New CustomKinFragLib notebook

.github/workflows/ci.yml

@@ -62,7 +62,7 @@ jobs:
         shell: bash -l {0}
         run: |
           echo "Download combinatorial library from zenodo..."
-          wget -q -O data/combinatorial_library/combinatorial_library.tar.bz2 https://zenodo.org/record/17368450/files/combinatorial_library.tar.bz2?download=1
+          wget -q -O data/combinatorial_library/combinatorial_library.tar.bz2 https://zenodo.org/record/18386001/files/combinatorial_library.tar.bz2?download=1
           ls -l data/combinatorial_library/
           echo "Decompress selected files..."
           tar -xvf data/combinatorial_library/combinatorial_library.tar.bz2 combinatorial_library/combinatorial_library_deduplicated.json combinatorial_library/chembl_standardized_inchi.csv
@@ -74,6 +74,6 @@ jobs:
         shell: bash -l {0}
         run: |
           PYTEST_ARGS="--nbval-lax --nbval-current-env --nbval-cell-timeout=3600"
-          PYTEST_IGNORE="--ignore=notebooks/custom_kinfraglib/2_4_custom_filters_paper.ipynb --ignore=notebooks/custom_kinfraglib/1_4_custom_filters_pairwise_retrosynthesizability.ipynb --ignore=notebooks/custom_kinfraglib/2_1_custom_filters_pipeline.ipynb" 
+          PYTEST_IGNORE="--ignore=notebooks/custom_kinfraglib/2_5_custom_filters_paper.ipynb --ignore=notebooks/custom_kinfraglib/1_4_custom_filters_pairwise_retrosynthesizability.ipynb --ignore=notebooks/custom_kinfraglib/2_1_custom_filters_pipeline.ipynb --ignore=notebooks/custom_kinfraglib/2_4_custom_filters_enumeration_analysis.ipynb" 
 
           pytest $PYTEST_ARGS $PYTEST_IGNORE

data/combinatorial_library/README.md

@@ -1,18 +1,20 @@
 # KinFragLib: Combinatorial library 
 
-[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.17368450.svg)](https://doi.org/10.5281/zenodo.17368450)
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.18386001.svg)](https://doi.org/10.5281/zenodo.18386001)
 
 This folder is meant for the metadata and properties of the KinFragLib combinatorial library, which is based on the KinFragLib fragment library at https://github.com/volkamerlab/KinFragLib. This dataset is used for the analysis of the combinatorial library.
 
-**Note**: Since this dataset contains large files, we provide it outside this repository at https://zenodo.org/records/17368450 (DOI: 10.5281/zenodo.17368450, v2.0.2).
+**Note**: Since this dataset contains large files, we provide it outside this repository at https://zenodo.org/records/18386001 (DOI: 10.5281/zenodo.18386001, v2.0.3).
 In order to run the analysis notebooks, please download this dataset to this folder. 
 
 ## Raw data
 
 - `combinatorial_library.json`: Full combinatorial library, please refer to `notebooks/kinfraglib/4_1_combinatorial_library_data_preparation.ipynb` at https://github.com/volkamerlab/KinFragLib for detailed information about this data format
 - `combinatorial_library_deduplicated.json`: Deduplicated combinatorial library (based on InChIs)
-- `chembl_standardized_inchi.csv`: Standardized ChEMBL 36 molecules in the form of InChI strings.
+- `chembl_standardized_inchi.csv`: Standardized ChEMBL36 molecules in the form of InChI strings.
 - `KLIFS_download_summary.csv`: PDB codes of all KLIFS structures used to generate the KinFragLib fragmentation library. 
+- `combinatorial_library_custom_sampled.sdf`: Combinatorial library created from a subset of CustomKinFragLib fragments. 
+- `combinatorial_library_rejected_sampled.sdf`: Combinatorial library created from a subset of fragments rejected by the CustomKinFragLib filtering pipeline 
 
 ## Processed data
 
@@ -26,3 +28,5 @@ Data extracted from `combinatorial_library_deduplicated.json`, performed in `not
 - `chembl_exact.json`: Ligands with exact matches in ChEMBL
 - `chembl_most_similar.json`: Most similar ligand in ChEMBL for each recombined ligand 
 - `chembl_highly_similar.json`: Most similar ligand in ChEMBL for each recombined ligand with similarity greater than 0.9.
+- `custom_enamine_search_sampled.csv`: Most similar molecule from Enamine REAL Space for each molecule in the CustomKinFragLib combinatorial library. 
+- `reference_enamine_search_sampled.csv`: Most similar molecule from Enamine REAL Space for each molecule in the rejected fragments combinatorial library. 

kinfraglib/utils.py

@@ -1328,16 +1328,20 @@ def construct_ligand(fragment_ids, bond_ids, fragment_library):
     ed_combo = Chem.EditableMol(combo)
     replaced_dummies = []
 
-    atoms = combo.GetAtoms()
-
     for bond in bond_ids:
 
-        dummy_1 = next(
-            atom for atom in combo.GetAtoms() if atom.GetProp("fragment_atom_id") == bond[0]
-        )
-        dummy_2 = next(
-            atom for atom in combo.GetAtoms() if atom.GetProp("fragment_atom_id") == bond[1]
-        )
+        # should be a one element lists
+        dummy_1_candidates = [atom for atom in combo.GetAtoms() if atom.GetProp("fragment_atom_id") == bond[0]]
+        dummy_2_candidates = [atom for atom in combo.GetAtoms() if atom.GetProp("fragment_atom_id") == bond[1]]
+
+        if len(dummy_1_candidates) == 0 or len(dummy_2_candidates) == 0:
+            raise RuntimeError(f'Dummy atoms for bond {bond} not found')
+        elif len(dummy_1_candidates) > 1 or len(dummy_2_candidates) > 1:
+            raise RuntimeError(f'This should not happen: Dummy atoms found for bond {bond} are unambigious')
+
+        dummy_1 = dummy_1_candidates[0]
+        dummy_2 = dummy_2_candidates[0]
+
         atom_1 = dummy_1.GetNeighbors()[0]
         atom_2 = dummy_2.GetNeighbors()[0]