.. highlight:: cpp

.. doctest::
  :hide:

  >>> import gemmi

Structure analysis
##################

.. _neighbor_search:

Neighbor search
===============

Fixed-radius near neighbor search is usually implemented using
the `cell lists <https://en.wikipedia.org/wiki/Cell_lists>`_ method,
also known as binning, bucketing, or the cell technique
(or cubing -- as it was called in an
`article <https://web.stanford.edu/class/sbio228/public/readings/Molecular_Simulation_I_Lecture4/Levinthal_SCIAM_66_Protein_folding.pdf>`_ from 1966).
The method is simple. The unit cell (or the area where the molecules are
located) is divided into small cells. The size of these cells depends
on the search radius. Each cell stores a list of atoms in its area;
these lists are used for fast lookup of atoms.

In Gemmi the cell technique is implemented in a class named `NeighborSearch`.
The implementation works with both crystal and non-crystal systems and:

* handles crystallographic symmetry (including non-standard settings with
  origin shift that are present in a couple of hundreds of PDB entries),
* handles strict NCS (MTRIX record in the PDB format that is not "given";
  in mmCIF it is the _struct_ncs_oper category),
* handles alternative locations (atoms from different conformers are not
  neighbors),
* can find neighbors any number of unit cells apart; surprisingly,
  molecules from different and non-neighboring unit cells can be
  in contact, either because of the molecule's shape (a single chain can be
  :ref:`longer than four unit cells <long_chain>`) or because of
  the non-optimal choice of symmetric images in the model
  (some PDB entries even have links between chains more than
  10 unit cells away which cannot be expressed in the 1555 type of notation).

Note that while an atom can be bonded with its own symmetric image,
it sometimes happens that an atom meant to be on a special position
is slightly off, and its symmetric images represent the same atom (so after
expanding the symmetry, we may have four nearby images, each with occupancy 0.25).
Such images will be returned by the NeighborSearch class as neighbors
and need to be filtered out by the users.

The NeighborSearch constructor divides the unit cell into bins.
For this, it needs to know the search radius for which we optimize bins,
as well as the unit cell. Since the system may be non-periodic,
the constructor also takes the model as an argument -- it is used to
calculate the bounding box for the model if there is no unit cell. The
reference to the model is stored and is also used if `populate()` is called.
The C++ signature (in `gemmi/neighbor.hpp`) of the constructor is::

  NeighborSearch::NeighborSearch(Model& model, const UnitCell& cell, double radius)

The cell lists need to be populated with items either by calling::

  void NeighborSearch::populate(bool include_h=true)

or by adding individual chains::

  void NeighborSearch::add_chain(const Chain& chain, bool include_h=true)

or by adding individual atoms::

  void NeighborSearch::add_atom(const Atom& atom, int n_ch, int n_res, int n_atom)

where `n_ch` is the index of the chain in the model, `n_res` is the index
of the residue in the chain, and `n_atom` is the index of the atom
in the residue.

An example in Python:

.. doctest::

  >>> import gemmi
  >>> st = gemmi.read_structure('../tests/1pfe.cif.gz')
  >>> ns = gemmi.NeighborSearch(st[0], st.cell, 5).populate(include_h=False)

Here we do the same using `add_chain()` instead of `populate()`:

.. doctest::

  >>> ns = gemmi.NeighborSearch(st[0], st.cell, 5)
  >>> for chain in st[0]:
  ...     ns.add_chain(chain, include_h=False)

And again the same, with complete control over which atoms are included:

.. doctest::

  >>> ns = gemmi.NeighborSearch(st[0], st.cell, 5)
  >>> for n_ch, chain in enumerate(st[0]):
  ...     for n_res, res in enumerate(chain):
  ...         for n_atom, atom in enumerate(res):
  ...             if not atom.is_hydrogen():
  ...                 ns.add_atom(atom, n_ch, n_res, n_atom)
  ...



