Ruff formatting

src/signatures/associations/genotype/samples.py

@@ -1,42 +1,92 @@
-# Import standard modules
-import sys, numpy as np, pandas as pd
-from signatures.associations.sampleCuration.pan_cancer_sample import getSamples
+import sys
 
+import numpy as np
+import pandas as pd
 
-if __name__=='__main__':
+from signatures.associations.sampleCuration.pan_cancer_sample import getSamples
 
+if __name__ == "__main__":
     # filename to save covariates into
     filename_output = sys.argv[1]
 
     # GET DATA - include hypermutated samples
     print("Load samples")
-    sample_df = getSamples(PC=True, signatures=True, primary=True, ethnicity=False, hyper=True, ER=False, MSI=False)
+    sample_df = getSamples(
+        PC=True,
+        signatures=True,
+        primary=True,
+        ethnicity=False,
+        hyper=True,
+        ER=False,
+        MSI=False,
+    )
 
     # Remove samples where sex doesn't match organ
-    sample_df = sample_df[ ( ~(sample_df.tumour_tissue.map(lambda x: x in ["Breast","Ovary","Uterus"])&(sample_df.is_female==-1)) ) &\
-                           ( ~(sample_df.tumour_tissue.map(lambda x: x in ["Testis", "Prostate"])&(sample_df.is_female==1)) ) ]
+    sample_df = sample_df[
+        (
+            ~(
+                sample_df.tumour_tissue.map(
+                    lambda x: x in ["Breast", "Ovary", "Uterus"]
+                )
+                & (sample_df.is_female == -1)
+            )
+        )
+        & (
+            ~(
+                sample_df.tumour_tissue.map(lambda x: x in ["Testis", "Prostate"])
+                & (sample_df.is_female == 1)
+            )
+        )
+    ]
     print("Removed male samples for Breast,Ovary,Uterus and female for Testis,Prostate")
     print(f"Number of samples: {len(sample_df)}")
 
     # Confounding variables
-    variables = ['log_age', 'is_female'] + [f"pc{i}" for i in range(1,4)] # logit_purity',
+    variables = ["log_age", "is_female"] + [
+        f"pc{i}" for i in range(1, 4)
+    ]  # logit_purity',
 
     # Construct X matrix
     print("Construct X")
-    X = pd.DataFrame({variable:sample_df[variable] for variable in variables+['tumour_tissue', 'tumour_group',
-                                                                              'participant_id', 'tumour_sample_platekey', 'germline_sample_platekey']})
+    X = pd.DataFrame(
+        {
+            variable: sample_df[variable]
+            for variable in variables
+            + [
+                "tumour_tissue",
+                "tumour_group",
+                "participant_id",
+                "tumour_sample_platekey",
+                "germline_sample_platekey",
+            ]
+        }
+    )
 
     # Get non-nan values
-    [print(key, np.sum(~np.isnan(X[key]))) for key in X.columns if X[key].dtype in [float, int]]
-    cuts = np.prod(np.array([~np.isnan(X[key]) for key in X.columns if X[key].dtype in [float, int]]),
-                   axis=0).astype(bool)
+    [
+        print(key, np.sum(~np.isnan(X[key])))
+        for key in X.columns
+        if X[key].dtype in [float, int]
+    ]
+    cuts = np.prod(
+        np.array(
+            [~np.isnan(X[key]) for key in X.columns if X[key].dtype in [float, int]]
+        ),
+        axis=0,
+    ).astype(bool)
     print(f"Non-nan: {np.sum(cuts)}")
 
