Example input

• OPTIM test sets from the group gitlab TESTS directory.
  

  Tests for can be found in subdirectories DOUBLE_ENDED, FRQS, GEOPT, PATH for a range of systems: AMBER12, AMBER9, AMBER9_genrigid, BLJ60, BOWMAN, LJ38, SD, TIP4P_genrigid, TIP4P_rbaa, TTM3, VARIABLES.
  

• Stillinger-Weber potential with periodic boundaries.
  
• Examples for Stillinger-Weber potential with periodic boundary conditions. The 64fixed directory corresponds to 64 atoms in a fixed cell and contains subdirectories illustrating Cartesian and fractional coordinate input. The 64variable directory contains subdirectories for variable cell optimisations at pressures of 0 GPa and 12 GPa, which both use fractional coordinate input.
  
  
• Examples illustrating the CHECKPERM keyword.
  

  OPTIM input and output for 1bdd, NmHR, and M1NT R134N mutant.
  

• ReaxFF force field.
  

  Force field parameters for ReaxFF from Development of ReaxFF Reactive Force Field for Aqueous Iron-Sulfur Clusters with Applications to Stability and Reactivity in Water. Journal of chemical information and modeling (2021) 61, 1204
  

• Example input and output for influenza matrix protein 1 N-terminal domain with AMBER ff19SB. Four examples for minima separated by 1, 2, 3, and 6 transition states.
  

  1 step OPTIM input and output.
  

  2 steps OPTIM input and output.
  

  3 steps OPTIM input and output.
  

  6 steps OPTIM input and output.
  

• Example input and output for a water molecule minimisation with ORCA.
  

  OPTIM input and output.
  

• Example input and output for cyclohexane boat to chair with ORCA.
  

  OPTIM input and output.
  

• Example input and output for bulk silicon double-ended search with DFTBP.
  

  OPTIM input and output.
  

• Example input and output for trpzip with a one step pathway between two local minima with AMBER20.
  

  OPTIM input and output.
  

• A detailed tutorial for exploring the energy landscape of peptides using GMIN/OPTIM/PATHSAMPLE programs interfaced with AMBER.
  Steps to obtain and further analyse the heat capacity as a function of temperature are also explained, as well as basin-hopping parallel tempering. The entire procedure is explained in steps starting from the sequence only.
• Example GPU and CPU input and output for ubiquitin with a one step pathway between two local minima with AMBER20. QCI is not really needed for this initial initial interpolation, but QCI and DNEB only examples are both provided here.
  

  GPU DNEB run, GPU QCI run, CPU DNEB run, CPU QCI run
  

• Example for water hexamer MBPOL intermolecular potential.
  

  Pathsample run for water hexamer including some OPTIM input and output.
  

• Example for QUIP Carbon GAP 20 potential.
  

  basic minimisation for seven carbon atoms; new GAP potential.
  

• Example for QUIP Carbon GAP 20 potential.
  

  basic minimisation for seven carbon atoms; old GAP potential.
  

• C₆₀ example for QUIP Carbon GAP 20 potential.
  

  transition state search and pathway for buckminsterfullerene.
  

• Example for QUIP silicon potential.
  

  basic minimisation for seven silicon atoms.
  

• Example for CASTEP on nest: pentaprismane.
  

  transition state search with path run in subdirectory.
  

• Example for the Tapered ReaxFF interface; a double-ended pathway search for a C₂₄ cluster with a tiny gap on a very noisy PES.
  

  NEWCONNECT run with 50 DNEB images.
  

• Another example for the Tapered ReaxFF interface; a double-ended pathway search for a C₂₄ cluster with a tiny gap on a very noisy PES.
  

  NEWCONNECT run with 200 DNEB images.
  

• Example for the xtb interface; a double-ended pathway search for the water₂₁ cluster.
  

  NEWCONNECT run with 2 DNEB images.
  

