gdsc

Name: gdsc
Author: QSong-github

132stars4forks

How to install gdsc?

qsong-github/drugclaw/gdsc

$npx -y skills add QSong-github/DrugClaw --skill gdsc

Installs into the current project.

Summary

| Field | Value | |---|---| | Category | Drug-centric | | Subcategory | Drug Molecular Property | | Source | Sanger / Wellcome Trust | | Datasets | screened_compounds (drug list), GDSC1/GDSC2 (dose-response), Cell Model Passports (cell-line annotations) | | URL | <htt

Files · 1

View on GitHub

SKILL.md

1	# 60_GDSC_GDSC2 — Genomics of Drug Sensitivity in Cancer
2
3	## Overview
4
5	\| Field \| Value \|
6	\|---\|---\|
7	\| Category \| Drug-centric \|
8	\| Subcategory \| Drug Molecular Property \|
9	\| Source \| Sanger / Wellcome Trust \|
10	\| Datasets \| screened_compounds (drug list), GDSC1/GDSC2 (dose-response), Cell Model Passports (cell-line annotations) \|
11	\| URL \| <https://www.cancerrxgene.org/> \|
12	\| Cell Models \| <https://cellmodelpassports.sanger.ac.uk/downloads> \|
13
14	GDSC contains pharmacological profiles for ~500 drugs tested in ~1,000 cancer cell lines. Queryable entities include drug names, gene targets, pathways, and cell-line identifiers.
15
16	## File Layout
17
18	```
19	DATA_DIR/
20	├── screened_compounds_rel_8.4.csv # drug list (~100 KB)
21	├── GDSC1_fitted_dose_response_27Oct23.xlsx # GDSC1 IC50/AUC (~80 MB, optional)
22	└── GDSC2_fitted_dose_response_27Oct23.xlsx # GDSC2 IC50/AUC (~50 MB, optional)
23	```
24
25	Default `DATA_DIR`:
26	```
27	resources_metadata/drug_molecular_property/GDSC
28	```
29
30	Override via environment variable: `export GDSC_DATA_DIR=/your/path`
31
32	## Dependencies
33
34	```bash
35	conda install openpyxl # or: pip install openpyxl
36	```
37
38	## Download & Query
39
40	The script auto-downloads all data files (drug list CSV + GDSC1/GDSC2 dose-response XLSX) on first run if the data directory is empty.
41
42	### CLI
43
44	```bash
45	# First run: auto-downloads all files, then queries default examples (Erlotinib, Nutlin, A549)
46	python 60_GDSC_GDSC2.py
47	```
48
49	If auto-download fails (e.g. no internet on HPC compute node), download manually from the repository root:
50
51	```bash
52	cd resources_metadata/drug_molecular_property/GDSC
53	wget 'https://ftp.sanger.ac.uk/pub/project/cancerrxgene/releases/current_release/screened_compounds_rel_8.4.csv'
54	wget 'https://cog.sanger.ac.uk/cancerrxgene/GDSC_data_8.5/GDSC1_fitted_dose_response_27Oct23.xlsx'
55	wget 'https://cog.sanger.ac.uk/cancerrxgene/GDSC_data_8.5/GDSC2_fitted_dose_response_27Oct23.xlsx'
56	```
57
58	### Python API
59
60	```python
61	from importlib.machinery import SourceFileLoader
62	mod = SourceFileLoader("gdsc", "60_GDSC_GDSC2.py").load_module()
63
64	# Single entity
65	results = mod.query_gdsc("Erlotinib")
66
67	# Multiple entities
68	results = mod.query_gdsc(["Nutlin", "A549", "EGFR"])
69
70	# Optional: manually trigger download
71	mod.download_gdsc_data()
72	```
73
74	### Return Format
75
76	```json
77	[
78	{
79	"source": "screened_compounds_rel_8.4.csv",
80	"match_count": 1,
81	"matches": [
82	{
83	"DRUG_NAME": "Erlotinib",
84	"TARGET": "EGFR",
85	"TARGET_PATHWAY": "EGFR signaling",
86	"PUBCHEM_ID": "176870",
87	"...": "..."
88	}
89	]
90	}
91	]
92	```
93
94	- Returns an empty list when no matches are found.
95	- Returns `{"error": "..."}` if the data directory is missing or empty.
96
97	### LLM Integration Example
98
99	```text
100	User: "What is the target of Erlotinib in GDSC?"
101	Agent: calls query_gdsc("Erlotinib")
102	→ source: screened_compounds_rel_8.4.csv, TARGET: EGFR, PATHWAY: EGFR signaling
103	→ "Erlotinib targets EGFR (EGFR signaling pathway) according to GDSC."
104	```

