Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Structure Modification · FreedomIntelligence | DecimalAI

Bio Pdb Structure Modification

NewSkillScoreMaintained

bio-pdb-structure-modification

Modify protein structures using Biopython Bio.PDB. Use when transforming coordinates, removing atoms or residues, adding new entities, modifying B-factors and occupancies, or building structures programmatically.

scienceby FreedomIntelligence

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/bio-pdb-structure-modification/SKILL.md

decimalai skills pull bio-pdb-structure-modification --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/bio-pdb-structure-modification/SKILL.md https://app.decimal.ai/s/bio-pdb-structure-modification/SKILL.md

Raw file: /s/bio-pdb-structure-modification/SKILL.md ↗ · machine-readable summary at /s/bio-pdb-structure-modification.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 57d ago

No — -100% lift+187% tokens0% turns0% of whole cases pass with skill (0/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✗	＝ Same ✗	—	—
case-08	✗→✗	＝ Same ✗	—	—
case-11	✗→✗	＝ Same ✗	—	—
case-22	✗→✗	＝ Same ✗	—	—
case-05	✗→✗	＝ Same ✗	—	—

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 7/30/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

FreedomIntelligence / OpenClaw-Medical-Skills

skills/bio-pdb-structure-modification/SKILL.md

View on GitHub ↗

Imported 8/1/2026Version v1

Skill contents

SKILL.md

Version Compatibility

Reference examples tested with: BioPython 1.83+, numpy 1.26+

Before using code patterns, verify installed versions match. If versions differ:

Python: pip show <package> then help(module.function) to check signatures

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.

Structure Modification

"Extract a chain from a PDB file" → Remove/add atoms, residues, or chains; transform coordinates; modify B-factors and occupancies; build structures programmatically.

Python: Bio.PDB.PDBIO() with Select subclass for filtering, Bio.PDB.Superimposer() for transforms

Transform coordinates, remove/add entities, modify properties, and build structures programmatically.

Required Imports

python
from Bio.PDB import PDBParser, PDBIO, StructureBuilder
from Bio.PDB.Structure import Structure
from Bio.PDB.Model import Model
from Bio.PDB.Chain import Chain
from Bio.PDB.Residue import Residue
from Bio.PDB.Atom import Atom
import numpy as np

Transforming Coordinates

python
from Bio.PDB import PDBParser, PDBIO
import numpy as np

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Translate all atoms
translation = np.array([10.0, 0.0, 0.0])
for atom in structure.get_atoms():
    atom.coord = atom.coord + translation

# Save transformed structure
io = PDBIO()
io.set_structure(structure)
io.save('translated.pdb')

Rotation Around Axis

python
from Bio.PDB import PDBParser
from Bio.PDB.vectors import rotaxis
import numpy as np

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Rotate around Z axis by 90 degrees
angle = np.radians(90)
axis = np.array([0, 0, 1])

# Get center of mass for rotation origin
coords = np.array([a.coord for a in structure.get_atoms()])
center = coords.mean(axis=0)

# Rotation matrix
cos_a = np.cos(angle)
sin_a = np.sin(angle)
rot_matrix = np.array([
    [cos_a, -sin_a, 0],
    [sin_a, cos_a, 0],
    [0, 0, 1]
])

# Apply rotation around center
for atom in structure.get_atoms():
    atom.coord = np.dot(rot_matrix, atom.coord - center) + center

Applying Transformation Matrix

python
from Bio.PDB import PDBParser
import numpy as np

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# 4x4 transformation matrix (rotation + translation)
# From superimposition or external source
transform = np.array([
    [1.0, 0.0, 0.0, 10.0],
    [0.0, 1.0, 0.0, 5.0],
    [0.0, 0.0, 1.0, 0.0],
    [0.0, 0.0, 0.0, 1.0]
])

rotation = transform[:3, :3]
translation = transform[:3, 3]

for atom in structure.get_atoms():
    atom.coord = np.dot(rotation, atom.coord) + translation

Center Structure at Origin

python
from Bio.PDB import PDBParser
import numpy as np

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Calculate center
coords = np.array([a.coord for a in structure.get_atoms()])
center = coords.mean(axis=0)

# Translate to origin
for atom in structure.get_atoms():
    atom.coord = atom.coord - center

Removing Atoms

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Remove hydrogens
for residue in structure.get_residues():
    atoms_to_remove = [a.id for a in residue if a.element == 'H']
    for atom_id in atoms_to_remove:
        residue.detach_child(atom_id)

io = PDBIO()
io.set_structure(structure)
io.save('no_hydrogens.pdb')