All these functions store `Mark`\ s in cell-lists. A Mark contains the
position of an atom's symmetry image and indices that point to the original
atom. Searching for neighbors returns Marks, from which we can obtain
original chains, residues, and atoms.

NeighborSearch has a couple of functions for searching.
The first one takes an atom as an argument::

  std::vector<Mark*> NeighborSearch::find_neighbors(const Atom& atom, float min_dist, float max_dist)

.. doctest::

  >>> ref_atom = st[0].sole_residue('A', gemmi.SeqId('3')).sole_atom('P')
  >>> marks = ns.find_neighbors(ref_atom, min_dist=0.1, max_dist=3)
  >>> len(marks)
  6

`find_neighbors()` checks the atom's altloc and only considers
atoms from the same conformation as potential neighbors.
In particular, if the altloc is empty, all atoms are considered.
A positive `min_dist` in the `find_neighbors()` call prevents
the atom whose neighbors we are searching for from being included
in the results (since the distance of an atom to itself is zero).

The second one takes the position and altloc as explicit arguments::

  std::vector<Mark*> NeighborSearch::find_atoms(const Position& pos, char altloc, float min_dist, float radius)

.. doctest::

  >>> point = gemmi.Position(20, 20, 20)
  >>> marks = ns.find_atoms(point, '\0', radius=3)
  >>> len(marks)
  7

To find only the nearest atom (regardless of altloc), use::

  Mark* find_nearest_atom(const Position& pos, float radius=INFINITY)

.. doctest::

  >>> ns.find_nearest_atom(point)
  <gemmi.NeighborSearch.Mark 3 of atom 0/7/9 element C>

All the above functions can search in a radius bigger than the radius passed
to the NeighborSearch constructor, but it requires checking more cells
(125+ instead of 27), which is usually not optimal.
On the other hand, it is usually also not optimal to use big cells
(radius ≫ 10Å). And very small ones (radius < 4Å) are also inefficient.

In C++ you may use a low-level function that takes a callback
as an argument (usage examples are in the source code)::

  template<typename T>
  void NeighborSearch::for_each(const Position& pos, char altloc, float radius, const T& func, int k=1)

Cell-lists contain `Mark`\ s. When searching for neighbors you get references
(in C++ -- pointers) to these marks.
`Mark` has a number of properties: `x`, `y`, `z`,
`altloc`, `element`, `image_idx` (index of the symmetry operation
that was used to generate this mark, 0 for identity),
`chain_idx`, `residue_idx` and `atom_idx`.

.. doctest::

  >>> mark = marks[0]
  >>> mark
  <gemmi.NeighborSearch.Mark 11 of atom 0/7/2 element O>
  >>> mark.pos
  <gemmi.Position(21.091, 18.2279, 17.841)>
  >>> mark.altloc
  '\x00'
  >>> mark.element
  gemmi.Element('O')
  >>> mark.image_idx
  11
  >>> mark.chain_idx, mark.residue_idx, mark.atom_idx
  (0, 7, 2)

The references to the original model and to atoms are not stored.
`Mark` has a method `to_cra()` that needs to be called with `Model`
as an argument to get a triple of Chain, Residue and Atom::

  CRA NeighborSearch::Mark::to_cra(Model& model) const

.. doctest::

  >>> cra = mark.to_cra(st[0])
  >>> cra.chain
  <gemmi.Chain A with 79 res>
  >>> cra.residue
  <gemmi.Residue 8(DC) with 19 atoms>
  >>> cra.atom
  <gemmi.Atom OP2 at (1.4, 15.9, -17.8)>

Note that `mark.pos` can be the position of a symmetric image of the atom.
In this example the "original" atom is in a different location:

.. doctest::

  >>> mark.pos
  <gemmi.Position(21.091, 18.2279, 17.841)>
  >>> cra.atom.pos
  <gemmi.Position(1.404, 15.871, -17.841)>

The symmetry operation that relates the original position and its
image is composed of two parts: one of symmetry transformations
in the unit cell and a shift of the unit cell that we often
call here the PBC (periodic boundary conditions) shift.

