How to Visualize Protein-Ligand Interactions in PyMOL: Binding Pockets, H-bonds and Publication Figures
Protein-ligand interaction figures are among the most published images in structural biology and drug discovery. This tutorial covers the complete workflow: loading the complex, showing the binding pocket, displaying H-bonds and contacts, measuring key distances, and producing a clean publication figure that communicates the binding mode clearly.
Interaction types — what PyMOL can show
Before opening PyMOL, it helps to know which interaction types you’re looking for. The six major categories in protein-ligand binding, and how PyMOL handles each:
Loading and preparing the complex
# Load from RCSB (structure must contain bound ligand)
fetch 4XYZ
# Or load a docking result / local file
load protein_ligand_complex.pdb
# Remove water and other heteroatoms that aren't the ligand
remove resn HOH
remove resn SO4+PEG+GOL+EDO # common crystallographic additives to remove
# Identify the ligand residue name
select ligands, hetatm
iterate ligands, print(resn, resi, chain)
# Name your ligand selection (replace LIG with actual 3-letter code)
select LIG, resn LIGiterate hetatm, print(resn) in the command line, which prints all heteroatom residue names. In the PDB, crystal additives like sulfate (SO4), glycerol (GOL), and PEG appear as hetatm alongside your actual ligand — remove them first before working on the binding site.
Showing the binding pocket
A clean binding site figure uses three layers: the gray protein context (cartoon or surface), the pocket residues shown as sticks, and the ligand displayed prominently. Build these layers in order.
# Start clean
as cartoon
hide everything, resn HOH
# Select binding site residues (all within 5 Å of ligand)
select pocket, byres (LIG around 5)
# Show pocket residues as sticks
show sticks, pocket
# Color protein context neutral gray
color gray70, polymer
# Show ligand prominently
show sticks, LIG
# Optional: surface of binding pocket only
show surface, pocket
set transparency, 0.4
color white, pocket # white surface lets colored sticks show through
# Zoom in on the binding site
zoom LIG, 5 # zoom to 5 Å buffer around ligandColoring the ligand
Ligand coloring is one of the most important visual choices in a binding site figure. The standard is CPK colors with a carbon color that contrasts the protein. Green carbons are the universal convention for ligands in structural biology publications:
# Standard: green carbons, CPK for all other atoms
util.cbag LIG
# Alternative carbon colors (when showing multiple ligands)
util.cbac LIG # cyan carbons — good for second molecule
util.cbay LIG # yellow carbons
util.cbam LIG # magenta carbons
# Color protein binding site residues differently from context
color gray70, polymer # neutral gray for whole protein
color slate, pocket # slightly blue for pocket residues
util.cbag LIG # green carbons for ligand
# Make ligand slightly larger sticks to distinguish from protein
set stick_radius, 0.25, LIG # default 0.25, protein sticks thinner
set stick_radius, 0.15, pocket # thinner protein sticksDisplaying H-bonds
Hydrogen bonds are the most important interactions to show in a binding site figure. PyMOL’s distance command with mode=2 identifies H-bonds automatically by checking donor–acceptor geometry and distance.
# Show all H-bonds between ligand and protein
distance hbonds, LIG, pocket, mode=2
# mode=2 uses PyMOL's built-in H-bond detection
# Shows dashed lines between H-bond donor/acceptor pairs
# Color H-bond lines
color yellow, hbonds # yellow is the most common publication color for H-bonds
color cyan, hbonds # cyan is also widely used and visible on gray backgrounds
# Hide the distance labels if you're annotating manually
hide labels, hbonds
# Show labels (distance in Ångströms) — include in figure
show labels, hbonds
set label_size, 12
set label_color, blackmode=2 H-bond detection uses a simple geometric criterion — it does not account for hydrogen atom positions (which may not be present in the PDB file) or electronic structure. Use it as a starting point. For a publication, verify all H-bonds manually by checking donor-acceptor distances and angles, or use PLIP (Protein-Ligand Interaction Profiler, accessible at plip-tool.org) for more accurate interaction detection and comparison.
Hydrophobic and van der Waals contacts
Hydrophobic contacts don’t have the same clear geometric signature as H-bonds — they’re simply close contacts between non-polar atoms. PyMOL can show them as distance lines with a wider cutoff:
# All contacts within 4 Å (captures both H-bonds and close contacts)
distance contacts_4A, LIG, pocket, 4.0
# Show only hydrophobic contacts (C-C interactions, exclude N, O, S)
select lig_carbon, LIG and elem C
select pock_carbon, pocket and elem C
distance hydrophobic, lig_carbon, pock_carbon, 4.5
# Color hydrophobic contacts differently from H-bonds
color gray60, hydrophobic
# Alternative: show hydrophobic surface coloring instead of distance lines
spectrum hydrophobicity, white_orange, pocket
# Alternatively, just visually show which residues line the hydrophobic pocket
select hydrophobic_residues, pocket and resn PHE+LEU+ILE+VAL+MET+TRP+ALA+PRO
color tv_orange, hydrophobic_residuesDistance measurements
For specific interactions you want to quantify — catalytic distances, key H-bond lengths, metal coordination bonds — the distance command creates labeled lines showing the exact distance in ångströms.
