Analysis > Trajectory Analysis... gathers in a single window the analyses that read the frames of an object (molecular dynamics, scans, NEB, umbrella sampling, imported trajectories). The analyses are grouped in a list on the left; the options of the selected analysis appear below it, and the results are shown on the right as a plot (with several Views for some analyses), a Table and a Summary (Figure 1).
Input (top of the window)
System / Object: the trajectory.
First frame / Last / Step: the frames analyzed (Last = -1 means the last frame).
Time per frame (ps): the time between frames; 0 plots against the frame number.
Atom set: the Whole object or the Current selection in the 3D view. Ignore water and ions leaves water and single-atom ions out of the atom set.
Click Run; long analyses run in the background and can be cancelled. Export Image... saves the plot and Save Data... the numbers.
Figure 1: Trajectory Analysis: DSSP map of chignolin (1UAO) along a 10 ps MD simulation, showing the β-hairpin (strand–turn–strand).
Available analyses
Structure
Radius of gyration: Rg of the atom set in every frame (optionally mass weighted).
Secondary structure (DSSP): DSSP (Kabsch and Sander, 1983) for the protein residues of the atom set, shown as a residue × frame map (Figure 1) and as the fraction of each structure type per residue. Missing amide hydrogens are placed by the DSSP rule. Clicking the map goes to that frame.
Solvent accessible surface (SASA): Shrake–Rupley SASA of the atom set alone (probe radius, default 1.4 Å; points per atom, default 100).
Interactions
Hydrogen bonds: GROMACS criterion (donor–acceptor distance, default 3.5 Å; H–D···A angle, default 30°), within the atom set, between the atom set and the rest, or between the atom set and water, with a minimum occupancy filter. Double-clicking a row selects its atoms and goes to a frame where the bond exists.
Contact / distance map: contact frequency, mean CA–CA distance, or the change in contacts between the two halves of the trajectory, for the residues of the atom set (contact cutoff, default 4.5 Å). Clicking a cell selects the two residues.
Interaction energy: Coulomb and Lennard-Jones energy between the atom set and a partner group (the rest of the system, with or without water and ions, or a selection), with the MM charges and parameters of the system.
Electron transfer pathways: the Pathways model (Beratan and Onuchic): the best donor-to-acceptor path through covalent bonds, hydrogen bonds and through-space jumps, and its coupling along the trajectory.
Dynamics
Principal components (PCA): frames are superposed on their average first; the variance of each component, the projections on PC1 and PC2, and buttons to animate a component or color the atoms by their weight in PC1 (Figure 2).
Cross-correlation (DCCM): Cij = <Δri·Δrj> / sqrt(<Δri2><Δrj2>) after superposition, as a heat map.
Clustering: GROMOS clustering (Daura et al., 1999) on the pairwise RMSD after superposition (RMSD cutoff, default 1.5 Å).
B-factors and average structure: RMSF, isotropic B = 8π2/3 <Δr2> and anisotropic U tensors, plus the average structure; the B-factors can be exported in a PDB file.
Quasi-harmonic entropy: quasi-harmonic frequencies and entropy from the mass-weighted covariance (Andricioaei and Karplus, 2001) and Schlitter's bound. It needs many more frames than three times the number of atoms.
Side-chain rotamers (chi1, chi2): the side-chain dihedrals of the residues along the trajectory.
Diffusion (MSD): mean square displacement of the water oxygens or of the atom set, over time origins and with periodic jumps removed; the diffusion coefficient comes from a linear fit, MSD = 6Dt. Set the time per frame for D in physical units.
Statistics
Statistics of a series: mean, standard deviation, block-averaged standard error (Flyvbjerg and Petersen, 1989) and running mean of any series produced by the other analyses, or loaded from a file (columns x y), optionally discarding the first part as equilibration.
Figure 2: Principal component analysis (CA atoms) of the same trajectory: variance per component and cumulative variance.