• Example input and output for Tryptophan Zipper 1 (PDB 1LE0) with AMBER12 and the NapShift Hybrid Potential.
  

  2 NEWCONNECT cycles using QCI with AMBER12 and NapShift.
  

• Example input and output for a 640-atom RNA with AMBER.
  

  NEWCONNECT run using QCI with 5 images.
  

• Example input and output for a coarse-grained protein model.
  

  S6 pathway for a protein bead model with 97 sites.
  

  QCI interpolation for a protein bead model with 92 sites, avoiding chain crossings
  

• REDOPATH and REDOPATHXYZ examples to create complete path.xyz files for visualisation for AMBER.

  REDOPATH for tau protein AMBER 12.
  

  REDOPATHXYZ for tau protein AMBER9.
  

  REDOPATHXYZ for tau protein AMBER12.
  
  

• Example RIGIDINIT input and output for rigidification of TRP groups in trpzip. A two step pathway between local minima is characterised using both DNEB and QCI.
  

  trpzip AMBER local rigid body pathway with DNEB.
  

  trpzip AMBER local rigid body pathway with QCI.
  

• Example VASP input from Dr Bora Karasulu. This is a test for Li ion migration pathways in a supercell with 63 Li atoms (one vacancy) and 32 sulphur atoms.
  

  Input and output for VASP runs.
  

• Example PUSHOPT runs for a machine learning example with BFGSTS and SEARCH 0 transition state seacrhes combined with pathways calculated using LBFGS, SEARCH 0, and Page-McIver steepest-descent.
  

  Input and output for PUSHOPT runs.
  

• Example MULTIJOB run for tighter minimisation of LJ₃₁ minima in either xyz or coordinate list format.
  

  Input and output for MULTIJOB run using either xyz or sequential coordinates. Ten structures are converged more tightly using BFGSMIN.
  

• Example MULTIJOB run for double-ended connections of LJ₃₁ minima in either xyz or coordinate list format using NEWCONNECT.
  

  Input and output for MULTIJOB run using either xyz or sequential coordinates. Ten pairs of minima are connected using the NEWCONNECT keyword after the initial pair. A path.info file containing data for each transition state is created.
  

• Example connection for bulk LJ₈₆₄ with periodic boundary conditions.
  

  NEWCONNECT run using keywords PERMDISTINIT and ATOMMATCHINIT with hybrid eigenvector-following transition state searches using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction and LBFGS energy minimisation for the pathways. Permutational alignment is only attempted for the end points at the beginning of the run.
  

• Examples for the LJ₃₈ cluster connecting the lowest and second-lowest minima.
  These are all double-ended transition state searches and pathways using NEWCONNECT.
  

  Parameters that produce a connected path from a single DNEB cycle.
  Parameters that produce a connected path more efficiently, in six DNEB cycles.
  A connected pathway where the two minima are not in the optimal permutational alignment. 56 DNEB cycles are needed and the path has 207 transition states.
  
  

• Example for the LJ₃₈ cluster calculating paths for an index 2 saddle following the two modes with negative eigenvalues.
  

  Input and output files for both modes.
  
  

• Example calculations for the LJ₃₈ cluster.
  Double-ended transition state searches and pathways using NEWCONNECT. In each case a path.info file is produced using the DUMPALLPATHS keyword for use by the PATHSAMPLE program.
  

  Eigenvector-following transition state searches with LBFGS energy minimisation for the pathways
  Hybrid eigenvector-following transition state searches with calculation of the eigenvalue and eigenvector corresponding to the uphill direction performed using an iterative approach for the full Hessian and LBFGS energy minimisation for the pathways
  Hybrid eigenvector-following transition state searches with calculation of the eigenvalue and eigenvector corresponding to the uphill direction using the DSYEVR routine and LBFGS energy minimisation for the pathways
  Hybrid eigenvector-following transition state searches using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction and LBFGS energy minimisation for the pathways
  