# Measure distance between specific atoms
# Full atom identifier: /object//chain/resname`resi/atom_name
distance d1, /4XYZ//A/LIG`401/O3, /4XYZ//A/SER`185/OG
# Simpler: click two atoms in the viewer (Measurement Wizard)
# Wizard → Measurement → then click two atoms
# Or use the distance command with selection expressions
distance d_key, (resn LIG and name O3), (resi 185 and name OG and chain A)
# Styling distance objects for publication
set dash_gap, 0.4 # gap between dashes (default 0.4)
set dash_length, 0.2 # length of each dash
set dash_width, 2.0 # line thickness
set label_size, 14 # distance label font size
# Show distances without labels (for cleaner figures annotated elsewhere)
hide labels| mode= value | What it detects |
|---|---|
| mode=0 (default) | All atom pairs within the cutoff distance — shows every contact |
| mode=1 | Contacts between different residues (excludes within-residue bonds) |
| mode=2 | H-bond geometry check (donor-acceptor pairs, correct angles) |
| mode=3 | Saltbridge detection (charged atoms only) |
| mode=4 | Covalent bonds (for showing explicit bond connections) |
2D interaction diagrams
PyMOL produces excellent 3D binding site figures but cannot generate 2D interaction diagrams — the schematic “ligand map” views showing all interactions as labeled arrows and symbols that are standard in medicinal chemistry papers. These require dedicated tools:
-
PLIPplip-tool.org — free web server that detects and reports all non-covalent interactions from a PDB or uploaded file. Also generates a PyMOL visualization script you can load directly. Best free option for interaction analysis.
-
LigPlot+ebi.ac.uk/thornton-srv/software/LigPlus — generates the classic 2D schematic diagrams of protein-ligand interactions. Free for academic use. Output is a clean 2D diagram suitable for supplementary figures.
-
Maestro / GlideSchrödinger’s commercial suite. Industry standard for interaction visualization in drug discovery. Beautiful 2D and 3D interaction figures. Academic license available through the PyMOL educational license.
-
ProteinsPlusproteins.plus — web-based platform that includes PoseView for 2D ligand interaction diagrams. Free, no registration required.
The standard workflow in a structural biology or drug discovery paper is to use PyMOL for the 3D figure panel and one of the above tools for a 2D schematic in the same figure or supplementary information.
Complete publication figure workflow
# ── Load and clean ──────────────────────────────────
fetch 4XYZ
remove resn HOH
remove resn SO4+PEG+GOL
# ── Define selections ───────────────────────────────
select LIG, resn LIG
select pocket, byres (LIG around 5)
# ── Representations ─────────────────────────────────
as cartoon
hide cartoon, pocket or LIG
show sticks, pocket
show sticks, LIG
# ── Colors ──────────────────────────────────────────
color gray70, polymer # neutral protein context
color slate, pocket # slightly blue pocket residues
util.cbag LIG # green carbons, CPK for ligand
set stick_radius, 0.15, pocket
set stick_radius, 0.25, LIG
# ── H-bonds ─────────────────────────────────────────
distance hbonds, LIG, pocket, mode=2
color yellow, hbonds
hide labels, hbonds # annotate manually in figure legend
# ── Key distances (add your specific atoms here) ────
distance d1, (resn LIG and name O3), (resi 185 and name OG and chain A)
color cyan, d1
set label_size, 13
# ── Optional: binding site surface ──────────────────
show surface, pocket
set transparency, 0.5
color white, pocket
# ── Figure settings ─────────────────────────────────
bg_color white
set ray_opaque_background, 1
set cartoon_fancy_helices, 1
set ambient, 0.40
set specular, 0.25
set light_count, 4
set ray_shadow, 1
set antialias, 2
# ── Orient (paste your get_view output here) ────────
zoom LIG, 5
# ── Export ──────────────────────────────────────────
png binding_site.png, width=1200, height=900, dpi=300, ray=1Protein-ligand visualization in one paragraph
The standard binding site figure uses three layers: gray cartoon protein context, slate-colored pocket residues as sticks, and a CPK-colored ligand with green carbons. Add H-bonds with distance hbonds, LIG, pocket, mode=2 and color them yellow. For specific key distances, use individual distance commands between named atoms. Keep hydrophobic contacts as residue coloring rather than distance lines — too many lines obscure the binding mode rather than clarifying it. For 2D schematic diagrams, use PLIP or LigPlot+ alongside your 3D PyMOL figure. Save everything as a .pml script so the figure can be regenerated after reviewer requests.