People who liked this also viewed

6 picks

Type

gdsc

132stars4forks

View on GitHub

How to install gdsc?

qsong-github/drugclaw/gdsc

$npx -y skills add QSong-github/DrugClaw --skill gdsc

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# 60_GDSC_GDSC2 — Genomics of Drug Sensitivity in Cancer
2
3	## Overview
4
5	\| Field \| Value \|
6	\|---\|---\|
7	\| Category \| Drug-centric \|
8	\| Subcategory \| Drug Molecular Property \|
9	\| Source \| Sanger / Wellcome Trust \|
10	\| Datasets \| screened_compounds (drug list), GDSC1/GDSC2 (dose-response), Cell Model Passports (cell-line annotations) \|
11	\| URL \| <https://www.cancerrxgene.org/> \|
12	\| Cell Models \| <https://cellmodelpassports.sanger.ac.uk/downloads> \|
13
14	GDSC contains pharmacological profiles for ~500 drugs tested in ~1,000 cancer cell lines. Queryable entities include drug names, gene targets, pathways, and cell-line identifiers.
15
16	## File Layout
17
18	```
19	DATA_DIR/
20	├── screened_compounds_rel_8.4.csv # drug list (~100 KB)
21	├── GDSC1_fitted_dose_response_27Oct23.xlsx # GDSC1 IC50/AUC (~80 MB, optional)
22	└── GDSC2_fitted_dose_response_27Oct23.xlsx # GDSC2 IC50/AUC (~50 MB, optional)
23	```
24
25	Default `DATA_DIR`:
26	```
27	resources_metadata/drug_molecular_property/GDSC
28	```
29
30	Override via environment variable: `export GDSC_DATA_DIR=/your/path`
31
32	## Dependencies
33
34	```bash
35	conda install openpyxl # or: pip install openpyxl
36	```
37
38	## Download & Query
39
40	The script auto-downloads all data files (drug list CSV + GDSC1/GDSC2 dose-response XLSX) on first run if the data directory is empty.
41
42	### CLI
43
44	```bash
45	# First run: auto-downloads all files, then queries default examples (Erlotinib, Nutlin, A549)
46	python 60_GDSC_GDSC2.py
47	```
48
49	If auto-download fails (e.g. no internet on HPC compute node), download manually from the repository root:
50
51	```bash
52	cd resources_metadata/drug_molecular_property/GDSC
53	wget 'https://ftp.sanger.ac.uk/pub/project/cancerrxgene/releases/current_release/screened_compounds_rel_8.4.csv'
54	wget 'https://cog.sanger.ac.uk/cancerrxgene/GDSC_data_8.5/GDSC1_fitted_dose_response_27Oct23.xlsx'
55	wget 'https://cog.sanger.ac.uk/cancerrxgene/GDSC_data_8.5/GDSC2_fitted_dose_response_27Oct23.xlsx'
56	```
57
58	### Python API
59
60	```python
61	from importlib.machinery import SourceFileLoader
62	mod = SourceFileLoader("gdsc", "60_GDSC_GDSC2.py").load_module()
63
64	# Single entity
65	results = mod.query_gdsc("Erlotinib")
66
67	# Multiple entities
68	results = mod.query_gdsc(["Nutlin", "A549", "EGFR"])
69
70	# Optional: manually trigger download
71	mod.download_gdsc_data()
72	```
73
74	### Return Format
75
76	```json
77	[
78	{
79	"source": "screened_compounds_rel_8.4.csv",
80	"match_count": 1,
81	"matches": [
82	{
83	"DRUG_NAME": "Erlotinib",
84	"TARGET": "EGFR",
85	"TARGET_PATHWAY": "EGFR signaling",
86	"PUBCHEM_ID": "176870",
87	"...": "..."
88	}
89	]
90	}
91	]
92	```
93
94	- Returns an empty list when no matches are found.
95	- Returns `{"error": "..."}` if the data directory is missing or empty.
96
97	### LLM Integration Example
98
99	```text
100	User: "What is the target of Erlotinib in GDSC?"
101	Agent: calls query_gdsc("Erlotinib")
102	→ source: screened_compounds_rel_8.4.csv, TARGET: EGFR, PATHWAY: EGFR signaling
103	→ "Erlotinib targets EGFR (EGFR signaling pathway) according to GDSC."
104	```

People who liked this also viewed

6 picks

Type