The first part is stored explicitly as `mark.image_idx`.
The corresponding transformation is:

.. doctest::

  >>> ns.get_image_transformation(mark.image_idx)  # doctest: +ELLIPSIS
  <gemmi.FTransform object at 0x...>
  >>> _.mat
  <gemmi.Mat33 [1, -1, 0]
               [0, -1, 0]
               [0, 0, -1]>

To find the full symmetry operation we need to determine the nearest
image under PBC:

.. doctest::

  >>> st.cell.find_nearest_pbc_image(point, cra.atom.pos, mark.image_idx)
  <gemmi.NearestImage 12_665 in distance 3.00>

To calculate only the distance to the atom, you can use the same function
with `mark.pos` and symmetry operation index 0. `mark.pos` represents
the position of the atom that has already been transformed by the symmetry
operation `mark.image_idx` (and shifted into the unit cell).

.. doctest::

  >>> st.cell.find_nearest_pbc_image(point, mark.pos, 0)
  <gemmi.NearestImage 1_555 in distance 3.00>
  >>> _.dist()  # doctest: +ELLIPSIS
  2.998659073040...

For more information see the :ref:`properties of NearestImage <nearestimage>`.

The neighbor search can also be used with small molecule structures.
Here, we have MgI\ :sub:`2`, with each Mg atom surrounded by 6 iodine atoms,
at a distance of 2.92Å:

.. doctest::

  >>> small = gemmi.read_small_structure('../tests/2013551.cif')
  >>> mg_site = small.sites[0]
  >>> ns = gemmi.NeighborSearch(small, 4.0).populate()
  >>> for mark in ns.find_site_neighbors(mg_site, min_dist=0.1):
  ...   site = mark.to_site(small)
  ...   print(site.label, 'symmetry op #%d' % mark.image_idx)
  I symmetry op #0
  I symmetry op #0
  I symmetry op #0
  I symmetry op #3
  I symmetry op #3
  I symmetry op #3


Contact search
==============

Contacts in a molecule or in a crystal can be found using the neighbor search
described in the previous section. However, to make it easier, we have
a dedicated class ContactSearch. It uses the neighbor search to find pairs
of atoms that are close to each other and applies the filters described below.

When constructing a ContactSearch object we set the overall maximum search
distance. This distance is stored as the `search_radius` property:

.. doctest::

  >>> cs = gemmi.ContactSearch(4.0)
  >>> cs.search_radius
  4.0

Additionally, we can set up per-element radii.
This excludes pairs of atoms in a distance larger than the sum
of their per-element radii.
The radii are initialized as a linear function of the
:ref:`covalent radius <covalent_radius>`:
*r* = *a* × *r*\ :sub:`cov` + *b*/2.

.. doctest::

  >>> cs.setup_atomic_radii(1.0, 1.5)

Then, each radius can be accessed and modified individually:

.. doctest::

  >>> cs.get_radius(gemmi.Element('Zr'))
  2.5
  >>> cs.set_radius(gemmi.Element('Hg'), 1.5)


Next, we have the `ignore` property that can take
one of the following values:

* ContactSearch.Ignore.Nothing -- no filtering here,
* ContactSearch.Ignore.SameResidue -- ignore atom pairs from the same residue,
* ContactSearch.Ignore.AdjacentResidues -- ignore atom pairs from the same or
  adjacent residues,
* ContactSearch.Ignore.SameChain -- show only inter-chain contacts
  (including contacts between different symmetry images of one chain),
* ContactSearch.Ignore.SameAsu -- show only contacts between different
  asymmetric units.

