{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Boltz-2 NIM Inference Example with Interactive Visualization\n", "Copyright (c) 2025, NVIDIA CORPORATION. Licensed under the Apache License, Version 2.0 (the \"License\") you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.\n", "\n", "This notebook demonstrates how to perform protein structure prediction using the NVIDIA BioNeMo Boltz-2 NIM and visualize the results interactively. For an example of a structure of a transcription factor and DNA complex refer to https://www.rcsb.org/structure/5GNJ.\n", "\n", "## Prerequisites\n", "You can host Boltz-2 NIM locally or use a NVIDIA-hosted NIM to run this workflow. Ensure the prerequisites in this section are met before running the workflow.\n", "Visit https://build.nvidia.com for instructions to run self-hosted or NVIDIA-hosted NIMs and system requirements for individual NIMs.\n", "\n", "### Launch the Boltz-2 NIM locally\n", "Execute the following code snippets in a bash terminal.\n", "```bash\n", "docker login nvcr.io\n", "Username: $oauthtoken\n", "Password: \n", "\n", "export NGC_API_KEY=\n", "export LOCAL_NIM_CACHE=~/.cache/nim\n", "mkdir -p $LOCAL_NIM_CACHE\n", "\n", "docker run -it \\\n", " --runtime=nvidia \\\n", " -p 8000:8000 \\\n", " -e NGC_API_KEY \\\n", " -v \"$LOCAL_NIM_CACHE\":/opt/nim/.cache \\\n", " nvcr.io/nim/mit/boltz2:1.0.0\n", "```\n", "\n", "--- \n", "\n", "### API Information\n", "- **Local Endpoint**: `http://localhost:8000/biology/mit/boltz2/predict`\n", "- **Documentation**: `http://localhost:8000/docs`\n", "- **Output Format**: mmCIF (macromolecular Crystallographic Information File)\n", "\n", "### Other Requirements\n", "- `httpx` for async HTTP requests (will be auto-installed)\n", "- `py3Dmol` for interactive 3D visualization (will be auto-installed)\n", "- Local Boltz-2 NIM running on port 8000" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Setup and Imports" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "āœ… httpx is already installed\n", "āœ… py3Dmol is already installed\n", "\n", "šŸŽ‰ All packages are ready!\n" ] } ], "source": [ "# Install required packages if not available\n", "import subprocess\n", "import sys\n", "\n", "def install_package(package):\n", " try:\n", " __import__(package)\n", " print(f\"āœ… {package} is already installed\")\n", " except ImportError:\n", " print(f\"šŸ“¦ Installing {package}...\")\n", " subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", package])\n", " print(f\"āœ… {package} installed successfully\")\n", "\n", "# Install required packages\n", "install_package(\"httpx\")\n", "install_package(\"py3Dmol\")\n", "\n", "print(\"\\nšŸŽ‰ All packages are ready!\")" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "All imports successful!\n" ] } ], "source": [ "import asyncio\n", "import json\n", "import os\n", "import time\n", "from pathlib import Path\n", "from datetime import datetime\n", "from typing import Dict, Any, Optional, List\n", "import httpx\n", "import py3Dmol\n", "from IPython.display import display, HTML\n", "\n", "print(\"All imports successful!