  Single-ended transition state searches and pathways. In each case a path.info file is produced using the DUMPALLPATHS keyword for use by the PATHSAMPLE program, in addition to the energy as a function of integrated path length (EofS) and a movie of the complete path (points.path.xyz)
  

  Eigenvector-following transition state search with Runge-Kutta energy minimisation for the pathway
  Eigenvector-following transition state search with Page-McIver energy minimisation for the pathway
  Eigenvector-following transition state search with LBFGS energy minimisation for the pathway
  Eigenvector-following transition state search with eigenvector-following energy minimisation for the pathway
  Eigenvector-following transition state search with Bulirsch-Stoer energy minimisation for the pathway
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction using the DSYEVR routine then Bulirsch-Stoer energy minimisation for the pathway
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction using the DSYEVR routine then Runge-Kutta energy minimisation for the pathway
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction using the DSYEVR routine then Page-McIver energy minimisation for the pathway
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction using the DSYEVR routine then LBFGS energy minimisation for the pathway
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction using the DSYEVR routine then eigenvector-following energy minimisation for the pathway
  Hybrid eigenvector-following transition state search using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction then Bulirsch-Stoer energy minimisation for the pathway
  Hybrid eigenvector-following transition state search using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction then Runge-Kutta energy minimisation for the pathway
  Hybrid eigenvector-following transition state search using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction then Page-McIver energy minimisation for the pathway
  Hybrid eigenvector-following transition state search using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction then LBFGS energy minimisation for the pathway
  Hybrid eigenvector-following transition state search using a gradient-only variational approach for the eigenvalue and eigenvector corresponding to the uphill direction then eigenvector-following energy minimisation for the pathway
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction performed using an iterative approach for the full Hessian and LBFGS energy minimisation for the pathways
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction performed using an iterative approach for the full Hessian and Bulirsch-Stoer energy minimisation for the pathways
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction performed using an iterative approach for the full Hessian and Runge-Kutta energy minimisation for the pathways
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction performed using an iterative approach for the full Hessian and Page-McIver energy minimisation for the pathways
  Hybrid eigenvector-following transition state search with calculation of the eigenvalue and eigenvector corresponding to the uphill direction performed using an iterative approach for the full Hessian and eigenvector-following energy minimisation for the pathways
  

• Factorised superposition approach (FSA) for computing binding free energies (AMBER9)
• REDOPATH movie-making run for a trapped ion cluster
• NEWCONNECT run for TIP4P water octamer
• NEWCONNECT run for TIP4P water octamer using the angle-axis scheme
• Harmonic normal mode analysis on a folded conformation of the met-enkephalin peptide (AMBER9)
• Harmonic normal mode analysis on a folded conformation of the met-enkephalin peptide (CHARMM19)
• Transition state benchmark for a amyloidogenic GNNQQNY peptide rearrangement (CHARMM19)
• Transition state benchmark for a tryptophan zipper 1 rearrangement (AMBER9)
• The MSB test sets for pathways in trpzip1 and GNNQQNY with CHARMM and AMBER. These compressed tar files contain a total of 2100 pathways, including OPTIM input and output. Each double-ended connection attempt is performed with quasi-continuous interpolation (QCI), DNEB interpolation, and the seven interpolation schemes described in M.S. Bauer, B. Strodel, S.N. Fejer, E.F. Koslover and D.J. Wales, J. Chem. Phys., 132, 054101 (2010).

  AMBER/GNNQQNY/TS1
  AMBER/GNNQQNY/TS2
  AMBER/GNNQQNY/TS3
  AMBER/trpzip/TS1
  AMBER/trpzip/TS2
  AMBER/trpzip/TS3
  CHARMM/GNNQQNY/TS1
  CHARMM/GNNQQNY/TS2
  CHARMM/GNNQQNY/TS3
  CHARMM/trpzip/TS1
  CHARMM/trpzip/TS2
  CHARMM/trpzip/TS3