Add new ChEMBL and UniProt datasets for AC50 activities, Phase 4 molecules, and nuclear proteins

Author: jbenjoseph

README.md

@@ -39,26 +39,29 @@ The default catalog includes local-file/HTTP datasets plus API presets useful in
 19. `chembl_activity_ic50_human`
 20. `chembl_activity_kd_human`
 21. `chembl_activity_ec50_human`
-22. `chembl_assays_binding_human`
-23. `chembl_assays_functional_human`
-24. `chembl_assays_adme_human`
-25. `chembl_targets_human_single_protein`
-26. `chembl_targets_human_protein_complex`
-27. `chembl_molecules_phase3plus`
-28. `chembl_molecules_black_box_warning`
-29. `chembl_mechanism_phase2plus`
-30. `chembl_drug_indications_phase2plus`
-31. `chembl_drug_indications_phase3plus`
-32. `uniprot_human_reviewed`
-33. `uniprot_human_receptors`
-34. `uniprot_human_membrane`
-35. `uniprot_human_kinases`
-36. `uniprot_human_gpcr`
-37. `uniprot_human_ion_channels`
-38. `uniprot_human_transporters`
-39. `uniprot_human_secreted`
-40. `uniprot_human_transcription_factors`
-41. `uniprot_human_enzymes`
+22. `chembl_activity_ac50_human`
+23. `chembl_assays_binding_human`
+24. `chembl_assays_functional_human`
+25. `chembl_assays_adme_human`
+26. `chembl_targets_human_single_protein`
+27. `chembl_targets_human_protein_complex`
+28. `chembl_molecules_phase3plus`
+29. `chembl_molecules_phase4`
+30. `chembl_molecules_black_box_warning`
+31. `chembl_mechanism_phase2plus`
+32. `chembl_drug_indications_phase2plus`
+33. `chembl_drug_indications_phase3plus`
+34. `uniprot_human_reviewed`
+35. `uniprot_human_receptors`
+36. `uniprot_human_membrane`
+37. `uniprot_human_nucleus`
+38. `uniprot_human_kinases`
+39. `uniprot_human_gpcr`
+40. `uniprot_human_ion_channels`
+41. `uniprot_human_transporters`
+42. `uniprot_human_secreted`
+43. `uniprot_human_transcription_factors`
+44. `uniprot_human_enzymes`
 
 Most of these are distributed through MoleculeNet/DeepChem mirrors and retain upstream licensing terms.
 ChEMBL and UniProt presets are fetched through their public REST APIs and cached locally as JSONL.