\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configuration" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Boltz-2 Endpoint: http://localhost:8000/biology/mit/boltz2/predict\n", "Health Check: http://localhost:8000/v1/health/live\n" ] } ], "source": [ "# Local Boltz-2 NIM endpoint\n", "BOLTZ2_URL = \"http://localhost:8000/biology/mit/boltz2/predict\"\n", "HEALTH_URL = \"http://localhost:8000/v1/health/live\"\n", "\n", "print(f\"Boltz-2 Endpoint: {BOLTZ2_URL}\")\n", "print(f\"Health Check: {HEALTH_URL}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Health Check\n", "\n", "Let's first verify the NIM is running and accessible:" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "āœ… Boltz-2 NIM is running and accessible\n" ] } ], "source": [ "async def check_nim_health():\n", " \"\"\"Check if the Boltz-2 NIM is running and accessible.\"\"\"\n", " try:\n", " async with httpx.AsyncClient(timeout=10.0) as client:\n", " response = await client.get(HEALTH_URL)\n", " if response.status_code == 200:\n", " print(\"āœ… Boltz-2 NIM is running and accessible\")\n", " return True\n", " else:\n", " print(f\"āš ļø Health check returned status {response.status_code}\")\n", " return False\n", " except Exception as e:\n", " print(f\"āŒ Cannot connect to Boltz-2 NIM: {e}\")\n", " return False\n", "\n", "# Check NIM health\n", "nim_healthy = await check_nim_health()" ] }, { "cell_type": "markdown", "metadata": { "jp-MarkdownHeadingCollapsed": true }, "source": [ "### API Client Functions" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "API client functions defined successfully!\n" ] } ], "source": [ "async def make_boltz2_prediction(request_data: Dict[str, Any], timeout: int = 300) -> Optional[Dict]:\n", " \"\"\"\n", " Make a prediction request to the local Boltz-2 NIM.\n", " \n", " Args:\n", " request_data: The prediction request payload\n", " timeout: Request timeout in seconds\n", " \n", " Returns:\n", " Response data or None if failed\n", " \"\"\"\n", " headers = {\n", " \"Content-Type\": \"application/json\"\n", " }\n", " \n", " async with httpx.AsyncClient(timeout=timeout) as client:\n", " print(f\"šŸš€ Making prediction request to {BOLTZ2_URL}\")\n", " print(f\"ā±ļø Timeout set to {timeout} seconds\")\n", " \n", " try:\n", " start_time = time.time()\n", " \n", " response = await client.post(BOLTZ2_URL, json=request_data, headers=headers)\n", " \n", " end_time = time.time()\n", " duration = end_time - start_time\n", " \n", " print(f\"šŸ“” Response received in {duration:.2f} seconds\")\n", " print(f\"šŸ“Š Status code: {response.status_code}\")\n", " \n", " if response.status_code == 200:\n", " print(\"āœ… Prediction successful!\")\n", " return response.json()\n", " else:\n", " print(f\"āŒ Prediction failed: {response.status_code}\")\n", " print(f\"Error details: {response.text}\")\n", " return None\n", " \n", " except httpx.TimeoutException:\n", " print(f\"ā° Request timed out after {timeout} seconds\")\n", " return None\n", " except Exception as e:\n", " print(f\"āŒ Request failed: {e}\")\n", " return None\n", "\n", "print(\"API client functions defined successfully!\")" ] }, { "cell_type": "markdown", "metadata": { "jp-MarkdownHeadingCollapsed": true }, "source": [ "### 3D Visualization Functions" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "3D visualization functions defined successfully!\n" ] } ], "source": [ "def visualize_structure(structure_data: str, title: str = \"Protein Structure\", \n", " width: int = 800, height: int = 600, \n", " style: str = \"cartoon\", color_scheme: str = \"spectrum\"):\n", " \"\"\"\n", " Visualize a protein structure using py3Dmol.