13  Demo: Protein Structure Visualization with PyMol

In this demo, we’ll create publication-quality visualizations of a protein structure—without prior PyMol knowledge. This demonstrates using Claude Code to learn and apply unfamiliar tools.

13.1 The Goal

We want to visualize the binding site of a protein-ligand complex and create a figure suitable for a paper or presentation. We’ll learn PyMol along the way.

13.2 Prerequisites

• PyMol installed (we’ll cover installation)
• Internet connection (to download structures)
• Basic understanding of proteins (helpful but not required)

13.3 Project Setup

mkdir ~/Projects/pymol-visualization
cd ~/Projects/pymol-visualization
claude

> Set up a project for protein structure visualization. I'll need to download PDB files and generate PyMol scripts.

13.4 Phase 1: Installing PyMol

First, let’s get PyMol installed:

> How do I install PyMol on my system? I'm on [Mac/Windows/Linux]

Mac (with Homebrew):

brew install pymol

Mac (alternative): Download from pymol.org (educational license available)

Linux (Ubuntu):

sudo apt-get install pymol

Windows: Download installer from pymol.org

Verify installation:

> Check if pymol is installed and working

pymol -c -q  # Should exit without errors

13.5 Phase 2: Choosing a Structure

Let’s pick an interesting protein. For this demo, we’ll use Human Carbonic Anhydrase II bound to a drug—it’s well-studied and has a clear binding site.

> Find a PDB structure for Human Carbonic Anhydrase II bound to a small molecule inhibitor

Claude might suggest PDB ID 1BN1 (CA-II with acetazolamide) or similar.

> Download the PDB file for 1BN1 to our data folder

import urllib.request
import os

os.makedirs('data', exist_ok=True)
url = "https://files.rcsb.org/download/1BN1.pdb"
urllib.request.urlretrieve(url, "data/1BN1.pdb")

13.6 Phase 3: First Visualization

Now let’s create our first image. We’ll ask Claude to help with PyMol commands.

> Create a PyMol script that:
> 1. Loads 1BN1.pdb
> 2. Colors the protein in a nice blue color
> 3. Shows the ligand (acetazolamide) as sticks in yellow
> 4. Saves an image

Claude generates scripts/basic_view.pml:

# Load structure
load data/1BN1.pdb

# Color protein
color marine, polymer

# Show ligand as sticks
select ligand, organic
show sticks, ligand
color yellow, ligand

# Clean up display
hide lines
show cartoon, polymer
set cartoon_fancy_helices, 1

# Set background
bg_color white

# Ray trace for high quality
ray 1200, 1200

# Save image
png results/basic_view.png

Run it:

> Run this PyMol script

pymol -c scripts/basic_view.pml  # -c for command line mode

13.7 Phase 4: Visualizing the Binding Site

The basic view shows the whole protein. Let’s focus on the binding site:

> Modify the script to:
> 1. Zoom in on the binding site (within 5 Angstroms of the ligand)
> 2. Show residues near the ligand as sticks
> 3. Color binding site residues by element (carbon gray, oxygen red, nitrogen blue)

Updated script:

# Load and basic setup
load data/1BN1.pdb
bg_color white

# Color protein
color marine, polymer
show cartoon, polymer

# Select and show ligand
select ligand, organic
show sticks, ligand
color yellow, ligand and elem C
util.cnc ligand  # Color by element except carbon

# Select binding site residues (within 5A of ligand)
select binding_site, byres (polymer within 5 of ligand)
show sticks, binding_site
color gray, binding_site and elem C
util.cnc binding_site

# Zoom to binding site
zoom binding_site

# Add surface representation (transparent)
show surface, binding_site
set transparency, 0.7, binding_site
color white, binding_site and surface

# High quality render
set ray_shadow, 0
set antialias, 2
ray 1600, 1600

png results/binding_site.png

13.8 Phase 5: Showing Interactions

Let’s visualize the key interactions between the ligand and protein:

> Add visualization of hydrogen bonds between the ligand and nearby residues

# Find and display hydrogen bonds
distance hbonds, ligand, binding_site, mode=2
set dash_color, black, hbonds
set dash_width, 2