src/refua_data/catalog.py

@@ -475,6 +475,35 @@ def _zinc_druglike_tranche_urls(
         ),
         tags=("api", "chembl", "human", "ec50", "potency"),
     ),
+    DatasetDefinition(
+        dataset_id="chembl_activity_ac50_human",
+        name="ChEMBL Human AC50 Activities",
+        description=(
+            "ChEMBL activity records for human targets with AC50 and pChEMBL "
+            "values, useful for concentration-response modeling."
+        ),
+        source="ChEMBL REST API",
+        homepage="https://www.ebi.ac.uk/chembl/",
+        license_name="ChEMBL data terms",
+        license_url="https://www.ebi.ac.uk/chembl/ws",
+        file_format="jsonl",
+        category="target_activity",
+        api=ApiDatasetConfig(
+            endpoint="https://www.ebi.ac.uk/chembl/api/data/activity.json",
+            params={
+                "target_organism": "Homo sapiens",
+                "standard_type": "AC50",
+                "pchembl_value__isnull": "false",
+            },
+            pagination="chembl",
+            items_path="activities",
+            page_size_param="limit",
+            page_size=1000,
+            max_pages=40,
+            max_rows=25_000,
+        ),
+        tags=("api", "chembl", "human", "ac50", "potency"),
+    ),
     DatasetDefinition(
         dataset_id="chembl_assays_binding_human",
         name="ChEMBL Human Binding Assays",
@@ -640,6 +669,31 @@ def _zinc_druglike_tranche_urls(
         ),
         tags=("api", "chembl", "clinical", "phase3plus"),
     ),
+    DatasetDefinition(
+        dataset_id="chembl_molecules_phase4",
+        name="ChEMBL Molecules Phase 4",
+        description=(
+            "ChEMBL molecules with max clinical phase >= 4, useful for "
+            "marketed-drug priors and late-stage benchmark sets."
+        ),
+        source="ChEMBL REST API",
+        homepage="https://www.ebi.ac.uk/chembl/",
+        license_name="ChEMBL data terms",
+        license_url="https://www.ebi.ac.uk/chembl/ws",
+        file_format="jsonl",
+        category="compound_library",
+        api=ApiDatasetConfig(
+            endpoint="https://www.ebi.ac.uk/chembl/api/data/molecule.json",
+            params={"max_phase__gte": "4"},
+            pagination="chembl",
+            items_path="molecules",
+            page_size_param="limit",
+            page_size=1000,
+            max_pages=30,
+            max_rows=20_000,
+        ),
+        tags=("api", "chembl", "clinical", "phase4"),
+    ),
     DatasetDefinition(
         dataset_id="chembl_molecules_black_box_warning",
         name="ChEMBL Molecules with Black Box Warning",
@@ -824,6 +878,34 @@ def _zinc_druglike_tranche_urls(
         ),
         tags=("api", "uniprot", "human", "membrane", "target_family"),
     ),
+    DatasetDefinition(
+        dataset_id="uniprot_human_nucleus",
+        name="UniProt Human Nuclear Proteins",
+        description=(
+            "Reviewed human proteins annotated with nuclear localization for "
+            "nucleus-focused target enrichment and biology workflows."
+        ),
+        source="UniProt REST API",
+        homepage="https://www.uniprot.org/help/api_queries",
+        license_name="UniProt terms",
+        license_url="https://www.uniprot.org/help/license",
+        file_format="jsonl",
+        category="target_families",
+        api=ApiDatasetConfig(
+            endpoint="https://rest.uniprot.org/uniprotkb/search",
+            params={
+                "query": "organism_id:9606 AND reviewed:true AND keyword:Nucleus",
+                "format": "json",
+            },
+            pagination="link_header",
+            items_path="results",
+            page_size_param="size",
+            page_size=500,
+            max_pages=20,
+            max_rows=8_000,
+        ),
+        tags=("api", "uniprot", "human", "nucleus", "target_family"),
+    ),
     DatasetDefinition(
         dataset_id="uniprot_human_kinases",
         name="UniProt Human Kinases",

tests/test_catalog.py

@@ -11,17 +11,20 @@ def test_default_catalog_contains_core_and_api_datasets() -> None:
     assert "chembl_activity_ki_human" in ids
     assert "chembl_activity_kd_human" in ids
     assert "chembl_activity_ec50_human" in ids
+    assert "chembl_activity_ac50_human" in ids
     assert "chembl_assays_functional_human" in ids
     assert "chembl_assays_adme_human" in ids
+    assert "chembl_molecules_phase4" in ids
     assert "chembl_molecules_black_box_warning" in ids
     assert "chembl_mechanism_phase2plus" in ids
     assert "chembl_drug_indications_phase2plus" in ids
     assert "chembl_drug_indications_phase3plus" in ids
     assert "uniprot_human_reviewed" in ids
     assert "uniprot_human_receptors" in ids
     assert "uniprot_human_membrane" in ids
+    assert "uniprot_human_nucleus" in ids
     assert "uniprot_human_secreted" in ids
     assert "uniprot_human_transcription_factors" in ids
     assert "uniprot_human_enzymes" in ids
     assert "chembl_targets_human_protein_complex" in ids
-    assert len(datasets) >= 39
+    assert len(datasets) >= 42