.. doctest::

  >>> cs.ignore = gemmi.ContactSearch.Ignore.AdjacentResidues

You can also ignore atoms that have occupancy below the specified
threshold:

.. doctest::

  >>> cs.min_occupancy = 0.01

Sometimes, it is handy to get each atom pair twice (as A-B and B-A).
In such cases, set the `twice` property to true. By default, it is false:

.. doctest::

  >>> cs.twice
  False

Now consider an atom near a special position, such as a rotation axis.
An atom that is intended to be on a rotation axis can be slightly off,
as macromolecular refinement programs typically don't constrain coordinates.
In such a case, the atom appears to be near its symmetry images, but that's
not a contact, only an artifact. On the other hand, it's entirely possible for
an atom to be near a symmetry axis and bonded to its symmetry image.
To distinguish these two cases, we use a configurable cutoff distance:

.. doctest::

  >>> cs.special_pos_cutoff_sq = 0.5 ** 2  # setting cut-off to 0.5A

The contact search uses an instance of NeighborSearch.

.. doctest::

  >>> st_virus = gemmi.read_structure('../tests/5cvz_final.pdb')
  >>> st_virus.setup_entities()
  >>> ns = gemmi.NeighborSearch(st_virus[0], st_virus.cell, 5).populate()

If you'd like to ignore hydrogens from the model,
call `ns.populate(include_h=False)`.

If you'd like to ignore waters, either remove waters from the Model
(function `remove_waters()`) or ignore results that contain waters.

The actual contact search is done by:

.. doctest::

  >>> results = cs.find_contacts(ns)

  >>> len(results)
  49
  >>> results[0]  # doctest: +ELLIPSIS
  <gemmi.ContactSearch.Result object at 0x...>

The ContactSearch.Result class has four properties:

.. doctest::

  >>> results[0].partner1
  <gemmi.CRA A/SER 21/OG>
  >>> results[0].partner2
  <gemmi.CRA A/TYR 24/N>
  >>> results[0].image_idx
  52
  >>> results[0].dist
  2.8613363437597505

* The first two properties are :ref:`CRA <CRA>`\ s for the involved atoms.

* The `image_idx` is an index of the symmetry image (both crystallographic
  symmetry and strict NCS count) -- it is 0 iff both atoms (`partner1`
  and `partner2`) are in the same unit.
  In this example, the value can be high because it is a structure of
  an icosahedral viral capsid with 240 identical units in the unit cell.

* The last property is the distance between atoms.

Atoms pointed to by `partner1` and `partner2` can be far apart
in the asymmetric unit:

.. doctest::

  >>> results[0].partner1.atom.pos
  <gemmi.Position(42.221, 46.34, 19.436)>
  >>> results[0].partner2.atom.pos
  <gemmi.Position(49.409, 39.333, 19.524)>

But you can find the position of the symmetry image of `partner2` that
is in contact with `partner1` with:

.. doctest::

  >>> st_virus.cell.find_nearest_pbc_position(results[0].partner1.atom.pos,
  ...                                         results[0].partner2.atom.pos,
  ...                                         results[0].image_idx)
  <gemmi.Position(42.6647, 47.5137, 16.8644)>

You could also find the symmetry image of `partner1`
that is near the original position of `partner2`:

.. doctest::

  >>> st_virus.cell.find_nearest_pbc_position(results[0].partner2.atom.pos,
  ...                                         results[0].partner1.atom.pos,
  ...                                         results[0].image_idx, inverse=True)
  <gemmi.Position(49.2184, 39.9091, 16.7278)>

See also the command-line program :ref:`gemmi-contact <gemmi-contact>`.

Gemmi also provides an undocumented class LinkHunt which matches
contacts to link definitions from :ref:`monomer library <CCD_etc>`
and to connections (LINK, SSBOND) from the structure.
If you find it useful, please contact the author.

