Merge pull request #140 from daisybio/development

New version v1.2.2

Author: JudithBernett

.github/workflows/run_tests.yml

@@ -66,7 +66,7 @@ jobs:
           print("::set-output name=result::{}".format(result))
 
       - name: Restore pre-commit cache
-        uses: actions/cache@v4.2.0
+        uses: actions/cache@v4.2.1
         if: matrix.session == 'pre-commit'
         with:
           path: ~/.cache/pre-commit

docs/conf.py

@@ -56,9 +56,9 @@
 # the built documents.
 #
 # The short X.Y version.
-version = "1.2.1"
+version = "1.2.2"
 # The full version, including alpha/beta/rc tags.
-release = "1.2.1"
+release = "1.2.2"
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

drevalpy/models/MOLIR/molir.py

@@ -165,11 +165,11 @@ def load_cell_line_features(self, data_path: str, dataset_name: str) -> FeatureD
         feature_dataset = get_multiomics_feature_dataset(
             data_path=data_path,
             dataset_name=dataset_name,
-            gene_list=None,
+            gene_lists=None,
             omics=self.cell_line_views,
         )
-        # log transformation
-        feature_dataset.apply(function=np.log, view="gene_expression")
+        # log transformation replaced with arcsinh transformation since log(0) is undefined
+        feature_dataset.apply(function=np.arcsinh, view="gene_expression")
         return feature_dataset
 
     def load_drug_features(self, data_path: str, dataset_name: str) -> FeatureDataset | None:

drevalpy/models/SimpleNeuralNetwork/multiomics_neural_network.py

@@ -178,7 +178,14 @@ def load_cell_line_features(self, data_path: str, dataset_name: str) -> FeatureD
         :return: FeatureDataset containing the cell line omics features, filtered through the
             drug target genes
         """
-        return get_multiomics_feature_dataset(data_path=data_path, dataset_name=dataset_name)
+        gene_lists = {
+            "gene_expression": "drug_target_genes_all_drugs",
+            "methylation": None,
+            "mutations": "drug_target_genes_all_drugs",
+            "copy_number_variation_gistic": "drug_target_genes_all_drugs",
+            "proteomics": "drug_target_genes_all_drugs_proteomics",
+        }
+        return get_multiomics_feature_dataset(data_path=data_path, gene_lists=gene_lists, dataset_name=dataset_name)
 
     def load_drug_features(self, data_path: str, dataset_name: str) -> FeatureDataset:
         """

drevalpy/models/SuperFELTR/superfeltr.py

@@ -250,10 +250,10 @@ def load_cell_line_features(self, data_path: str, dataset_name: str) -> FeatureD
         :returns: FeatureDataset containing the cell line gene expression features, mutations, and copy number variation
         """
         feature_dataset = get_multiomics_feature_dataset(
-            data_path=data_path, dataset_name=dataset_name, gene_list=None, omics=self.cell_line_views
+            data_path=data_path, dataset_name=dataset_name, gene_lists=None, omics=self.cell_line_views
         )
         # log transformation
-        feature_dataset.apply(function=np.log, view="gene_expression")
+        feature_dataset.apply(function=np.arcsinh, view="gene_expression")
         return feature_dataset
 
     def load_drug_features(self, data_path: str, dataset_name: str) -> FeatureDataset | None:

drevalpy/models/baselines/multi_omics_random_forest.py

@@ -55,7 +55,14 @@ def load_cell_line_features(self, data_path: str, dataset_name: str) -> FeatureD
         :returns: FeatureDataset containing the cell line omics features, filtered through the
             drug target genes
         """
-        return get_multiomics_feature_dataset(data_path=data_path, dataset_name=dataset_name)
+        gene_lists = {
+            "gene_expression": "drug_target_genes_all_drugs",
+            "methylation": None,
+            "mutations": "drug_target_genes_all_drugs",
+            "copy_number_variation_gistic": "drug_target_genes_all_drugs",
+            "proteomics": "drug_target_genes_all_drugs_proteomics",
+        }
+        return get_multiomics_feature_dataset(data_path=data_path, gene_lists=gene_lists, dataset_name=dataset_name)
 
     def train(
         self,

drevalpy/models/utils.py

@@ -139,35 +139,43 @@ def load_drug_fingerprint_features(data_path: str, dataset_name: str, default_ra
 def get_multiomics_feature_dataset(
     data_path: str,
     dataset_name: str,
-    gene_list: str | None = "drug_target_genes_all_drugs",
+    gene_lists: dict | None = None,
     omics: list[str] | None = None,
 ) -> FeatureDataset:
     """
     Get multiomics feature dataset for the given list of OMICs.
 
     :param data_path: path to the data, e.g., data/
     :param dataset_name: name of the dataset, e.g., GDSC2
-    :param gene_list: list of genes to include, e.g., landmark_genes
+    :param gene_lists: dictionary of names of lists of genes to include, for each omics type,
+                e.g., {"gene_expression": "landmark_genes"}, if None, all features are not reduced
     :param omics: list of omics to include, e.g., ["gene_expression", "methylation"]
     :returns: FeatureDataset with the multiomics features
     :raises ValueError: if no omics features are found
     """
     if omics is None:
         omics = ["gene_expression", "methylation", "mutations", "copy_number_variation_gistic", "proteomics"]
+
+    if gene_lists is None:
+        gene_lists = {o: None for o in omics}
+
+    if not np.all([k in omics for k in gene_lists.keys()]):
+        raise ValueError("Gene lists must be provided for all omics types.")
+
     feature_dataset = None
     for omic in omics:
         if feature_dataset is None:
             feature_dataset = load_and_reduce_gene_features(
                 feature_type=omic,
-                gene_list=None if omic == "methylation" else gene_list,
+                gene_list=gene_lists[omic],
                 data_path=data_path,
                 dataset_name=dataset_name,
             )
         else:
             feature_dataset.add_features(
                 load_and_reduce_gene_features(
                     feature_type=omic,
-                    gene_list=None if omic == "methylation" else gene_list,
+                    gene_list=gene_lists[omic],
                     data_path=data_path,
                     dataset_name=dataset_name,
                 )