mdtraj Trajectory Analysis

Overview

mdtraj is a dependency-light Python library for analyzing MD trajectories. Reads DCD/XTC/TRR/NetCDF/H5/AMBER/GROMACS/CHARMM/OpenMM into a Trajectory object backed by NumPy arrays, then exposes geometry, RMSD/Rg/RMSF/SASA, contacts, hydrogen bonds, torsions, and 8-state DSSP as pure-Python functions.

Units: mdtraj uses nm and ps internally. Multiply distances by 10 for Å, divide time by 1000 for ns. Torsions are in radians — np.degrees().

When to Use

• RMSD vs time, Rg, per-residue RMSF for stability across MD replicates
• Residue-residue contact frequency maps from a trajectory ensemble
• Backbone phi/psi for Ramachandran (general, Gly, Pro)
• 8-state DSSP per residue per frame for secondary-structure time series
• Lightweight ad-hoc analyses where MDAnalysis's full framework is overkill
• NMR observables (J-couplings, chemical shifts via SHIFTX2)
• Use mdanalysis-trajectory instead when you need MDAnalysis's selection grammar, AnalysisBase parallelism, or LAMMPS/NAMD-specific readers
• Use OpenMM/GROMACS to run the simulation — this skill is post-simulation only