\n", " \n", " Args:\n", " structure_data: mmCIF structure data as string\n", " title: Title for the visualization\n", " width: Viewer width in pixels\n", " height: Viewer height in pixels\n", " style: Visualization style ('cartoon', 'stick', 'sphere', 'line')\n", " color_scheme: Color scheme ('spectrum', 'chain', 'residue', 'atom')\n", " \n", " Returns:\n", " py3Dmol viewer object\n", " \"\"\"\n", " # Create viewer\n", " viewer = py3Dmol.view(width=width, height=height)\n", " \n", " # Add structure\n", " viewer.addModel(structure_data, 'cif')\n", " \n", " # Set style based on parameters\n", " if style == \"cartoon\":\n", " viewer.setStyle({'cartoon': {'color': color_scheme}})\n", " elif style == \"stick\":\n", " viewer.setStyle({'stick': {'color': color_scheme}})\n", " elif style == \"sphere\":\n", " viewer.setStyle({'sphere': {'color': color_scheme}})\n", " elif style == \"line\":\n", " viewer.setStyle({'line': {'color': color_scheme}})\n", " else:\n", " # Default to cartoon\n", " viewer.setStyle({'cartoon': {'color': color_scheme}})\n", " \n", " # Center and zoom\n", " viewer.zoomTo()\n", " \n", " # Add title\n", " display(HTML(f\"

{title}

\"))\n", " \n", " return viewer\n", "\n", "def visualize_protein_ligand_complex(structure_data: str, title: str = \"Protein-Ligand Complex\",\n", " width: int = 800, height: int = 600):\n", " \"\"\"\n", " Specialized visualization for protein-ligand complexes.\n", " \n", " Args:\n", " structure_data: mmCIF structure data as string\n", " title: Title for the visualization\n", " width: Viewer width in pixels\n", " height: Viewer height in pixels\n", " \n", " Returns:\n", " py3Dmol viewer object\n", " \"\"\"\n", " viewer = py3Dmol.view(width=width, height=height)\n", " \n", " # Add structure\n", " viewer.addModel(structure_data, 'cif')\n", " \n", " # Style protein as cartoon\n", " viewer.setStyle({'and': [{'resn': ['ALA', 'ARG', 'ASN', 'ASP', 'CYS', 'GLN', 'GLU', 'GLY', 'HIS', 'ILE', \n", " 'LEU', 'LYS', 'MET', 'PHE', 'PRO', 'SER', 'THR', 'TRP', 'TYR', 'VAL']}]}, \n", " {'cartoon': {'color': 'spectrum'}})\n", " \n", " # Style ligands as sticks\n", " viewer.setStyle({'and': [{'not': {'resn': ['ALA', 'ARG', 'ASN', 'ASP', 'CYS', 'GLN', 'GLU', 'GLY', 'HIS', 'ILE', \n", " 'LEU', 'LYS', 'MET', 'PHE', 'PRO', 'SER', 'THR', 'TRP', 'TYR', 'VAL']}}]}, \n", " {'stick': {'color': 'red', 'radius': 0.3}})\n", " \n", " # Center and zoom\n", " viewer.zoomTo()\n", " \n", " # Add title\n", " display(HTML(f\"

{title}

\"))\n", " \n", " return viewer\n", "\n", "def create_multi_view_visualization(structures: List[str], titles: List[str] = None,\n", " width: int = 400, height: int = 400):\n", " \"\"\"\n", " Create side-by-side visualization of multiple structures.\n", " \n", " Args:\n", " structures: List of mmCIF structure data strings\n", " titles: List of titles for each structure\n", " width: Width of each viewer\n", " height: Height of each viewer\n", " \n", " Returns:\n", " List of py3Dmol viewer objects\n", " \"\"\"\n", " if titles is None:\n", " titles = [f\"Structure {i+1}\" for i in range(len(structures))]\n", " \n", " viewers = []\n", " \n", " # Create HTML for side-by-side layout\n", " html_content = \"
\"\n", " \n", " for i, (structure, title) in enumerate(zip(structures, titles)):\n", " viewer = py3Dmol.view(width=width, height=height)\n", " viewer.addModel(structure, 'cif')\n", " viewer.setStyle({'cartoon': {'color': 'spectrum'}})\n", " viewer.zoomTo()\n", " \n", " viewers.append(viewer)\n", " \n", " # Add title for each structure\n", " display(HTML(f\"

{title}

