diffdock

Name: diffdock
Author: LeonChaoX

698stars61forks

How to install diffdock?

leonchaox/qinyan-academic-skills/diffdock

$npx -y skills add LeonChaoX/qinyan-academic-skills --skill diffdock

Installs into the current project.

Summary

Diffusion-based molecular docking. Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening, for structure-based drug design. Not for affinity prediction.

Files · 1

View on GitHub

SKILL.md

1	# DiffDock: Molecular Docking with Diffusion Models
2
3	## Overview
4
5	DiffDock is a diffusion-based deep learning tool for molecular docking that predicts 3D binding poses of small molecule ligands to protein targets. It represents the state-of-the-art in computational docking, crucial for structure-based drug discovery and chemical biology.
6
7	Core Capabilities:
8	- Predict ligand binding poses with high accuracy using deep learning
9	- Support protein structures (PDB files) or sequences (via ESMFold)
10	- Process single complexes or batch virtual screening campaigns
11	- Generate confidence scores to assess prediction reliability
12	- Handle diverse ligand inputs (SMILES, SDF, MOL2)
13
14	Key Distinction: DiffDock predicts binding poses (3D structure) and confidence (prediction certainty), NOT binding affinity (ΔG, Kd). Always combine with scoring functions (GNINA, MM/GBSA) for affinity assessment.
15
16	## When to Use This Skill
17
18	This skill should be used when:
19
20	- "Dock this ligand to a protein" or "predict binding pose"
21	- "Run molecular docking" or "perform protein-ligand docking"
22	- "Virtual screening" or "screen compound library"
23	- "Where does this molecule bind?" or "predict binding site"
24	- Structure-based drug design or lead optimization tasks
25	- Tasks involving PDB files + SMILES strings or ligand structures
26	- Batch docking of multiple protein-ligand pairs
27
28	## Installation and Environment Setup
29
30	### Check Environment Status
31
32	Before proceeding with DiffDock tasks, verify the environment setup:
33
34	```bash
35	# Use the provided setup checker
36	python scripts/setup_check.py
37	```
38
39	This script validates Python version, PyTorch with CUDA, PyTorch Geometric, RDKit, ESM, and other dependencies.
40
41	### Installation Options
42
43	Option 1: Conda (Recommended)
44	```bash
45	git clone https://github.com/gcorso/DiffDock.git
46	cd DiffDock
47	conda env create --file environment.yml
48	conda activate diffdock
49	```
50
51	Option 2: Docker
52	```bash
53	docker pull rbgcsail/diffdock
54	docker run -it --gpus all --entrypoint /bin/bash rbgcsail/diffdock
55	micromamba activate diffdock
56	```
57
58	Important Notes:
59	- GPU strongly recommended (10-100x speedup vs CPU)
60	- First run pre-computes SO(2)/SO(3) lookup tables (~2-5 minutes)
61	- Model checkpoints (~500MB) download automatically if not present
62
63	## Core Workflows
64
65	### Workflow 1: Single Protein-Ligand Docking
66
67	Use Case: Dock one ligand to one protein target
68
69	Input Requirements:
70	- Protein: PDB file OR amino acid sequence
71	- Ligand: SMILES string OR structure file (SDF/MOL2)
72
73	Command:
74	```bash
75	python -m inference \
76	--config default_inference_args.yaml \
77	--protein_path protein.pdb \
78	--ligand "CC(=O)Oc1ccccc1C(=O)O" \
79	--out_dir results/single_docking/
80	```
81
82	Alternative (protein sequence):
83	```bash
84	python -m inference \
85	--config default_inference_args.yaml \
86	--protein_sequence "MSKGEELFTGVVPILVELDGDVNGHKF..." \
87	--ligand ligand.sdf \
88	--out_dir results/sequence_docking/
89	```
90
91	Output Structure:
92	```
93	results/single_docking/
94	├── rank_1.sdf # Top-ranked pose
95	├── rank_2.sdf # Second-ranked pose
96	├── ...
97	├── rank_10.sdf # 10th pose (default: 10 samples)
98	└── confidence_scores.txt
99	```
100
101	### Workflow 2: Batch Processing Multiple Complexes
102
103	Use Case: Dock multiple ligands to proteins, virtual screening campaigns
104
105	Step 1: Prepare Batch CSV
106
107	Use the provided script to create or validate batch input:
108
109	```bash
110	# Create template
111	python scripts/prepare_batch_csv.py --create --output batch_input.csv
112
113	# Validate existing CSV
114	python scripts/prepare_batch_csv.py my_input.csv --validate
115	```
116
117	CSV Format:
118	```csv
119	complex_name,protein_path,ligand_description,protein_sequence
120	complex1,protein1.pdb,CC(=O)Oc1ccccc1C(=O)O,
121	complex2,,COc1ccc(C#N)cc1,MSKGEELFT...
122	complex3,protein3.pdb,ligand3.sdf,
123	```
124
125	Required Columns:
126	- `complex_name`: Unique identifier
127	- `protein_path`: PDB file path (leave empty if using sequence)
128	- `ligand_description`: SMILES string or ligand file path
129	- `protein_sequence`: Amino acid sequence (leave empty if using PDB)
130
131	Step 2: Run Batch Docking
132
133	```bash
134	python -m inference \
135	--config default_inference_args.yaml \
136	--protein_ligand_csv batch_input.csv \
137	--out_dir results/batch/ \
138	--batch_size 10
139	```
140
141	For Large Virtual Screening (>100 compounds):
142
143	Pre-compute protein embeddings for faster processing:
144	```bash
145	# Pre-compute embeddings
146	python datasets/esm_embedding_preparation.py \
147	--protein_ligand_csv screening_input.csv \
148	--out_file protein_embeddings.pt
149
150	# Run with pre-computed embeddings
151	python -m inference \
152	--config default_inference_args.yaml \
153	--protein_ligand_csv screening_input.csv \
154	--esm_embeddings_path protein_embeddings.pt \
155	--out_dir results/screening

