Structure alignment analysis

Structure Alignment & Analysis

Workflows for structural comparison of proteins and other molecules.

Send all cmd.* code via: ~/.pymol-agent-bridge/bin/pymol-agent-bridge exec "..." (or heredoc for multi-line). See @pymol-fundamentals for details.

Alignment Methods

align - Sequence + Structure

Best for: Homologous proteins with similar sequences.

# Basic alignment (mobile moves to target)
result = cmd.align("mobile", "target")

# Result tuple: (RMSD, aligned_atoms, cycles, RMSD_pre, aligned_pairs, raw_score, align_length)
print("RMSD: " + str(round(result[0], 2)) + " A")
print("Atoms aligned: " + str(result[1]))
print("Refinement cycles: " + str(result[2]))

super - Structure Only

Best for: Proteins with low sequence identity but structural similarity.

result = cmd.super("mobile", "target")
print("RMSD: " + str(round(result[0], 2)) + " A")
print("Atoms aligned: " + str(result[1]))

cealign - Combinatorial Extension

Best for: Detecting distant structural homology.

result = cmd.cealign("target", "mobile")
print("RMSD: " + str(round(result["RMSD"], 2)) + " A")
print("Aligned length: " + str(result["alignment_length"]))

Note: cealign argument order is (target, mobile), opposite of align/super.

RMSD Calculations

After Alignment

# All atoms
rms = cmd.rms_cur("mobile", "target")

# CA atoms only (standard for backbone comparison)
rms = cmd.rms_cur("mobile and name CA", "target and name CA")
print("CA RMSD: " + str(round(rms, 2)) + " A")

Without Moving Structures

# Calculate RMSD without transformation
rms = cmd.rms("mobile", "target", matchmaker=4)

Loading Multiple Structures

From PDB

cmd.fetch("1tim")
cmd.fetch("8tim")
print("Loaded: " + str(cmd.get_names()))

Naming Convention

cmd.fetch("1tim", name="reference")
cmd.fetch("8tim", name="variant")

Visualization

Color by Structure

cmd.color("green", "reference")
cmd.color("cyan", "variant")
cmd.show("cartoon")
cmd.hide("lines")

Show Alignment

# Create alignment object for visualization
cmd.align("variant", "reference", object="alignment")
cmd.show("cgo", "alignment")  # Shows aligned residue connections

Highlight Differences

# Color by RMSD (requires script)
# Simpler approach: color conserved vs variable regions
cmd.select("core", "variant within 1 of reference")
cmd.select("variable", "variant and not core")
cmd.color("white", "core")
cmd.color("red", "variable")

Partial Alignment

Align Specific Regions

# Align only chain A
result = cmd.align("mobile and chain A", "target and chain A")

# Align specific domain
result = cmd.align("mobile and resi 1-100", "target and resi 1-100")

# Align only CA atoms
result = cmd.align("mobile and name CA", "target and name CA")

Exclude Flexible Regions

# Exclude loops from alignment
result = cmd.align(
    "mobile and ss h+s",  # Only helices and sheets
    "target and ss h+s"
)

Multi-Structure Comparison

Sequential Alignment

# Align all to reference
cmd.fetch("1tim", name="ref")
cmd.fetch("2tim", name="var1")
cmd.fetch("3tim", name="var2")

for structure in ["var1", "var2"]:
    result = cmd.align(structure, "ref")
    print(structure + " RMSD: " + str(round(result[0], 2)))

Color Scheme

colors = ["green", "cyan", "magenta", "yellow", "orange"]
structures = ["ref", "var1", "var2"]
for i, struct in enumerate(structures):
    cmd.color(colors[i], struct)

Complete Workflow

Standard Structure Comparison

# 1. Load structures
cmd.delete("all")
cmd.fetch("1tim", name="reference")
cmd.fetch("8tim", name="variant")

# 2. Align
result = cmd.align("variant", "reference")
print("RMSD: " + str(round(result[0], 2)) + " A")
print("Atoms: " + str(result[1]))

# 3. Style
cmd.show("cartoon")
cmd.hide("lines")
cmd.color("green", "reference")
cmd.color("cyan", "variant")

# 4. Calculate backbone RMSD
ca_rms = cmd.rms_cur("variant and name CA", "reference and name CA")
print("CA RMSD: " + str(round(ca_rms, 2)) + " A")

# 5. View
cmd.center("all")
cmd.zoom("all")
cmd.bg_color("white")

Interpretation

RMSD (CA)	Interpretation
< 0.5 A	Nearly identical
0.5-1.0 A	Very similar (same fold, minor differences)
1.0-2.0 A	Similar fold, some conformational differences
2.0-3.0 A	Same fold family, significant differences
> 3.0 A	Different conformations or distant homologs

Publication-Quality Figures

Settings for Print

cmd.bg_color("white")
cmd.set("ray_shadows", 0)
cmd.set("spec_reflect", 0.2)
cmd.set("ambient", 0.5)
cmd.set("antialias", 2)
cmd.set("ray_trace_mode", 1)
cmd.set("cartoon_fancy_helices", 1)

Multi-View Strategy

Always create multiple views to tell the complete story:

Overview - Shows full aligned structures:

cmd.orient()
cmd.ray(1200, 900)
cmd.png("comparison_overview.png")

Side View - Rotated 90° to show depth:

cmd.turn("y", 90)
cmd.ray(1200, 900)
cmd.png("comparison_side.png")

Detail View - Focus on key difference regions:

cmd.center("mobile and resi 65-85")  # Hinge region, mutation site, etc.
cmd.zoom("all and resi 60-90", buffer=8)
cmd.ray(1200, 900)
cmd.png("comparison_detail.png")

Color Schemes for Comparison

Standard two-structure comparison:

• Structure 1: marine (blue)
• Structure 2: salmon (pink/orange)

Multi-structure comparison:

• Reference: green
• Variants: cyan, magenta, yellow, orange

IMPORTANT: Always use cmd.ray(w, h) then cmd.png(path) without dimensions. Using cmd.png(path, w, h) causes a view inflation bug.

Tips

• Use align for closely related proteins
• Use super when sequence identity is low
• Use cealign for detecting distant relationships
• Always report CA RMSD for standardization
• Color structures distinctly for clarity
• rms_cur calculates RMSD of current positions (after alignment)
• Multiple views tell a more complete story than a single image