     # Apply cuts
     X = X[cuts]
 
     # Save
-    X['sample_id'] = X.participant_id.astype(str)+"_"+X.tumour_sample_platekey+"_"+X.germline_sample_platekey
-    X.rename(columns={"tumour_tissue":"group"}, inplace=True)
+    X["sample_id"] = (
+        X.participant_id.astype(str)
+        + "_"
+        + X.tumour_sample_platekey
+        + "_"
+        + X.germline_sample_platekey
+    )
+    X.rename(columns={"tumour_tissue": "group"}, inplace=True)
     print(X.columns)
     X[["sample_id", "group"] + variables].to_csv(filename_output, index=False, sep="\t")

src/signatures/associations/genotype/samples_and_targets.py

@@ -1,98 +1,204 @@
-# Import standard modules
-import sys, os, numpy as np, pandas as pd
+import os
 import re
-from signatures.associations.sampleCuration.pan_cancer_sample import getSamples
-from dotenv import load_dotenv
+import sys
 import warnings
 
-warnings.filterwarnings('ignore')
+import numpy as np
+import pandas as pd
+from dotenv import load_dotenv
 
-load_dotenv()
-DATA_DIR = os.getenv('DATA_DIR')
+from signatures.associations.sampleCuration.pan_cancer_sample import getSamples
+
+warnings.filterwarnings("ignore")
 
-if __name__=="__main__":
+load_dotenv()
+DATA_DIR = os.getenv("DATA_DIR")
 
+if __name__ == "__main__":
     samples_file, targets_file = sys.argv[1:]
 
     # Get sample dataframe
     print("Load samples")
     sample_df = getSamples(hyper=True)
-    sample_df['sample_id'] = sample_df.participant_id.astype(str)+"_"+sample_df.tumour_sample_platekey+"_"+sample_df.germline_sample_platekey
+    sample_df["sample_id"] = (
+        sample_df.participant_id.astype(str)
+        + "_"
+        + sample_df.tumour_sample_platekey
+        + "_"
+        + sample_df.germline_sample_platekey
+    )
 
     # Remove samples where sex doesn't match organ
-    sample_df = sample_df[ ( ~(sample_df.tumour_tissue.map(lambda x: x in ["Ovary","Uterus"])&(sample_df.is_female==-1)) ) &\
-                           ( ~(sample_df.tumour_tissue.map(lambda x: x in ["Testis", "Prostate"])&(sample_df.is_female==1)) ) ]
+    sample_df = sample_df[
+        (
+            ~(
+                sample_df.tumour_tissue.map(lambda x: x in ["Ovary", "Uterus"])
+                & (sample_df.is_female == -1)
+            )
+        )
+        & (
+            ~(
+                sample_df.tumour_tissue.map(lambda x: x in ["Testis", "Prostate"])
+                & (sample_df.is_female == 1)
+            )
+        )
+    ]
     print("Removed male samples for Ovary, Uterus and female for Testis, Prostate")
     print(f"Number of samples: {len(sample_df)}")
 
     # Get genotype grid
-    genotype_df = pd.read_csv(f"{DATA_DIR}/pancancer_mutationTimeR_updated_CNACalls_2023_09_16.tsv",
-                                usecols=['participant_id', 'tumour_sample_platekey', 'germline_sample_platekey', 'is_WGD'], sep="\t") # 'ploidy',
+    genotype_df = pd.read_csv(
+        f"{DATA_DIR}/pancancer_mutationTimeR_updated_CNACalls_2023_09_16.tsv",
+        usecols=[
+            "participant_id",
+            "tumour_sample_platekey",
+            "germline_sample_platekey",
+            "is_WGD",
+        ],
+        sep="\t",
+    )  # 'ploidy',
     # genotype_df['ploidy'] = (genotype_df.ploidy>2.5).astype(int)
-    genotype_df['is_WGD'] = genotype_df.is_WGD.astype(int)
-    genotype_df['sample_id'] = genotype_df.participant_id.astype(str)+"_"+genotype_df.tumour_sample_platekey+"_"+genotype_df.germline_sample_platekey
-    genotype_df = genotype_df.drop(np.intersect1d(genotype_df.keys(),
-                                                    ['participant_id', 'tumour_sample_platekey', 'germline_sample_platekey',
-                                                     'day_sampling', 'sample_platekey', 'tumour_pseudo_id']), axis=1)
+    genotype_df["is_WGD"] = genotype_df.is_WGD.astype(int)
+    genotype_df["sample_id"] = (
+        genotype_df.participant_id.astype(str)
+        + "_"
+        + genotype_df.tumour_sample_platekey
+        + "_"
+        + genotype_df.germline_sample_platekey
+    )
+    genotype_df = genotype_df.drop(
+        np.intersect1d(
+            genotype_df.keys(),
+            [
+                "participant_id",
+                "tumour_sample_platekey",
+                "germline_sample_platekey",
+                "day_sampling",
+                "sample_platekey",
+                "tumour_pseudo_id",
+            ],
+        ),
+        axis=1,
+    )
 
