v0.6

New features:

• add argument salt_concentration - salt concentration in mol/L passed to gmx genion -conc (default: not set, only charge neutralization is performed)
• add argument ion_pname - positive ion name passed to gmx genion -pname (default NA)
• add argument ion_nname - negative ion name passed to gmx genion -nname (default CL)
• add argument water_model - water model passed to gmx pdb2gmx -water (default tip3p)
• update GROMACS to 2025.2 and MDAnalysis to 2.10.0
• protein RMSD is reported as both backbone (N/CA/C/O, the backbone superposition/fit group) and CA (the protein Cα subset)
• add run_prolif water-mediated and halogen bond analysis (requires >= ProLIF 2.1.0 version)
• run_prolif major speed-up: restrict the ProLIF protein to binding-site residues (union of residues within the cutoff of the ligand over the trajectory) instead of converting the whole protein to RDKit every frame; results are identical (~15–20× faster for large solvated systems). New argument binding_site_cutoff (default 12 Å; set 0 to use the full protein)
• run_prolif add ProLIF argument parallel_strategy (default chunk)
• run_prolif speed up --water_bridge analysis : restrict waters to an updating per-frame selection near the ligand instead of converting every water every frame. New argument water_cutoff (default 8 Å, automatically widened for higher water_bridge_order; set 0 to use all waters)
• run_prolif add argument ligand_sdf: optional reference .sdf/.mol bond-order template for the ligand (usually unnecessary — ligand chemistry is inferred from the topology and the interpreted SMILES is logged for verification)
• run_prolif/prolif_drawmap show the occupancy percentage above each dot in prolif_output_occupancyX.png by default (use --no-show_percentage to disable)
  Fixed bugs:
• change ActiveSite RMSD definition: the pocket is now the complete backbone of protein residues within active_site_dist of the ligand heavy atoms (previously only backbone atoms lying within the cutoff), so values are not directly comparable to earlier versions
• RMSF is now computed on Cα atoms (gmx rmsf selects C-alpha, used for both the fit and the per-residue output) instead of Protein
• fix last_frame.pdb: the final frame was selected by passing the number of frames to gmx trjconv -dump, which interprets the value as a time in ps and extracted an early frame instead of the last one; the actual final trajectory time is now parsed from gmx check and passed to -dump (a clear error is raised if it cannot be determined)
• last_frame.pdb is now written as a visualization-ready snapshot: molecules are made whole across periodic boundaries and the protein–ligand–cofactor complex is centered in a compact unit cell (gmx trjconv -pbc mol -center -ur compact), with the full System (water and ions included) written
• add start_frame.pdb - the first production frame written with the same visualization-ready treatment as last_frame.pdb (-pbc mol -center -ur compact, full System), for visual comparison of the start and end of the simulation
• improve RMSD plot readability: move the legend outside the axes (below, left-aligned) so it no longer overlaps the curves, and draw lines with transparency (new alpha argument to plot_rmsd, default 0.7) so overlapping RMSD curves remain visible