> Also show the zinc ion that's important for catalysis, as a sphere

# Show zinc
select zinc, elem Zn
show spheres, zinc
color gray50, zinc
set sphere_scale, 0.4, zinc

13.9 Phase 6: Publication Quality

Let’s refine for publication:

> Make these changes for publication quality:
> 1. Increase overall size and resolution
> 2. Add smooth lighting
> 3. Remove the PyMol axes
> 4. Use a cleaner font for any labels

# Publication settings
set ray_trace_mode, 1
set ray_trace_frames, 1
set ray_shadows, 0
set orthoscopic, on

# Remove axes
set axes, 0

# Smooth appearance
set antialias, 2
set spec_reflect, 0.5
set spec_power, 200

# High resolution output
ray 2400, 2400
png results/figure_1_binding_site.png, dpi=300

13.10 Phase 7: Multiple Views

For a paper, we often need multiple perspectives:

> Create a script that generates 3 views:
> 1. Overview of the full protein with binding site highlighted
> 2. Close-up of binding site
> 3. Rotated view showing the depth of the binding pocket

Claude creates a comprehensive script that: - Sets up a common base configuration - Saves the session - Generates each view with appropriate rotation and zoom - Outputs as both PNG and PDF

13.11 Phase 8: Making It Reproducible

> Create a main.py script that:
> 1. Downloads the PDB if not present
> 2. Runs all PyMol scripts
> 3. Generates all figures
> 4. Works from the command line

#!/usr/bin/env python
"""Generate protein structure figures."""

import os
import subprocess
import urllib.request

PDB_ID = "1BN1"

def download_pdb():
    """Download PDB file if not present."""
    path = f"data/{PDB_ID}.pdb"
    if not os.path.exists(path):
        os.makedirs("data", exist_ok=True)
        url = f"https://files.rcsb.org/download/{PDB_ID}.pdb"
        print(f"Downloading {PDB_ID}...")
        urllib.request.urlretrieve(url, path)
    return path

def run_pymol_script(script):
    """Run a PyMol script in command-line mode."""
    result = subprocess.run(
        ["pymol", "-c", "-q", script],
        capture_output=True,
        text=True
    )
    if result.returncode != 0:
        print(f"Error running {script}:")
        print(result.stderr)
    return result.returncode == 0

def main():
    os.makedirs("results", exist_ok=True)

    # Download structure
    download_pdb()

    # Generate figures
    scripts = [
        "scripts/overview.pml",
        "scripts/binding_site.pml",
        "scripts/rotated_view.pml",
    ]

    for script in scripts:
        print(f"Running {script}...")
        run_pymol_script(script)

    print("Done! Figures saved to results/")

if __name__ == "__main__":
    main()

13.12 Learning PyMol Along the Way

Throughout this demo, Claude has been teaching us PyMol:

13.12.1 Key Concepts Learned

Concept	What We Learned
Selections	select name, criteria to create named groups
Representations	cartoon, sticks, surface, spheres
Coloring	color, util.cnc for element colors
Distance	distance for measuring and showing bonds
Views	zoom, orient, turn, move
Ray tracing	ray for publication-quality rendering

13.12.2 Asking for Help

When stuck, ask Claude:

> I want the protein backbone to be semi-transparent. How do I do that in PyMol?

> What's the difference between 'show' and 'set' in PyMol?

13.13 Exploring Further

> What other cool visualizations could we do with this structure?

Ideas: - Electrostatic surface potential - Hydrophobicity mapping - Animation of binding - Multiple structures superimposed

13.14 Final Cleanup

> Create a README that documents:
> 1. What structures we visualized
> 2. How to run the scripts
> 3. What each output figure shows

> Commit everything with a descriptive message

13.15 What You’ve Learned

By completing this demo, you’ve:

• Installed and used PyMol without prior experience
• Learned PyMol syntax through practical application
• Created publication-quality figures programmatically
• Built a reproducible pipeline for structure visualization

13.16 Applying This to Your Work

Try with your own protein of interest:

> Help me visualize [your protein]. I want to show [specific feature].

Claude will adapt the approach to your structure.

13.17 Next Steps

Try another demo or continue to Part 4: Paper Writing Track.