\"))\n", " viewer.show()\n", " \n", " return viewers\n", "\n", "print(\"3D visualization functions defined successfully!\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Prepare Request Data\n", "\n", "Based on the API schema, here are the key parameters:\n", "\n", "**Required**:\n", "- `polymers`: List of polymers (DNA, RNA, or Protein) - max 5\n", "\n", "**Optional**:\n", "- `ligands`: List of ligands - max 5\n", "- `constraints`: Pocket or bond constraints\n", "- `recycling_steps`: 1-10 (default: 3)\n", "- `sampling_steps`: 10-1000 (default: 50)\n", "- `diffusion_samples`: 1-5 (default: 1)\n", "- `step_scale`: 0.5-5.0 (default: 1.638)\n", "- `without_potentials`: boolean (default: false)\n", "- `output_format`: \"mmcif\" (default)\n", "- `concatenate_msas`: boolean (default: false)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "šŸ“‹ Request Summary:\n", " Recycling steps: 3\n", " Sampling steps: 50\n", " Diffusion samples: 1\n", " Step scale: 1.638\n", " Output format: mmcif\n" ] } ], "source": [ "# Example protein sequence\n", "# sequence = \"MKTVRQERLKSIVRILERSKEPVSGAQLAEELSVSRQVIVQDIAYLRSLGYNIVATPRGYVLAGG\"\n", "\n", "# Prepare the request payload according to the API schema\n", "request_data = {\n", " \"polymers\": [\n", " {\n", " \"id\": \"A\",\n", " \"molecule_type\": \"protein\", \n", " \"sequence\": \"MGREEPLNHVEAERQRREKLNQRFYALRAVVPNVSKMDKASLLGDAIAYINELKSKVVKTESEKLQIKNQLEEVKLELAGRLEHHHHHH\",\n", " \"cyclic\": False,\n", " \"modifications\": [] # No modifications\n", " },\n", " {\n", " \"id\": \"B\",\n", " \"molecule_type\": \"protein\", \n", " \"sequence\": \"MGREEPLNHVEAERQRREKLNQRFYALRAVVPNVSKMDKASLLGDAIAYINELKSKVVKTESEKLQIKNQLEEVKLELAGRLEHHHHHH\",\n", " \"cyclic\": False,\n", " \"modifications\": [] # No modifications\n", " },\n", " {\n", " \"id\": \"C\",\n", " \"molecule_type\": \"dna\", \n", " \"sequence\": \"TGGGTCACGTGTTCC\",\n", " \"cyclic\": False,\n", " \"modifications\": [] # No modifications\n", " },\n", " {\n", " \"id\": \"D\",\n", " \"molecule_type\": \"dna\", \n", " \"sequence\": \"AGGAACACGTGACCC\",\n", " \"cyclic\": False,\n", " \"modifications\": [] # No modifications\n", " }\n", " ],\n", " \"constraints\": [], # No constraints\n", " \"recycling_steps\": 3, # Default value\n", " \"sampling_steps\": 50, # Default value\n", " \"diffusion_samples\": 1, # Default value\n", " \"step_scale\": 1.638, # Default value\n", " \"without_potentials\": False, # Include potentials\n", " \"output_format\": \"mmcif\", # mmCIF format\n", " \"concatenate_msas\": False # Don't concatenate MSAs\n", "}\n", "\n", "print(f\"šŸ“‹ Request Summary:\")\n", "# print(f\" Protein sequence length: {len(sequence)} amino acids\")\n", "# print(f\" Ligand: {request_data['ligands'][0]['smiles']} (Aspirin)\")\n", "print(f\" Recycling steps: {request_data['recycling_steps']}\")\n", "print(f\" Sampling steps: {request_data['sampling_steps']}\")\n", "print(f\" Diffusion samples: {request_data['diffusion_samples']}\")\n", "print(f\" Step scale: {request_data['step_scale']}\")\n", "print(f\" Output format: {request_data['output_format']}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Make the Prediction Request" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "šŸŽÆ Starting Boltz-2 prediction at 2025-06-05 11:38:15.259851\n", "šŸš€ Making prediction request to http://localhost:8000/biology/mit/boltz2/predict\n", "ā±ļø Timeout set to 600 seconds\n", "šŸ“” Response received in 6.35 seconds\n", "šŸ“Š Status code: 200\n", "āœ… Prediction successful!