Removing Residues

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Remove water molecules
for model in structure:
    for chain in model:
        residues_to_remove = [r.id for r in chain if r.id[0] == 'W']
        for res_id in residues_to_remove:
            chain.detach_child(res_id)

io = PDBIO()
io.set_structure(structure)
io.save('no_water.pdb')

Removing a Chain

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Remove chain B
model = structure[0]
if model.has_id('B'):
    model.detach_child('B')

io = PDBIO()
io.set_structure(structure)
io.save('without_chain_B.pdb')

Modifying B-factors

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Set all B-factors to same value
for atom in structure.get_atoms():
    atom.bfactor = 20.0

# Or set based on residue property (e.g., conservation score)
conservation_scores = {100: 9.0, 101: 5.0, 102: 3.0}  # resnum -> score
for residue in structure.get_residues():
    resnum = residue.id[1]
    score = conservation_scores.get(resnum, 5.0)
    for atom in residue:
        atom.bfactor = score * 10  # Scale to B-factor range

io = PDBIO()
io.set_structure(structure)
io.save('modified_bfactor.pdb')

Modifying Occupancy

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Set occupancy for specific chain
for atom in structure[0]['A'].get_atoms():
    atom.occupancy = 0.5

io = PDBIO()
io.set_structure(structure)
io.save('modified_occupancy.pdb')

Renumbering Residues

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

chain = structure[0]['A']

# Renumber sequentially starting from 1
new_residues = []
for i, residue in enumerate(chain, start=1):
    hetfield, _, icode = residue.id
    new_id = (hetfield, i, icode)
    residue.id = new_id
    new_residues.append(residue)

io = PDBIO()
io.set_structure(structure)
io.save('renumbered.pdb')

Changing Chain ID

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Rename chain A to X
model = structure[0]
chain = model['A']
chain.id = 'X'

io = PDBIO()
io.set_structure(structure)
io.save('renamed_chain.pdb')

Building Structure from Scratch

python
from Bio.PDB.Structure import Structure
from Bio.PDB.Model import Model
from Bio.PDB.Chain import Chain
from Bio.PDB.Residue import Residue
from Bio.PDB.Atom import Atom
from Bio.PDB import PDBIO
import numpy as np

# Create hierarchy
structure = Structure('new_struct')
model = Model(0)
chain = Chain('A')
residue = Residue((' ', 1, ' '), 'ALA', '')

# Add atoms
ca = Atom('CA', np.array([0.0, 0.0, 0.0]), 20.0, 1.0, ' ', 'CA', 1, 'C')
cb = Atom('CB', np.array([1.5, 0.0, 0.0]), 20.0, 1.0, ' ', 'CB', 2, 'C')

residue.add(ca)
residue.add(cb)
chain.add(residue)
model.add(chain)
structure.add(model)

io = PDBIO()
io.set_structure(structure)
io.save('new_structure.pdb')

Using StructureBuilder

python
from Bio.PDB import StructureBuilder, PDBIO
import numpy as np

sb = StructureBuilder.StructureBuilder()
sb.init_structure('built')
sb.init_model(0)
sb.init_chain('A')
sb.init_seg(' ')

# Add residue with atoms
sb.init_residue('ALA', ' ', 1, ' ')
sb.init_atom('N', np.array([-1.0, 0.0, 0.0]), 20.0, 1.0, ' ', 'N', 1, 'N')
sb.init_atom('CA', np.array([0.0, 0.0, 0.0]), 20.0, 1.0, ' ', 'CA', 2, 'C')
sb.init_atom('C', np.array([1.0, 0.0, 0.0]), 20.0, 1.0, ' ', 'C', 3, 'C')
sb.init_atom('O', np.array([1.5, 1.0, 0.0]), 20.0, 1.0, ' ', 'O', 4, 'O')

sb.init_residue('GLY', ' ', 2, ' ')
sb.init_atom('N', np.array([1.5, -1.0, 0.0]), 20.0, 1.0, ' ', 'N', 5, 'N')
sb.init_atom('CA', np.array([2.5, -1.0, 0.0]), 20.0, 1.0, ' ', 'CA', 6, 'C')
sb.init_atom('C', np.array([3.5, -1.0, 0.0]), 20.0, 1.0, ' ', 'C', 7, 'C')
sb.init_atom('O', np.array([4.0, 0.0, 0.0]), 20.0, 1.0, ' ', 'O', 8, 'O')

structure = sb.get_structure()

io = PDBIO()
io.set_structure(structure)
io.save('built_structure.pdb')