     # Get the number of chromothripsis and chromoplexy mutations
     sv_dir = "/re_gecip/cancer_pan/rhoulston/analysisResults/SVClassification/"
     sv_counts = pd.DataFrame()
     for cohort in os.listdir(sv_dir):
-        if cohort!='README.md':
-            sv_counts = pd.concat((sv_counts,
-                                pd.read_csv(f"{sv_dir}/{cohort}/summaryLevel/sampleClusterClassCounts.tsv", sep="\t")))
-    sv_counts['sample_id'] = sv_counts.participant_id.astype(str)+"_"+sv_counts.tumour_sample_platekey+"_"+sv_counts.germline_sample_platekey
+        if cohort != "README.md":
+            sv_counts = pd.concat(
+                (
+                    sv_counts,
+                    pd.read_csv(
+                        f"{sv_dir}/{cohort}/summaryLevel/sampleClusterClassCounts.tsv",
+                        sep="\t",
+                    ),
+                )
+            )
+    sv_counts["sample_id"] = (
+        sv_counts.participant_id.astype(str)
+        + "_"
+        + sv_counts.tumour_sample_platekey
+        + "_"
+        + sv_counts.germline_sample_platekey
+    )
 
     # sv_counts = sv_counts[~pd.isnull(sv_counts['n_chromothripsis'])]
     # cluster_types = ['n_chromoplexy', 'n_chromothripsis', 'n_tandem_duplication']
-    cluster_types = ['n_tandem_duplication']
+    cluster_types = ["n_tandem_duplication"]
 
     for cluster_type in cluster_types:
-        sv_counts[re.sub("^n_", "is_", cluster_type)] = (sv_counts[cluster_type]>0).astype(int)
-    genotype_df = pd.merge(genotype_df,
-                           sv_counts[['sample_id']+cluster_types+[re.sub("^n_", "is_", cluster_type) for cluster_type in cluster_types]],
-                           on='sample_id', how='inner')
+        sv_counts[re.sub("^n_", "is_", cluster_type)] = (
+            sv_counts[cluster_type] > 0
+        ).astype(int)
+    genotype_df = pd.merge(
+        genotype_df,
+        sv_counts[
+            ["sample_id"]
+            + cluster_types
+            + [re.sub("^n_", "is_", cluster_type) for cluster_type in cluster_types]
+        ],
+        on="sample_id",
+        how="inner",
+    )
 