\n", "šŸ’¾ Results saved to: boltz2_prediction_20250605_113821.json\n" ] } ], "source": [ "# Only proceed if NIM is healthy\n", "if nim_healthy:\n", " print(f\"šŸŽÆ Starting Boltz-2 prediction at {datetime.now()}\")\n", " # print(f\"šŸ§¬ Sequence: {sequence[:30]}...{sequence[-10:]}\")\n", " \n", " # Make the prediction\n", " prediction_result = await make_boltz2_prediction(request_data, timeout=600) # 10 minute timeout\n", " \n", " if prediction_result:\n", " # Save the result\n", " timestamp = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n", " output_file = f\"boltz2_prediction_{timestamp}.json\"\n", " \n", " with open(output_file, 'w') as f:\n", " json.dump(prediction_result, f, indent=2)\n", " \n", " print(f\"šŸ’¾ Results saved to: {output_file}\")\n", " else:\n", " print(\"āŒ Prediction failed\")\n", " prediction_result = None\n", "else:\n", " print(\"āŒ Cannot proceed - NIM is not accessible\")\n", " prediction_result = None" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Analyze Results and Create Visualizations" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "šŸ”¬ === PREDICTION RESULTS ANALYSIS ===\n", "šŸ“Š Available data fields: ['structures', 'metrics', 'confidence_scores']\n", "\n", "šŸ—ļø Structure Analysis:\n", " Number of structures: 1\n", "\n", " Structure 1:\n", " Format: mmcif\n", " Size: 173710 characters\n", " Name: \n", " Source: oJeTY8_model_0.cif\n", "\n", "šŸŽÆ Confidence Analysis:\n", " Number of scores: 1\n", " Average confidence: 0.949\n", " Min confidence: 0.949\n", " Max confidence: 0.949\n", " All scores: ['0.949']\n", "šŸ’¾ Structure 1 saved to: boltz2_structure_1_20250605_113821.cif\n" ] } ], "source": [ "if prediction_result:\n", " print(\"\\nšŸ”¬ === PREDICTION RESULTS ANALYSIS ===\")\n", " \n", " # Show available keys in the response\n", " print(f\"šŸ“Š Available data fields: {list(prediction_result.keys())}\")\n", " \n", " # Analyze structures\n", " if 'structures' in prediction_result:\n", " structures = prediction_result['structures']\n", " print(f\"\\nšŸ—ļø Structure Analysis:\")\n", " print(f\" Number of structures: {len(structures)}\")\n", " \n", " for i, structure in enumerate(structures):\n", " print(f\"\\n Structure {i+1}:\")\n", " print(f\" Format: {structure.get('format', 'Unknown')}\")\n", " print(f\" Size: {len(structure.get('structure', ''))} characters\")\n", " print(f\" Name: {structure.get('name', 'Unnamed')}\")\n", " print(f\" Source: {structure.get('source', 'Unknown')}\")\n", " \n", " # Analyze confidence scores\n", " if 'confidence_scores' in prediction_result:\n", " scores = prediction_result['confidence_scores']\n", " print(f\"\\nšŸŽÆ Confidence Analysis:\")\n", " print(f\" Number of scores: {len(scores)}\")\n", " \n", " if scores:\n", " print(f\" Average confidence: {sum(scores)/len(scores):.3f}\")\n", " print(f\" Min confidence: {min(scores):.3f}\")\n", " print(f\" Max confidence: {max(scores):.3f}\")\n", " print(f\" All scores: {[f'{s:.3f}' for s in scores]}\")\n", " \n", " # Show metrics if available\n", " if 'metrics' in prediction_result and prediction_result['metrics']:\n", " print(f\"\\nšŸ“ˆ Runtime Metrics:\")\n", " for key, value in prediction_result['metrics'].items():\n", " print(f\" {key}: {value}\")\n", " \n", " # Save individual structures\n", " structure_files = []\n", " if 'structures' in prediction_result:\n", " for i, structure in enumerate(prediction_result['structures']):\n", " if structure.get('format') == 'mmcif':\n", " structure_file = f\"boltz2_structure_{i+1}_{timestamp}.cif\"\n", " with open(structure_file, 'w') as f:\n", " f.write(structure['structure'])\n", " structure_files.append(structure_file)\n", " print(f\"šŸ’¾ Structure {i+1} saved to: {structure_file}\")\n", " \n", "else:\n", " print(\"āŒ No results to analyze - prediction failed or was not attempted.