Adding a Residue to Existing Chain

python
from Bio.PDB import PDBParser, PDBIO
from Bio.PDB.Residue import Residue
from Bio.PDB.Atom import Atom
import numpy as np

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

chain = structure[0]['A']

# Create new residue
new_residue = Residue((' ', 999, ' '), 'ALA', '')
ca = Atom('CA', np.array([50.0, 50.0, 50.0]), 20.0, 1.0, ' ', 'CA', 9999, 'C')
new_residue.add(ca)

# Add to chain
chain.add(new_residue)

io = PDBIO()
io.set_structure(structure)
io.save('with_new_residue.pdb')

Copying a Chain

python
from Bio.PDB import PDBParser, PDBIO
from Bio.PDB.Chain import Chain
from Bio.PDB.Residue import Residue
from Bio.PDB.Atom import Atom
import copy

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Deep copy chain A as chain B
original_chain = structure[0]['A']
new_chain = Chain('B')

for residue in original_chain:
    new_residue = Residue(residue.id, residue.resname, residue.segid)
    for atom in residue:
        new_atom = Atom(
            atom.name, atom.coord.copy(), atom.bfactor, atom.occupancy,
            atom.altloc, atom.fullname, atom.serial_number, atom.element
        )
        new_residue.add(new_atom)
    new_chain.add(new_residue)

structure[0].add(new_chain)

io = PDBIO()
io.set_structure(structure)
io.save('duplicated_chain.pdb')

Extract and Save Specific Residues

python
from Bio.PDB import PDBParser, PDBIO
from Bio.PDB.Structure import Structure
from Bio.PDB.Model import Model
from Bio.PDB.Chain import Chain

parser = PDBParser(QUIET=True)
structure = parser.get_structure('protein', 'protein.pdb')

# Extract residues 50-100
new_structure = Structure('subset')
new_model = Model(0)
new_chain = Chain('A')

for residue in structure[0]['A']:
    if 50 <= residue.id[1] <= 100 and residue.id[0] == ' ':
        new_chain.add(residue.copy())

new_model.add(new_chain)
new_structure.add(new_model)

io = PDBIO()
io.set_structure(new_structure)
io.save('residues_50_100.pdb')

Merge Two Structures

python
from Bio.PDB import PDBParser, PDBIO

parser = PDBParser(QUIET=True)
struct1 = parser.get_structure('s1', 'structure1.pdb')
struct2 = parser.get_structure('s2', 'structure2.pdb')

# Add chains from struct2 to struct1 (rename to avoid conflicts)
for chain in struct2[0]:
    new_id = chr(ord(chain.id) + 10)  # Offset chain ID
    chain.id = new_id
    struct1[0].add(chain)

io = PDBIO()
io.set_structure(struct1)
io.save('merged.pdb')

Related Skills

structure-io - Parse and write structure files
structure-navigation - Access chains, residues, atoms
geometric-analysis - Calculate distances, angles, RMSD
sequence-manipulation/seq-objects - Generate sequences from modified structures

Illustrative Want to see what bio-pdb-structure-modification would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

