trypzip

Challenge: improving an initial trpyzip path

<img src="trypzip_endpoints.png" width="100%", height="100%">

Before you attempt this challenge, it is highly recommended that you first run through the OPTIM and PATHSAMPLE examples for the tetra-ALA peptide here.

The 12 residue tryptophan zipper peptide has been extensively used as a model system for protein folding and force field benchmarking. The beta-hairpin it forms is stabilised by the packing of four tryptophan rings in the native state, forming a 'steric zipper' motif.

In this directory you will find a PATHSAMPLE database containing only the minima and transition states along an initial discrete path from an unfolded to the native state as shown above. The pathway has not been at all optimised and as such is likely to be overly long and contain high barriers.

Your job is to expand this PATHSAMPLE database using a range of techniques (some of which are outlined in pathdata_annotated) to further sample the energy landscape of this peptide, improving the fastest path and converging the folding rate as you go.

There are keywords in these input files that you will not necessarily have seen before so check them on the PATHSAMPLE website.

Unlike other examples, this challenge is not designed to be something you can complete in an hour so don't be discouraged!

Requirements

In order to successfully attempt this challenge, the following need to be in your PATH:

• a PATHSAMPLE binary
• an A12OPTIM binary
• a disconnectionDPS binary

Directory contents

Both this directory and the backup in ./input contain all the files you need to run PATHSAMPLE to expand the database

As PATHSAMPLE acts as a driver for A12OPTIM (i.e. it starts A12OPTIM jobs), there are also A12OPTIM input files present.

PATHSAMPLE input files

• pathdata - Every PATHSAMPLE job requires a pathdata file containing the keywords used to specify what we would like the run to achieve. This example will require us to run it twice with different keywords, hence there are two sections at the bottom of pathdata, one initially commented out (starting with '! ')

• pathdata_annotated - The PATHSAMPLE keywords we are using in this example are detailed in pathdata_annotated. This file is not required, it is provided for reference only. For information on the full set of PATHSAMPLE keywords available, check the PATHSAMPLE website

• min.A - Defines which minima should be considered part of the A group - the products or reactants depending on how we specify the DIRECTION in pathdata

• min.B - Defines which minima should be considered part of the B group - the products or reactants depending on how we specify the DIRECTION in pathdata

• perm.allow - Specifies which atoms in trypzip should be considered identical with respect to permutational isomerisation. This ensures that we do not consider two minima that differ by a rotation of a methyl group to be different. It is possible to generate these files automatically from a PDB using the Python script here:

SCRIPTS/make_perm.allow/perm-pdb.py file.pdb AMBER 

• perm.allow_annotated - Contains details of how the perm.allow groups are constructed. For more information, see the OPTIM website

Both PATHSAMPLE and A12OPTIM need to be able to distinguish permutational isomers and so both require perm.allow to be present.

PATHSAMPLE database files

• min.data - Contains the energy, logarithm of the product of positive Hessian eigenvalues, symmetry and moments of inertia for each minimum. A minimum is identified by its line number in this file

• ts.data - Contains the energy, logarithm of the product of positive Hessian eigenvalues, symmetry, minima numbers that it connects and moments of intertia for each transition state. A transition state is identified by its line number in this file

• points.min - Contains the coordinates for each minimum in a binary format to keep the file size low

• points.ts - Contains the coordinates for each transition state in the same binary format

OPTIM input files

• odata.connect - Contains the A12OPTIM keywords used for jobs started by PATHSAMPLE.

• odata.connect_annotated - The A12OPTIM keywords present in odata.connect are detailed in odata.connect_annotated. For information on the full set of keywords available, check the OPTIM website

• coords.prmtop - The symmetrised (see the note below!) AMBER topology file for trypzip using parameters from the AMBER ff99SB force field

• coords.inpcrd - Coordinates for the trypzip atoms in our system in AMBER restart format. These are only used to allocate arrays during setup and the coordinates themselves are overwritten with those in start.X by PATHSAMPLE automatically as it starts A12OPTIM jobs in the next example

• start - Placeholder trypzip coordinates that are also overwritten with those in start.X by PATHSAMPLE.

• min.in - The AMBER force field parameters to use to calculate the energy and gradient.

• min.in_annotated - Not used during the run. Contains additional information about the AMBER parameters used in this exammple. See the AMBER manual for more information

• perm.allow - discussed above

• perm.allow_annotated - discussed above

disconnectionDPS input files

• dinfo - Contains the keywords that control the appearence of the disconnectivty graph produced when disconnectionDPS is run in the current directory

• dinfo_annotated - The disconnectionDPS keywords used in this example are detailed in dinfo_annotated. For more information and a full keyword listing, see the top of the disconnectionDPS.f90 source file, available in the source tar file on the Wales Group website

Getting started

Before you start, take a minute to look through pathdata_annotated and odata.connect_annotated and make sure you understand roughly the purpose of each keyword. You may also need to slightly alter your pathdata file to ensure that the EXEC keyword points to a valid A12OPTIM binary.

The PATHSAMPLE input in pathdata is initially set up to perform a Dijkstra analysis, returning the fastest path: <img src="trypzip_initial_path.png" width="100%", height="100%">

The first thing you need to do is decide which strategy you will use to select minima to connect to most efficiently expand the database.

pathdata contains a range of options with the parameters intentionally left blank:

! STEP 2: decide on a strategy to expand the database
PAIRLIST 1
! DUMMYRUN

! Option A - SHORTCUT x
! SHORTCUT x
! CYCLES 10

! Option B - SHORTCUT x BARRIER
! SHORTCUT x BARRIER
! CYCLES 10

! Option C - UNTRAP einc thresh (edeltamin) (elowbar) (ehighbar)
! UNTRAP einc thresh (edeltamin) (elowbar) (ehighbar)
! CYCLES 10

! Option D - FREEPAIRS x1 x2 x3
! FREEPAIRS x1 x2 x3
! CYCLES 10

! Option E - CONNECTREGION min1 min2
! CONNECTREGION min1 min2
! CYCLES 10

! Option F - ADDPATH path.info.x
! ADDPATH path.info.toadd
! CYCLES 0

The approach you use and the parameters you choose should be informed by the system you are studying. Without giving too much away, here are a few tips:

• look at the initial path carefully

• you can run disconnectionDPS at any time (remember IDENTIFY?) to make a disconnectivity graph and view it with gv

• you can extract minima from the PATHSAMPLE database using EXTRACTMIN minid in pathdata

• if you extract two minima, you can connect them manually with A12OPTIM as for tetra-ALA here and then add the resulting pathway back into the database using ADDPATH

• the DUMMYRUN keyword is your friend, especially when tuning parameters for efficient selection - consider uncommenting it!

Remember - this is a challenge so don't worry if you don't make much progress - it's designed to be hard!

Finally, to give you an idea of the task ahead, here are the disconnectivity graphs for the initial path on the left and the converged landscape on the right - good luck! <img src="tree_trypzip_initial_vs_converged.png" width="100%", height="100%">