Building DNA Structures¶

This guide covers all the ways to create DNA structures with MDNA: from simple sequences to custom-shaped assemblies.

Basic Construction¶

From a Sequence¶

import mdna

dna = mdna.make(sequence='GCGCGCGCGC')
dna.describe()
dna.draw()

From a Number of Base Pairs¶

When no sequence is provided, a random sequence is generated:

dna = mdna.make(n_bp=50)
print(dna.sequence)

From Existing Data¶

Load from an MDTraj trajectory or PDB file:

import mdtraj as md

# From a PDB file
dna = mdna.load(filename='my_dna.pdb')

# From an MDTraj trajectory
traj = md.load('trajectory.xtc', top='topology.pdb')
dna = mdna.load(traj=traj, chainids=[0, 1])

Or from precomputed reference frames:

import numpy as np
frames = np.load('frames.npy')  # shape (n_bp, 4, 3)
dna = mdna.load(frames=frames, sequence='ATCG...')

Custom Shapes¶

MDNA uses spline interpolation through control points to define arbitrary DNA paths.

Built-in Shapes¶

The Shapes class provides common geometries:

# Straight line (default)
points = mdna.Shapes.line(length=5)

# Circle
points = mdna.Shapes.circle(radius=3)

# Helix / supercoil
points = mdna.Shapes.helix(height=3, pitch=5, radius=7, num_turns=4)

Use them with make():

dna = mdna.make(n_bp=300, control_points=mdna.Shapes.helix())
dna.draw()

Custom Control Points¶

Define any 3D path with at least 4 points:

import numpy as np

control_points = np.array([
    [0, 0, 0],
    [30, 10, -10],
    [50, 10, 20],
    [3, 4, 30]
])

dna = mdna.make(n_bp=100, control_points=control_points)
dna.draw()

Tip

When only control_points are provided (no sequence or n_bp), MDNA infers n_bp from the spline arc length using the standard 0.34 nm rise per base pair.

Circular DNA¶

Create closed minicircles with optional linking number control:

# Relaxed minicircle
dna = mdna.make(n_bp=200, circular=True)
print('Lk, Wr, Tw:', dna.get_linking_number())
dna.draw()

# Supercoiled minicircle (overwound by 8 turns)
dna = mdna.make(n_bp=200, circular=True, dLk=8)
print('Lk, Wr, Tw:', dna.get_linking_number())

Note

When minimizing supercoiled DNA, use equilibrate_writhe=True to allow the writhe to equilibrate:

dna.minimize(equilibrate_writhe=True)

Minimization¶

Monte Carlo relaxation resolves steric clashes and elastic strain:

dna = mdna.make(n_bp=100)
dna.minimize(
    temperature=300,      # Kelvin
    exvol_rad=2.0,        # Excluded volume radius
    endpoints_fixed=True  # Pin the ends
)

You can also fix specific base pairs:

dna.minimize(fixed=[0, 1, 2, 97, 98, 99])

To view the Monte Carlo trajectory:

mc_traj = dna.get_MC_traj()

Extending DNA¶

Add base pairs to the 3' or 5' end of an existing structure:

dna = mdna.make(n_bp=50)

# Extend forward (3' direction)
dna.extend(sequence='G' * 30, forward=True)

# Extend backward (5' direction)
dna.extend(sequence='C' * 20, forward=False)

dna.draw()

Connecting Two DNA Structures¶

Join two Nucleic objects with an automatically optimized linker:

import numpy as np

dna0 = mdna.make(sequence='AAAAAAAAA')
dna1 = mdna.make(
    sequence='GGGGGGGGG',
    control_points=mdna.Shapes.line(1) + np.array([4, 0, -5])
)

# connect() finds the optimal number of linker base pairs
connected = mdna.connect(dna0, dna1)
connected.describe()
connected.draw()

The connect() function:

Computes the twist difference between the two end frames
Finds the optimal number of linking base pairs to achieve neutral twist
Interpolates a connecting path between the two structures
Optionally minimizes the resulting assembly

Exporting Structures¶

Save as PDB¶

dna.save_pdb('output.pdb')

Get MDTraj Trajectory¶

traj = dna.get_traj()
traj.save('output.h5')

Direct PDB Generation¶

For one-shot PDB generation without creating a Nucleic object:

traj = mdna.sequence_to_pdb(
    sequence='CGCGAATTCGCG',
    filename='my_dna',
    output='GROMACS'
)

Summary¶

Task	Function / Method
Generate from sequence/shape	`mdna.make()`
Load from trajectory/frames	`mdna.load()`
Relax structure	`nucleic.minimize()`
Extend DNA	`nucleic.extend()`
Join two structures	`mdna.connect()`
Export PDB	`nucleic.save_pdb()`
One-shot PDB	`mdna.sequence_to_pdb()`
MD simulation	`mdna.sequence_to_md()`