7/26/2026, 09:18 UTC · 23 cases

Lift

-100%

0% vs 100%

Agent turns

1.0 vs 1.0

Tokens

+187%

4,982.87 vs 1,736.48

fail → fail · 23

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-06 We want to set the occupancy of all atoms in Chain A to 0.5 using BioPython. Someone suggested setting `atom.occ = 0.5` or `atom.occupancy_value = 0.5`. What is the correct property name on the Atom instance?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-08 We want to rename Chain 'A' to Chain 'X' in a PDB model using BioPython. Someone suggested calling `model.rename_chain('A', 'X')`. What attribute on the Chain object itself should be assigned to change its identifier?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-11 We have an existing PDB structure loaded in BioPython and want to append a newly constructed Residue object to Chain A. What method on the Chain object attaches this new residue?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-22 We need to extract crystallographic resolution and experimental method metadata from an MMCIF file using BioPython. Should we alter structural objects in memory? How should this header metadata be accessed?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-05 We have a dictionary mapping residue numbers to conservation scores. We want to update the temperature factors (B-factors) in a BioPython PDB structure so visualization tools color by score. A colleague suggested modifying `atom.temp_factor`. What attribute on Atom objects should be updated?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-15 We need to center a protein structure at the coordinate origin (0, 0, 0) using BioPython before running rigid-body docking. How is the center of mass / centroid calculated from atom coordinates, and how is the shift applied?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-19 We want to set the occupancy of every atom specifically in Chain 'A' to 1.0 while leaving other chains untouched using BioPython. How do we traverse atoms in Chain 'A' directly?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-17 After completing coordinate shifts and residue updates on a BioPython Structure object, what IO class exports the updated object back to a PDB format file on disk, and what are its key output methods?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 We want to find all amino acid residues located within 5.0 Angstroms of a bound ligand in a PDB structure using BioPython. Should we modify coordinates or detach non-neighboring residues? How should spatial neighbor querying be performed?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-18 To rotate a protein structure around the Z-axis centered at its centroid using BioPython and NumPy, how is the rotation matrix applied relative to the centroid position?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-12 We need to duplicate Chain A as Chain B in a model using BioPython. Passing `structure[0]['A']` directly to `model.add()` fails due to parent link collisions and ID conflicts. How should new Atom instances be instantiated when copying residues over?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 We want to strip all solvent/water molecules from a structure model using BioPython Bio.PDB. A team member suggested checking `residue.resname == 'HOH'` and calling `chain.remove(residue)`. What is the canonical method on Chain objects to remove water residues based on their hetero-flag?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-16 We have a 4x4 transformation matrix (3x3 rotation block and 1x3 translation vector) from an external alignment tool. How should this matrix transformation be applied to atom coordinates in BioPython?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-04 We need to remove chain 'B' from the first model of a PDB structure in BioPython. A developer wrote `del model['B']`. What is the recommended Bio.PDB Model method to safely drop chain 'B' after checking if it exists?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-01 We need a Python script using BioPython to shift every atom's 3D coordinates in a loaded PDB structure by +15.0 along the X-axis before writing out the file. An associate suggested doing string operations on raw PDB file text lines (columns 31-54). How should this coordinate translation be performed programmatically on the Bio.PDB object hierarchy?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-23 In BioPython, when updating a residue ID tuple `(hetfield, resnum, icode)` during residue renumbering, what value should `hetfield` hold for standard amino acid residues?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-20 We need to calculate the backbone phi and psi torsion angles for all amino acid residues in a PDB file using BioPython. Should we modify atom coordinate vectors manually to compute these angles? What standard Bio.PDB module should be used instead?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-14 We want to merge chains from two separate PDB files (`struct1.pdb` and `struct2.pdb`) into a single model using BioPython. If both structures have a Chain 'A', what must be modified on the chain from `struct2` before adding it to `struct1[0]`?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-13 We want to extract residues numbered 50 through 100 from Chain A of a PDB file and save them as a new standalone structure using BioPython. How should the target residues be copied into a new Chain object?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-10 We want to build a protein backbone programmatically using BioPython's helper builder class that manages entity initialization step-by-step. Which BioPython class provides step-by-step methods like `init_structure`, `init_model`, `init_chain`, `init_residue`, and `init_atom`?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-07 We need to renumber residues in Chain A sequentially from 1 to N using BioPython. A junior dev tried setting `residue.number = i` or `residue.resnum = i`. How must a residue's identifier be updated in Bio.PDB?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-09 We need to construct a minimal PDB structure object from scratch in Python without reading an existing file. We want to instantiate the object hierarchy level by level. What are the four core structural classes imported from Bio.PDB that form this hierarchy, and how are child objects attached?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-02 In a BioPython structural script, we want to strip all hydrogen atoms from a PDB structure prior to simulation setup. Someone recommended using `residue.remove_atom()` or filtering out 'H' strings from exported text. What is the standard Bio.PDB method on a Residue object to detach an atom by its ID?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of -100 percentage points is the difference between those two pass rates over the 23 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 100%
Pass rate with the skill: 0%
Cases attempted: 23
Cases counted toward lift: 23
Cases that regressed: 2
Skill version measured: v1
Run completed: 7/26/2026, 09:18 UTC

The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested 56 days ago

Other measured skills in the registry, with their headline benchmark lift.