Molecules Header for YAML Files

Everything under the molecules defines what molecules are in your systems. You can specify your own molecule names. Because of this user defined names in the syntax examples are marked as {UserDefinedMolecule}.

You can define as many {UserDefinedMolecule} as you like. These molecules will be used in other YAML headed sections.

Unlike the primary options for YAML files, many of these settings are optional and do nothing if not specified. The mandatory/optional of each setting (and what conditionals), as well as the default behavior of each setting is explicitly stated in the setting’s description.

All of the molecules will be built, even if they are not used in a later system, so ensure your molecules do not have errors or are commented out.


Specifying Molecules

filepath

molecules:
  {UserDefinedMolecule}:
    filepath: benzene.pdb

Filepath of the molecule. Path is relative to the directory the YAML script is in. Depending on what type of molecule you pass in (ligand vs. protein) will determine what mandatory arguments are required. For small molecules, you need a way to specify the charges, proteins however can rely on built in force field parameters.

MANDATORY but exclusive with smiles and name

Valid Filetypes: PDB, mol2, sdf, cvs

Note: If CVS is specified and there is only one moleucle, the first column must be a SMILES string. If multiple molecules are to be used (for the Combinatorial <combinatorial> ability), then each row is its own molecule where the second column is the SMILES string.

smiles

molecules:
  {UserDefinedMolecule}:
    smiles: c1ccccc1

YANK can process SMILES stings to build the molecule as well. Usually only recommended for small ligands. Requires that the OpenEye Toolkits are installed.

MANDATORY but exclusive with filepath and name

name

molecules:
  {UserDefinedMolecule}:
    name: benzene

YANK can process raw molecule name if the OpenEye Toolkits are installed

MANDATORY but exclusive with filepath and smiles

strip_protons

molecules:
  {UserDefinedMolecule}:
    strip_protons: no

Specifies if LEaP will re-add all hydrogen atoms. This is helpful if the PDB contains atom names for hydrogens that AMBER does not recognize. Primarily for proteins, not small molecules.

OPTIONAL and defaults to no

Valid Options: [no]/yes

pdbfixer

molecules:
  {UserDefinedMolecule}:
    pdbfixer:
      replace_nonstandard_residues: no
      remove_heterogens: none
      add_missing_residues: no
      add_missing_atoms: none
      apply_mutations:
        mutations: T315I
        chain_id: A

Specifies whether PDBFixer should be used to modify the molecule. Can only be used on proteins, on files with .pdb file extensions.

Replacing nonstandard residues

molecules:
  {UserDefinedMolecule}:
    pdbfixer:
      replace_nonstandard_residues: yes

Options are:

  • yes : replace nonstandard amino acid residues will be replaced with one of the 20 standard amino acids according to the scheme used by PDBFixer
  • no : don’t replace residues

OPTIONAL with default value of no

Valid Options: [no]/yes

Removing heterogens

molecules:
  {UserDefinedMolecule}:
    pdbfixer:
      remove_heterogens: all

Valid options: [none] | water | all

This directs PDBFixer to remove some heterogen residues from the PDB file:

  • all : all heterogens (residues that are not one of the standard amino acids) will be removed
  • water : only water residues will be removed
  • none : no residues will be removed

OPTIONAL with default value of none

Valid Options: [none]/water/all

Adding missing residues and atoms

Add missing atoms (including entire residues, loops, and termini).

WARNING: PDBFixer uses a very simple approach to adding missing residues that will only produce sensible geometries in the simplest of cases. Use this option with caution.

molecules:
  {UserDefinedMolecule}:
    pdbfixer:
      add_missing_residues: yes
      add_missing_atoms: heavy
      ph: 7.4

add_missing_residues specifies whether missing residues should be added

  • no : only missing atoms in existing residues will be added (DEFAULT)
  • yes : missing residues specified in SEQRES will be added

add_missing_atoms specifies which detected missing atoms should be added:

  • none : no missing atoms will be added
  • heavy : only heavy atoms will be added (DEFAULT)
  • hydrogens : add only hydrogens (not recommended)
  • all : all missing atoms (including hydrogens) will be added (not recommended)

ph specifies the pH to be used for adding hydrogens (default: 7.4)

OPTIONAL

Mutations

Make the directed mutations to amino acid residues.

molecules:
  {UserDefinedMolecule}:
    pdbfixer:
      apply_mutations:
        mutations: T315I
        chain_id: A

Mutations are specified using the format <original-one-letter-code><resid><new-one-letter-code>, with the character / being an optional separator if multiple mutations are desired.

The initial PDB file numbering is used for residue identifier resid.

Examples for specifying mutations::

  • T315I : a single mutation that changes Thr at resid 315 to Ile
  • L858R/T790M : a double mutation

If chain_id is not specified, it defaults to null (no chain designator).

PDBFixer is applied after strip_protons if both are requested.

OPTIONAL

modeller

molecules:
  {UserDefinedMolecule}:
    modeller:
      apply_mutations:
        mutations: T315I
        chain_id: A

Specifies whether modeller should be used to model in mutations. Can only be used on proteins, on files with .pdb file extensions.

This feature requires the Modeller from the Sali Lab, which can be fetched from Omnia’s Conda channel. You will need to provide your own license however.

Mutations

Make the directed mutations to amino acid residues.

molecules:
  {UserDefinedMolecule}:
    modeller:
      apply_mutations:
        mutations: T315I
        chain_id: A

Mutations are specified using the format <original-one-letter-code><resid><new-one-letter-code>. Only single mutations are supported currently.

The initial PDB file numbering is used for residue identifier resid.

Examples for specifying mutations::

  • T315I : a single mutation that changes Thr at resid 315 to Ile

chain_id can be either an integer or string. Strings are case sensitive! If chain_id is not specified, it defaults to 0 (the first chain in the file will be selected).