Superposition
=============

Gemmi includes the `QCP method <https://theobald.brandeis.edu/qcp/>`_
(Liu P, Agrafiotis DK, & Theobald DL, 2010)
for superposing two lists of points in 3D.
The C++ function `superpose_positions()` takes two arrays of positions
and an optional array of weights. Before applying this function to chains
it is necessary to determine pairs of corresponding atoms.
Here, as a minimal example, we superpose the backbone of the third residue:

.. doctest::

  >>> model = gemmi.read_structure('../tests/4oz7.pdb')[0]
  >>> res1 = model['A'][2]
  >>> res2 = model['B'][2]
  >>> atoms = ['N', 'CA', 'C', 'O']
  >>> gemmi.superpose_positions([res1.sole_atom(a).pos for a in atoms],
  ...                           [res2.sole_atom(a).pos for a in atoms])  #doctest: +ELLIPSIS
  <gemmi.SupResult object at 0x...>
  >>> _.rmsd
  0.006558389527590043

To make it easier, we also have a higher-level function
`calculate_superposition()` that operates on `ResidueSpan`\ s.
This function first performs the sequence alignment.
Then the matching residues are superposed, using either
all atoms in both residues, or only Cα atoms (for peptides)
and P atoms (for nucleotides).
Atoms without matching counterparts are ignored.
The returned object (SupResult) contains the RMSD and the transformation
(rotation matrix + translation vector) that superposes the second span
onto the first one.

Note that the RMSD can be defined in two ways: the sum of squared deviations
is divided either by 3N (as in PyMOL and QCP) or by N (as in SciPy).
Gemmi returns the former. To get the latter, multiply it by √3.

Here is a usage example:

.. doctest::

  >>> model = gemmi.read_structure('../tests/4oz7.pdb')[0]
  >>> polymer1 = model['A'].get_polymer()
  >>> polymer2 = model['B'].get_polymer()
  >>> ptype = polymer1.check_polymer_type()
  >>> sup = gemmi.calculate_superposition(polymer1, polymer2, ptype, gemmi.SupSelect.CaP)
  >>> sup.count  # number of atoms used
  10
  >>> sup.rmsd
  0.1462689168993659
  >>> sup.transform.mat
  <gemmi.Mat33 [-0.0271652, 0.995789, 0.0875545]
               [0.996396, 0.034014, -0.0777057]
               [-0.0803566, 0.085128, -0.993124]>
  >>> sup.transform.vec
  <gemmi.Vec3(-17.764, 16.9915, -1.77262)>

The arguments to `calculate_superposition()` are:

- two `ResidueSpan`\ s,
- polymer type (to avoid determining it when it's already known).
  The information of whether it's a protein or nucleic acid is used
  during sequence alignment (to detect gaps between residues in the polymer --
  it helps in rare cases when the sequence alignment alone is ambiguous),
  and it decides whether to use Cα or P atoms (see the next point),
- atom selection: one of `SupSelect.CaP` (only Cα or P atoms),
  `SupSelect.MainChain` or `SupSelect.All` (all atoms),
- (optionally) altloc -- the conformer choice.
  By default, atoms with non-blank altloc are ignored.
  With altloc='A', only the A conformer is considered
  (atoms with altloc either blank or A). Etc.
- (optionally) trim_cycles (default: 0) --  maximum number of outlier
  rejection cycles. This option was inspired by the align command in PyMOL.

  .. doctest::

    >>> sr = gemmi.calculate_superposition(polymer1, polymer2, ptype,
    ...                                    gemmi.SupSelect.All, trim_cycles=5)
    >>> sr.count
    73
    >>> sr.rmsd
    0.18315488879658484

- (optionally) trim_cutoff (default: 2.0) --  outlier rejection cutoff in RMSD,

To calculate current RMSD between atoms (without superposition),
call `calculate_current_rmsd()`. It takes the same arguments
except the ones for trimming:

  .. doctest::
  
    >>> gemmi.calculate_current_rmsd(polymer1, polymer2, ptype, gemmi.SupSelect.CaP).rmsd
    19.660883858565462

The calculated superposition can be applied to a span of residues,
changing the atomic positions in-place:

.. doctest::

  >>> polymer2[2].sole_atom('CB')  # before
  <gemmi.Atom CB at (-30.3, -10.6, -11.6)>
  >>> polymer2.transform_pos_and_adp(sup.transform)
  >>> polymer2[2].sole_atom('CB')  # after
  <gemmi.Atom CB at (-28.5, -12.6, 11.2)>
  >>> # it is now nearby the corresponding atom in chain A:
  >>> polymer1[2].sole_atom('CB')
  <gemmi.Atom CB at (-28.6, -12.7, 11.3)>

.. doctest::
  :hide:

  >>> gemmi.calculate_current_rmsd(polymer1, polymer2, ptype, gemmi.SupSelect.CaP).rmsd  # doctest: +ELLIPSIS
  0.1462689168994...