\")\n", " structure_files = []" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### šŸŽØ Interactive 3D Visualization\n", "\n", "Now let's visualize the predicted structures interactively!" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "šŸŽØ Creating interactive 3D visualizations...\n", "\n", "\n", "šŸ“ Structure 1 Visualization:\n" ] }, { "data": { "text/html": [ "

Boltz-2 Predicted Structure 1 (Confidence: 0.949)

" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "application/3dmoljs_load.v0": "
\n

3Dmol.js failed to load for some reason. Please check your browser console for error messages.

\n
\n", "text/html": [ "
\n", "

3Dmol.js failed to load for some reason. Please check your browser console for error messages.

\n", "
\n", "" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "šŸŽ›ļø Interactive controls:\n", " ā€¢ Mouse: Rotate structure\n", " ā€¢ Scroll: Zoom in/out\n", " ā€¢ Right-click + drag: Pan\n" ] } ], "source": [ "if prediction_result and 'structures' in prediction_result:\n", " structures = prediction_result['structures']\n", " \n", " print(\"šŸŽØ Creating interactive 3D visualizations...\\n\")\n", " \n", " # Visualize each structure\n", " for i, structure in enumerate(structures):\n", " if structure.get('format') == 'mmcif':\n", " structure_data = structure['structure']\n", " \n", " # Get confidence score for this structure if available\n", " confidence_info = \"\"\n", " if 'confidence_scores' in prediction_result and i < len(prediction_result['confidence_scores']):\n", " confidence = prediction_result['confidence_scores'][i]\n", " confidence_info = f\" (Confidence: {confidence:.3f})\"\n", " \n", " title = f\"Boltz-2 Predicted Structure {i+1}{confidence_info}\"\n", " \n", " # Create main visualization\n", " print(f\"\\nšŸ“ Structure {i+1} Visualization:\")\n", " \n", " # Check if we have ligands in the request\n", " has_ligands = 'ligands' in request_data and len(request_data['ligands']) > 0\n", " \n", " if has_ligands:\n", " # Use protein-ligand complex visualization\n", " viewer = visualize_protein_ligand_complex(\n", " structure_data, \n", " title=title + \" - Protein-Ligand Complex\",\n", " width=900, \n", " height=600\n", " )\n", " else:\n", " # Use standard protein visualization\n", " viewer = visualize_structure(\n", " structure_data, \n", " title=title,\n", " width=900, \n", " height=600,\n", " style=\"cartoon\",\n", " color_scheme=\"spectrum\"\n", " )\n", " \n", " viewer.show()\n", " \n", " print(f\"\\nšŸŽ›ļø Interactive controls:\")\n", " print(f\" ā€¢ Mouse: Rotate structure\")\n", " print(f\" ā€¢ Scroll: Zoom in/out\")\n", " print(f\" ā€¢ Right-click + drag: Pan\")\n", " \n", "else:\n", " print(\"āŒ No structures available for visualization\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### šŸ’¾ Export Visualization\n", "\n", "Save visualizations as images (requires additional setup):" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "šŸ“ø To save visualizations as images:\n", "\n", "1. Right-click