Examples for specifying chain_id: * 0 : the first chain in the file will be selected * B : chain B will be selected

Modeller is applied after both strip_protons and ‘’pdbfixer’’ if either (or both) are requested.

OPTIONAL

select

molecules:
  {UserDefinedMolecule}:
    filepath: clinical-kinase-inhibitors.csv
    antechamber:
        charge_method: bcc
    select: !Combinatorial [0, 3]

The “select” keyword works the same way if you specify a pdb, mol2, sdf, or cvs file containing multiple structures. select has 3 modes:

  1. select: all includes all the molecules in the given file.
  2. select: <Integer> picks the molecule in the file with index <Integer>
  3. select: !Combinatorial: <List of Ints> pick specific indices in the file. See Combinatorial options for more information.

Indexing starts at 0.

OPTIONAL with default value of all

Valid Options: [all]/<Integer>/<Combinatorial List of ints>


Assigning Missing Parameters

antechamber

molecules:
  {UserDefinedMolecule}:
    filepath: benzene.mol2
    antechamber:
      charge_method: bcc
      net_charge = -1

Pass the molecule through AmberTools ANTECHAMBER to assign missing parameters such as torsions and angle terms.

charge_method is a required sub-directive which allows assigning missing charges to the molecule. It is either given a known charge method to ANTECHAMBER method or null to skip assigning charges. The later is helpful when you already have the charges, but are missing other parameters.

The net_charge keyword passed to antechamber has no effect if the charge_method is null (e.g., when using OpenEye charges) or if epik is used to enumerate protonation states. In these cases, the net charge is handled automatically.

OPTIONAL

PARTIALLY EXCLUSIVE If you have acess to the OpenEye toolkits and want to use them to assign partial charges to the atoms through the openeye command, then you should set charge_method to null. ANTECHAMBER can still get the other missing parameters such as torsions and angles.

OPTIONALLY SUPERSEDED by leap or the leap argument in systems. If the parameter files you feed into either leap argument have the charges and molecular parameters already included (such as standard protein residues in many force fields), then there is no need to invoke this command. If the force fields you give to the leap commands are missing parameters though, you should call this.

openeye

molecules:
  {UserDefinedMolecule}:
    filepath: benzene.mol2
    openeye:
      quacpac: am1-bcc

Use the OpenEye Toolkits if installed to determine molecular charge. Only the current options as shown are permitted and must be specified as shown.

OPTIONAL

PARTIALLY EXCLUSIVE If you want to use antechamber to assign partial charges, do not use this command. However, if you want to use antechamber to only get other missing parameters such as torsions and angles, use this command but set charge_method to null in antechamber

OPTIONALLY SUPERSEDED by leap or the leap argument in systems. If the parameter files you feed into either leap argument have the charges and molecular parameters already included (such as standard protein residues in many force fields), then there is no need to invoke this command. If the force fields you give to the leap commands are missing partial charges though, you should call this.


Assigning Extra Information

leap

molecules:
  {UserDefinedMolecule}:
    leap:
      parameters: [mymol.frcmod, mymol.off]

Load molecule-specific force field parameters into the molecule. These can be created from any source so long as leap can parse them. It is possible to assign partial charges with the files read in this way, which would supersede the options of antechamber and openeye.

This command has only one mandatory subargument parameters, which can accept both single files as a string, or can accept a comma separated list of files enclosed by [ ]. Filepaths are relative to either the AmberTools default paths or to the folder the YAML script is in.

Note: Proteins do not necessarily need this command if the force fields given to the leap argument in systems will fully describe them.

OPTIONAL

epik

molecules:
  {UserDefinedMolecule}:
     epik:
       select: 0
       ph: 7.6
       ph_tolerance: 3.0
       tautomerize: no

Run Schrodinger’s tool Epik with to select the most likely protonation state for the molecule in solution. Parameters in this call are direct reflections of the function to invoke epik from OpenMolTools. Each of the parameters in this list (with the exception of select) are optional.

We note that the option ph_tolerance set to a value here of 3.0, the pH range which will be searched will be pH +- 3.0, which is a 7 log unit range, which may take a some time to enumerate, although will likely be less than the simulation overall. Should you feel this time is too long, you might consider reducing the ph_tolerance.

OPTIONAL

regions

molecules:
  {UserDefinedMolecule}:
     regions:
        {UserDefinedRegion}: region_string
        ...

Define molecular regions in the molecule which can be used in upcoming features such as defining restraint regions in more general ways, or specific atom subsets you want to track through the yank.yank.Topography object which is stored as part of the simulation’s metadata, accessible through yank.sampling.Reporter.

Any number of user defined regions can be specified for every molecule, so long as their name is unique between all molecules which ultimately wind up in a system. E.g. If you have 2 ligands you want to bind to a receptor in a combinatorial setup, both ligands can have a region named “my_region” since they will never be in the same system together. However, the receptor cannot have a region named “my_region” as well, as that will be ambiguous as to which region, ligand or receptor, to define.

The regions apply only to the molecule the regions section is under, so even if the atom index changes in the yank.yank.Topography, the atomic indices defined in the region entry will be converted.

The region definition supports multiple selection formats:

  • DSL String: An MDTraj DSL string which identifies will identify a region.
  • Future Ability SMARTS String: Molecular selection format similar to regular expression for strings, but for molecules instead. This feature is not in yet, but is planned. The regions framework is the pre-cursor to this feature. See Daylight’s website for more information on SMARTS.
  • List of Ints: Select atoms by integers, this applies only to the final system, so numbers will probably not align with the atom numbers from the input files.
  • Single Int: Same as the list of ints, but with a single entry, subject to same rules

OPTIONAL