     # Get the number of Kataegis events
-    ktg_dir = f"{DATA_DIR}/kataegis_analysis/results/SigProfilerClusters_persample/summary/"
-    ktg_counts = pd.read_csv(f"{ktg_dir}/All_10983_clustered_events_table.tsv", sep="\t",usecols=['sample_id', 'kataegis'])
-    ktg_counts.rename({'kataegis':'n_kataegis'}, axis=1, inplace=True)
-    ktg_counts = ktg_counts[~pd.isnull(ktg_counts['n_kataegis'])]
-    cluster_types = ['n_kataegis']
+    ktg_dir = (
+        f"{DATA_DIR}/kataegis_analysis/results/SigProfilerClusters_persample/summary/"
+    )
+    ktg_counts = pd.read_csv(
+        f"{ktg_dir}/All_10983_clustered_events_table.tsv",
+        sep="\t",
+        usecols=["sample_id", "kataegis"],
+    )
+    ktg_counts.rename({"kataegis": "n_kataegis"}, axis=1, inplace=True)
+    ktg_counts = ktg_counts[~pd.isnull(ktg_counts["n_kataegis"])]
+    cluster_types = ["n_kataegis"]
     for cluster_type in cluster_types:
-        ktg_counts[re.sub("^n_", "is_", cluster_type)] = (ktg_counts[cluster_type]>0).astype(int)
-    genotype_df = pd.merge(genotype_df,
-                           ktg_counts[['sample_id']+cluster_types+[re.sub("^n_", "is_", cluster_type) for cluster_type in cluster_types]],
-                           on='sample_id', how='inner')
+        ktg_counts[re.sub("^n_", "is_", cluster_type)] = (
+            ktg_counts[cluster_type] > 0
+        ).astype(int)
+    genotype_df = pd.merge(
+        genotype_df,
+        ktg_counts[
+            ["sample_id"]
+            + cluster_types
+            + [re.sub("^n_", "is_", cluster_type) for cluster_type in cluster_types]
+        ],
+        on="sample_id",
+        how="inner",
+    )
 
     # Sample list which is also in NCRAS
     sample_list = np.intersect1d(genotype_df.sample_id, sample_df.sample_id)
     sample_df = sample_df.set_index("sample_id").loc[sample_list]
     genotype_df = genotype_df.set_index("sample_id").loc[sample_list]
 
     # Confounding variables
-    variables = ['log_age', 'is_female'] + [f"pc{i}" for i in range(1,4)]
+    variables = ["log_age", "is_female"] + [f"pc{i}" for i in range(1, 4)]
 
     # Construct X matrix
     print("Construct X")
-    X = pd.DataFrame({variable:sample_df[variable] for variable in variables+['tumour_tissue', 'tumour_group',
-                                                                              'participant_id',
-                                                                              'tumour_sample_platekey',
-                                                                              'germline_sample_platekey']})
+    X = pd.DataFrame(
+        {
+            variable: sample_df[variable]
+            for variable in variables
+            + [
+                "tumour_tissue",
+                "tumour_group",
+                "participant_id",
+                "tumour_sample_platekey",
+                "germline_sample_platekey",
+            ]
+        }
+    )
 
     # Get non-nan values
-    [print(key, np.sum(~np.isnan(X[key]))) for key in X.columns if X[key].dtype in [float, int]]
-    cuts = np.prod(np.array([~np.isnan(X[key]) for key in X.columns if X[key].dtype in [float, int]]),
-                   axis=0).astype(bool)
+    [
+        print(key, np.sum(~np.isnan(X[key])))
+        for key in X.columns
+        if X[key].dtype in [float, int]
+    ]
+    cuts = np.prod(
+        np.array(
+            [~np.isnan(X[key]) for key in X.columns if X[key].dtype in [float, int]]
+        ),
+        axis=0,
+    ).astype(bool)
     print(f"Non-nan: {np.sum(cuts)}")
 
     # Apply cuts
     X = X[cuts]
 
     # Save
-    X['sample_id'] = X.participant_id.astype(str)+"_"+X.tumour_sample_platekey+"_"+X.germline_sample_platekey
-    X.rename(columns={"tumour_group":"group"}, inplace=True)
+    X["sample_id"] = (
+        X.participant_id.astype(str)
+        + "_"
+        + X.tumour_sample_platekey
+        + "_"
+        + X.germline_sample_platekey
+    )
+    X.rename(columns={"tumour_group": "group"}, inplace=True)
     print(X.columns)
     X[["sample_id", "group"] + variables].to_csv(samples_file, index=False, sep="\t")