People who liked this also viewed

6 picks

Type

find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 20262.6M27k
grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026633k182k
vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026573k29k
improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026522k182k
setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026454k182k
triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026424k182k

diffdock

698stars61forks

View on GitHub

How to install diffdock?

leonchaox/qinyan-academic-skills/diffdock

$npx -y skills add LeonChaoX/qinyan-academic-skills --skill diffdock

Installs into the current project.

Summary

Diffusion-based molecular docking. Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening, for structure-based drug design. Not for affinity prediction.

Files · 1

View on GitHub

SKILL.md

1	# DiffDock: Molecular Docking with Diffusion Models
2
3	## Overview
4
5	DiffDock is a diffusion-based deep learning tool for molecular docking that predicts 3D binding poses of small molecule ligands to protein targets. It represents the state-of-the-art in computational docking, crucial for structure-based drug discovery and chemical biology.
6
7	Core Capabilities:
8	- Predict ligand binding poses with high accuracy using deep learning
9	- Support protein structures (PDB files) or sequences (via ESMFold)
10	- Process single complexes or batch virtual screening campaigns
11	- Generate confidence scores to assess prediction reliability
12	- Handle diverse ligand inputs (SMILES, SDF, MOL2)
13
14	Key Distinction: DiffDock predicts binding poses (3D structure) and confidence (prediction certainty), NOT binding affinity (ΔG, Kd). Always combine with scoring functions (GNINA, MM/GBSA) for affinity assessment.
15
16	## When to Use This Skill
17
18	This skill should be used when:
19
20	- "Dock this ligand to a protein" or "predict binding pose"
21	- "Run molecular docking" or "perform protein-ligand docking"
22	- "Virtual screening" or "screen compound library"
23	- "Where does this molecule bind?" or "predict binding site"
24	- Structure-based drug design or lead optimization tasks
25	- Tasks involving PDB files + SMILES strings or ligand structures
26	- Batch docking of multiple protein-ligand pairs
27
28	## Installation and Environment Setup
29
30	### Check Environment Status
31
32	Before proceeding with DiffDock tasks, verify the environment setup:
33
34	```bash
35	# Use the provided setup checker
36	python scripts/setup_check.py
37	```
38
39	This script validates Python version, PyTorch with CUDA, PyTorch Geometric, RDKit, ESM, and other dependencies.
40
41	### Installation Options
42
43	Option 1: Conda (Recommended)
44	```bash
45	git clone https://github.com/gcorso/DiffDock.git
46	cd DiffDock
47	conda env create --file environment.yml
48	conda activate diffdock
49	```
50
51	Option 2: Docker
52	```bash
53	docker pull rbgcsail/diffdock
54	docker run -it --gpus all --entrypoint /bin/bash rbgcsail/diffdock