.. _selections:

Selections
==========

Gemmi selection syntax is based on the selection syntax from MMDB,
which is sometimes called CID (Coordinate ID). The MMDB syntax is described
at the bottom of
the `pdbcur documentation <http://legacy.ccp4.ac.uk/html/pdbcur.html>`_.

The selection has a form of slash-separated parts:
/models/chains/residues/atoms. Leading and trailing parts can be omitted
when it's not ambiguous. An empty field means that all items match,
with two exceptions. The empty chain part (e.g. `/1//`) matches only
a chain without an ID (blank chainID in the PDB format;
not spec-conformant, but possible in working files). The empty altloc
(examples will follow) matches atoms with a blank altLoc field.
Gemmi (but not MMDB) can take additional properties
added at the end after a semicolon (;).

Let us go through the individual filters first:

* `/1` -- selects model 1 (if the PDB file doesn't have MODEL records,
  it is assumed that the only model is model 1).
* `//D` (or just `D`) -- selects chain D.
* `//*/10-30` (or `10-30`) -- residues with sequence IDs from 10 to 30.
* `//*/10A-30A` (or `10A-30A` or `///10.A-30.A` or `10.A-30.A`) --
  sequence ID can include insertion code. The MMDB syntax has a dot between
  the sequence sequence number and the insertion code.
  In Gemmi, the dot is optional.
* `//*/(ALA)` (or `(ALA)`) -- selects residues with a given name.
* `//*//CB` (or `CB:*` or `CB[*]`) -- selects atoms with a given name.
* `//*//[P]` (or just `[P]`) -- selects phosphorus atoms.
* `//*//:B` (or `:B`) -- selects atoms with altloc B.
* `//*//:` (or `:`) -- selects atoms without altloc.
* `//*//;q<0.5` (or `;q<0.5`) -- selects atoms with occupancy below 0.5
  (inspired by PyMOL, where it'd be `q<0.5`).
* `//*//;b>40` (or `;b>40`) -- selects atoms with isotropic B-factor
  above a given value.
* `;polymer` or `;solvent` -- selects polymer or solvent residues
  (if the PDB file doesn't have TER records, call setup_entities() first).
* `*` -- selects all atoms.

The syntax supports also comma-separated lists and negations with `!`:

* `(!ALA)` -- all residues but alanine,
* `[C,N,O]` -- all C, N and O atoms,
* `[!C,N,O]` -- all atoms except C, N and O,
* `:,A` -- altloc either empty or A (which makes one conformation),
* `/1/A,B/20-40/CA[C]:,A` -- multiple selection criteria, all of them
  must be fulfilled.
* `(CYS);!polymer` -- select cysteine ligands

**Incompatibility** with MMDB.
In MMDB, if the chemical element is specified (e.g. `[C]` or `[*]`),
the alternative location indicator defaults to "" (no altloc),
rather than to "*" (any altloc). This might be surprising.
In Gemmi, if ':' is absent the altloc is not checked ("*").

Note: the selections in Gemmi are not widely used yet and the API may evolve.

