decomposition tool
 Home-page: https://github.com/AlexandrovLab/SigProfilerAssignment.git
 Author: Raviteja Vangara
 Author-email: rvangara@health.ucsd.edu
 License: UCSD
-Platform: UNKNOWN
 Description-Content-Type: text/markdown
 License-File: LICENSE.txt
 
@@ -175,5 +174,3 @@ This software and its documentation are copyright 2022 as a part of the SigProfi
 
 ## <a name="contact"></a> Contact Information
 Please address any queries or bug reports to Raviteja Vangara at rvangara@health.ucsd.edu or Marcos Díaz-Gay at mdiazgay@health.ucsd.edu.
-
-

{SigProfilerAssignment-0.0.29 → SigProfilerAssignment-0.0.31}/SigProfilerAssignment.egg-info/SOURCES.txt

@@ -14,7 +14,6 @@ SigProfilerAssignment.egg-info/dependency_links.txt
 SigProfilerAssignment.egg-info/not-zip-safe
 SigProfilerAssignment.egg-info/requires.txt
 SigProfilerAssignment.egg-info/top_level.txt
-SigProfilerAssignment/DecompositionPlots/.DS_Store
 SigProfilerAssignment/DecompositionPlots/PlotDecomposition.py
 SigProfilerAssignment/DecompositionPlots/PlotDecomposition_CNV48.py
 SigProfilerAssignment/DecompositionPlots/PlotDecomposition_DBS78.py

{SigProfilerAssignment-0.0.29 → SigProfilerAssignment-0.0.31}/SigProfilerAssignment.egg-info/requires.txt

@@ -1,6 +1,6 @@
 scipy>=1.6.3
 numpy>=1.21.2
-pandas>=1.2.4
+pandas<2.0.0,>=1.2.4
 SigProfilerMatrixGenerator>=1.2.14
 sigProfilerPlotting>=1.3.11
 statsmodels>=0.9.0

SigProfilerAssignment-0.0.31/setup.py

@@ -0,0 +1,66 @@
+from setuptools import setup
+import shutil
+import os
+
+# remove the dist folder first if exists
+if os.path.exists("dist"):
+    shutil.rmtree("dist")
+
+VERSION = "0.0.31"
+
+
+def write_version_py(filename="SigProfilerAssignment/version.py"):
+    # Copied from numpy setup.py
+    cnt = """
+# THIS FILE IS GENERATED FROM SigProfilerAssignment SETUP.PY
+short_version = '%(version)s'
+version = '%(version)s'
+Update = 'v0.0.31: Resolve issue with export probability per mutation for exome data'
+
+    
+    """
+    fh = open(filename, "w")
+    fh.write(
+        cnt
+        % {
+            "version": VERSION,
+        }
+    )
+    fh.close()
+
+
+with open("README.md") as f:
+    long_description = f.read()
+
+requirements = [
+    "scipy>=1.6.3",
+    "numpy>=1.21.2",
+    "pandas>=1.2.4,<2.0.0",
+    "SigProfilerMatrixGenerator>=1.2.14",
+    "sigProfilerPlotting>=1.3.11",
+    "statsmodels>=0.9.0",
+    "scikit-learn>=0.24.2",
+    "psutil>=5.6.1",
+    "reportlab>=3.5.42",
+    "PyPDF2>=3.0.0",
+    "alive_progress>=2.4.1",
+    "pdf2image>=1.16.0",
+    "PyMuPDF>=1.21.0",
+]
+
+write_version_py()
+setup(
+    name="SigProfilerAssignment",
+    version=VERSION,
+    description="Mutational signatures attribution and decomposition tool",
+    long_description=long_description,
+    long_description_content_type="text/markdown",  # This is important!
+    url="https://github.com/AlexandrovLab/SigProfilerAssignment.git",
+    author="Raviteja Vangara",
+    author_email="rvangara@health.ucsd.edu",
+    license="UCSD",
+    packages=["SigProfilerAssignment"],
+    install_requires=requirements,
+    include_package_data=True,
+    zip_safe=False,
+)

SigProfilerAssignment-0.0.29/SigProfilerAssignment/version.py

@@ -1,8 +0,0 @@
-
-# THIS FILE IS GENERATED FROM SigProfilerAssignment SETUP.PY
-short_version = '0.0.29'
-version = '0.0.29'
-Update = 'v0.0.29: Stop decomposition plotting with custom signature database and update parameters for sample reconstruction plots'
-
-    
-    

SigProfilerAssignment-0.0.29/setup.py

@@ -1,58 +0,0 @@
-from setuptools import setup
-import shutil
-import os
-
-#remove the dist folder first if exists
-if os.path.exists("dist"):
-    shutil.rmtree("dist")
-
-VERSION = '0.0.29'
-
-
-def write_version_py(filename='SigProfilerAssignment/version.py'):
-    # Copied from numpy setup.py
-    cnt = """
-# THIS FILE IS GENERATED FROM SigProfilerAssignment SETUP.PY
-short_version = '%(version)s'
-version = '%(version)s'
-Update = 'v0.0.29: Stop decomposition plotting with custom signature database and update parameters for sample reconstruction plots'
-
-    
-    """
-    fh = open(filename, 'w')
-    fh.write(cnt % {'version': VERSION,})
-    fh.close()
-
-with open('README.md') as f:
-	long_description = f.read()
-
-requirements=[
-          'scipy>=1.6.3',
-          'numpy>=1.21.2',
-          'pandas>=1.2.4',
-          'SigProfilerMatrixGenerator>=1.2.14', 
-          'sigProfilerPlotting>=1.3.11',
-          'statsmodels>=0.9.0',
-          'scikit-learn>=0.24.2',
-          'psutil>=5.6.1',
-          'reportlab>=3.5.42',
-          'PyPDF2>=3.0.0',
-          'alive_progress>=2.4.1',
-          'pdf2image>=1.16.0',
-          'PyMuPDF>=1.21.0'
-           ]
-
-write_version_py() 
-setup(name='SigProfilerAssignment',
-      version=VERSION,
-      description='Mutational signatures attribution and decomposition tool',
-      long_description=long_description,
-      long_description_content_type='text/markdown',  # This is important!	
-      url="https://github.com/AlexandrovLab/SigProfilerAssignment.git",
-      author='Raviteja Vangara',
-      author_email='rvangara@health.ucsd.edu',
-      license='UCSD',
-      packages=['SigProfilerAssignment'],
-      install_requires=requirements,
-      include_package_data=True,      
-      zip_safe=False)