55	micromamba activate diffdock
56	```
57
58	Important Notes:
59	- GPU strongly recommended (10-100x speedup vs CPU)
60	- First run pre-computes SO(2)/SO(3) lookup tables (~2-5 minutes)
61	- Model checkpoints (~500MB) download automatically if not present
62
63	## Core Workflows
64
65	### Workflow 1: Single Protein-Ligand Docking
66
67	Use Case: Dock one ligand to one protein target
68
69	Input Requirements:
70	- Protein: PDB file OR amino acid sequence
71	- Ligand: SMILES string OR structure file (SDF/MOL2)
72
73	Command:
74	```bash
75	python -m inference \
76	--config default_inference_args.yaml \
77	--protein_path protein.pdb \
78	--ligand "CC(=O)Oc1ccccc1C(=O)O" \
79	--out_dir results/single_docking/
80	```
81
82	Alternative (protein sequence):
83	```bash
84	python -m inference \
85	--config default_inference_args.yaml \
86	--protein_sequence "MSKGEELFTGVVPILVELDGDVNGHKF..." \
87	--ligand ligand.sdf \
88	--out_dir results/sequence_docking/
89	```
90
91	Output Structure:
92	```
93	results/single_docking/
94	├── rank_1.sdf # Top-ranked pose
95	├── rank_2.sdf # Second-ranked pose
96	├── ...
97	├── rank_10.sdf # 10th pose (default: 10 samples)
98	└── confidence_scores.txt
99	```
100
101	### Workflow 2: Batch Processing Multiple Complexes
102
103	Use Case: Dock multiple ligands to proteins, virtual screening campaigns
104
105	Step 1: Prepare Batch CSV
106
107	Use the provided script to create or validate batch input:
108
109	```bash
110	# Create template
111	python scripts/prepare_batch_csv.py --create --output batch_input.csv
112
113	# Validate existing CSV
114	python scripts/prepare_batch_csv.py my_input.csv --validate
115	```
116
117	CSV Format:
118	```csv
119	complex_name,protein_path,ligand_description,protein_sequence
120	complex1,protein1.pdb,CC(=O)Oc1ccccc1C(=O)O,
121	complex2,,COc1ccc(C#N)cc1,MSKGEELFT...
122	complex3,protein3.pdb,ligand3.sdf,
123	```
124
125	Required Columns:
126	- `complex_name`: Unique identifier
127	- `protein_path`: PDB file path (leave empty if using sequence)
128	- `ligand_description`: SMILES string or ligand file path
129	- `protein_sequence`: Amino acid sequence (leave empty if using PDB)
130
131	Step 2: Run Batch Docking
132
133	```bash
134	python -m inference \
135	--config default_inference_args.yaml \
136	--protein_ligand_csv batch_input.csv \
137	--out_dir results/batch/ \
138	--batch_size 10
139	```
140
141	For Large Virtual Screening (>100 compounds):
142
143	Pre-compute protein embeddings for faster processing:
144	```bash
145	# Pre-compute embeddings
146	python datasets/esm_embedding_preparation.py \
147	--protein_ligand_csv screening_input.csv \
148	--out_file protein_embeddings.pt
149
150	# Run with pre-computed embeddings
151	python -m inference \
152	--config default_inference_args.yaml \
153	--protein_ligand_csv screening_input.csv \
154	--esm_embeddings_path protein_embeddings.pt \
155	--out_dir results/screening

People who liked this also viewed

6 picks

Type