on any 3D visualization above\n", "2. Select 'Save image as...' or 'Copy image'\n", "3. Choose your desired location and format\n", "\n", "Alternatively, you can:\n", "ā€¢ Take screenshots of the visualizations\n", "ā€¢ Use browser developer tools to export canvas\n", "ā€¢ Load the .cif files in external software like PyMOL or ChimeraX\n", "\n", "šŸ“ Structure files saved for external visualization:\n", " ā€¢ boltz2_structure_1_20250605_113821.cif\n", "\n", "These can be opened in:\n", " ā€¢ PyMOL: pymol structure_file.cif\n", " ā€¢ ChimeraX: open structure_file.cif\n", " ā€¢ VMD: vmd structure_file.cif\n" ] } ], "source": [ "# Note: Image export requires additional browser setup\n", "# This cell provides instructions for manual export\n", "\n", "print(\"šŸ“ø To save visualizations as images:\")\n", "print(\"\\n1. Right-click on any 3D visualization above\")\n", "print(\"2. Select 'Save image as...' or 'Copy image'\")\n", "print(\"3. Choose your desired location and format\")\n", "print(\"\\nAlternatively, you can:\")\n", "print(\"ā€¢ Take screenshots of the visualizations\")\n", "print(\"ā€¢ Use browser developer tools to export canvas\")\n", "print(\"ā€¢ Load the .cif files in external software like PyMOL or ChimeraX\")\n", "\n", "if structure_files:\n", " print(f\"\\nšŸ“ Structure files saved for external visualization:\")\n", " for file in structure_files:\n", " print(f\" ā€¢ {file}\")\n", " print(\"\\nThese can be opened in:\")\n", " print(\" ā€¢ PyMOL: pymol structure_file.cif\")\n", " print(\" ā€¢ ChimeraX: open structure_file.cif\")\n", " print(\" ā€¢ VMD: vmd structure_file.cif\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Summary\n", "\n", "This notebook demonstrates comprehensive Boltz-2 NIM usage with interactive visualization:\n", "\n", "### āœ… **Key Features:**\n", "1. **Local NIM Integration** - Direct connection to your local Boltz-2 instance\n", "2. **Health Checking** - Verifies NIM availability before making requests\n", "3. **Interactive 3D Visualization** - py3Dmol integration for structure viewing\n", "4. **Multiple Visualization Styles** - Cartoon, stick, sphere, and line representations\n", "5. **Biomolecular Complex Prediction** - Specialized visualization for complexes\n", "6. **Confidence Score Analysis** - Visual and statistical confidence assessment\n", "7. **File Output** - Saves both JSON results and individual mmCIF structure files\n", "8. **Parameter Flexibility** - Easy to adjust prediction quality vs. speed\n", "\n", "### ļæ½ļæ½ **API Parameters:**\n", "- **recycling_steps**: 1-10 (affects accuracy, default: 3)\n", "- **sampling_steps**: 10-1000 (affects quality, default: 50)\n", "- **diffusion_samples**: 1-5 (multiple predictions, default: 1)\n", "- **step_scale**: 0.5-5.0 (temperature, default: 1.638)\n", "\n", "### šŸ“ **Output Files:**\n", "- `boltz2_prediction_YYYYMMDD_HHMMSS.json` - Complete API response\n", "- `boltz2_structure_N_YYYYMMDD_HHMMSS.cif` - Individual structure files\n", "\n", "### šŸš€ **Next Steps:**\n", "1. Experiment with different protein sequences\n", "2. Try various ligands using SMILES notation\n", "3. Adjust parameters for your speed/quality needs\n", "4. Export structures for external analysis\n", "5. Compare multiple predictions side-by-side\n", "6. Analyze confidence scores for structure quality assessment" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.14" } }, "nbformat": 4, "nbformat_minor": 4 }