A selection is a standalone object with a list of filters that
can be applied to any Structure, Model or its part.
Empty selection matches all atoms:

.. doctest::

  >>> sel = gemmi.Selection()  # empty - no filters

Selection initialized with a string:

.. doctest::

  >>> sel = gemmi.Selection('A/1-4/N9')

parses the string and creates filters corresponding to each part
of the selection,
which can then be used to iterate over the selected items in the hierarchy:

.. doctest::

  >>> st = gemmi.read_structure('../tests/1pfe.cif.gz')
  >>> for model in sel.models(st):
  ...     print('Model', model.name)
  ...     for chain in sel.chains(model):
  ...         print('-', chain.name)
  ...         for residue in sel.residues(chain):
  ...             print('   -', str(residue))
  ...             for atom in sel.atoms(residue):
  ...                 print('          -', atom.name)
  ...
  Model 1
  - A
     - 1(DG)
            - N9
     - 2(DC)
     - 3(DG)
            - N9
     - 4(DT)

`str()` creates a CID string from the selection:

.. doctest::

  >>> sel.str()
  '//A/1.-4./N9'

A helper function `first()` returns the first matching atom:

.. doctest::

  >>> # find the first Cl atom - returns model and CRA (chain, residue, atom)
  >>> gemmi.Selection('[CL]').first(st)
  (<gemmi.Model 1 with 2 chain(s)>, <gemmi.CRA A/CL 20/CL>)

Selection can be used to create a new structure (or model) with a copy
of the selected parts. In this example, we copy alpha-carbon atoms:

.. doctest::

  >>> st = gemmi.read_structure('../tests/1orc.pdb')
  >>> st[0].count_atom_sites()
  559
  >>> selection = gemmi.Selection('CA[C]')

  >>> # create a new structure
  >>> ca_st = selection.copy_structure_selection(st)
  >>> ca_st[0].count_atom_sites()
  64

  >>> # create a new model
  >>> ca_model = selection.copy_model_selection(st[0])
  >>> ca_model.count_atom_sites()
  64

Selection can also be used to remove atoms.
Here, we remove atoms with a B-factor above 50:

.. doctest::

  >>> sel = gemmi.Selection(';b>50')
  >>> sel.remove_selected(ca_st)
  >>> ca_st[0].count_atom_sites()
  61
  >>> sel.remove_selected(ca_model)
  >>> ca_model.count_atom_sites()
  61

We can also do the opposite and remove atoms that are not selected:

.. doctest::

  >>> sel.remove_not_selected(ca_model)
  >>> ca_model.count_atom_sites()
  0

:ref:`Previously <model_count_atom>`, we introduced a couple functions
that take selection as an argument. As an example, we can use one
of them to count heavy atoms in polymers:

.. doctest::

  >>> st[0].count_occupancies(gemmi.Selection('[!H,D];polymer'))
  496.0

Each residue and atom has a flag that can be set manually
and used to create a selection. In the following example
we select residues within a radius of 8Å from a selected point:

.. doctest::

  >>> selected_point = gemmi.Position(20, 40, 30)
  >>> ns = gemmi.NeighborSearch(st[0], st.cell, 8.0).populate()
  >>> # First, a flag is set for neighbouring residues.
  >>> for mark in ns.find_atoms(selected_point):
  ...     mark.to_cra(st[0]).residue.flag = 's'
  >>> # Then, we select residues with this flag.
  >>> selection = gemmi.Selection().set_residue_flags('s')
  >>> # Next, we can use this selection.
  >>> selection.copy_model_selection(st[0]).count_atom_sites()
  121

Note: NeighborSearch searches for atoms in all symmetry images.
This is why it takes UnitCell as a parameter.
To search only in atoms directly listed in the file, pass an empty cell
(`gemmi.UnitCell()`).

Instead of selecting whole residues, we can select atoms.
Here, we select atoms in the radius of 8Å from a selected point:

.. doctest::

  >>> # selected_point and ns are reused from the previous example
  >>> # First, a flag is set for neighbouring atoms.
  >>> for mark in ns.find_atoms(selected_point):
  ...     mark.to_cra(st[0]).atom.flag = 's'
  >>> # Then, we select atoms with this flag.
  >>> selection = gemmi.Selection().set_atom_flags('s')
  >>> # Next, we can use this selection.
  >>> selection.copy_model_selection(st[0]).count_atom_sites()
  59

.. _graph_analysis:

Graph analysis
==============

This section shows how to analyze chemical molecules (``gemmi.ChemComp``)
using external libraries dedicated to graph analysis.
We don't plan to implement graph algorithms within gemmi,
except for the simplest ones, like breadth-first search,
which is used in a couple of functions
( ``Restraints::find_shortest_path()``, ``BondIndex::graph_distance()``).

Here is how to set up a graph in the Boost Graph Library
(`BGL <http://boost.org/libs/graph>`_) in C++:

.. literalinclude:: ../examples/with_bgl.cpp
   :lines: 9-10,13-41

Here, we set up a `NetworkX <https://networkx.org/>`_ graph in Python:

.. doctest::
  :skipif: networkx is None

  >>> import networkx

  >>> G = networkx.Graph()
  >>> block = gemmi.cif.read('../tests/SO3.cif')[-1]
  >>> so3 = gemmi.make_chemcomp_from_block(block)
  >>> for atom in so3.atoms:
  ...     G.add_node(atom.id, Z=atom.el.atomic_number)
  ...
  >>> for bond in so3.rt.bonds:
  ...     G.add_edge(bond.id1.atom, bond.id2.atom)  # ignoring bond type
  ...

Now, as a quick example, we can count automorphisms:

.. doctest::
  :skipif: networkx is None

  >>> import networkx.algorithms.isomorphism as iso
  >>> GM = iso.GraphMatcher(G, G, node_match=iso.categorical_node_match('Z', 0))
  >>> # expecting 3! automorphisms (permutations of the three oxygens)
  >>> sum(1 for _ in GM.isomorphisms_iter())
  6

The median number of automorphisms of molecules in the CCD is only 4.
However, the highest number of isomorphisms as of 2023 (ignoring hydrogens,
bond orders, chiralities) is a staggering 6879707136
for `T8W <https://www.rcsb.org/ligand/T8W>`_.
This value can be calculated almost instantly with nauty, which
returns a set of *generators* of the automorphism group
and the sets of equivalent vertices called *orbits*,
rather than listing all automorphisms.
Nauty is a software for "determining the automorphism group
of a vertex-coloured graph, and for testing graphs for isomorphism".
We can use it in Python through the pynauty module.

Here we set up a `pynauty <https://github.com/pdobsan/pynauty>`_
graph from the `so3` object prepared in the previous example:

.. doctest::
  :skipif: pynauty is None

  >>> import pynauty

  >>> n_vertices = len(so3.atoms)
  >>> adjacency = {n: [] for n in range(n_vertices)}
  >>> indices = {atom.id: n for n, atom in enumerate(so3.atoms)}
  >>> elements = {atom.el.atomic_number for atom in so3.atoms}
  >>> # The order in dict in Python 3.6+ is the insertion order.
  >>> coloring = {elem: set() for elem in sorted(elements)}
  >>> for n, atom in enumerate(so3.atoms):
  ...   coloring[atom.el.atomic_number].add(n)
  ...
  >>> for bond in so3.rt.bonds:
  ...   n1 = indices[bond.id1.atom]
  ...   n2 = indices[bond.id2.atom]
  ...   adjacency[n1].append(n2)
  ...
  >>> G = pynauty.Graph(n_vertices, adjacency_dict=adjacency, vertex_coloring=coloring.values())

The colors of vertices in this graph correspond to elements.
Pynauty takes a list of sets of vertices with the same color,
without knowing which color corresponds to which element.
We sorted the elements to ensure that two graphs with the same atoms
have the same coloring. However, SO3 and PO3 graphs would also have
the same coloring and would be reported as isomorphic.
So it is necessary to check if the elements in molecules are the same.

You may also want to encode other properties in the graph, such as bond orders,
atom charges and chiralities, as described in
`this paper <https://jcheminf.biomedcentral.com/articles/10.1186/s13321-023-00692-1>`_.
Here, we only present a simplified, minimal example.
Now, to finish this example, let's get the order of the isomorphism group
(which should be the same as in the NetworkX example, i.e. 6):

.. doctest::
  :skipif: pynauty is None

  >>> (gen, grpsize1, grpsize2, orbits, numorb) = pynauty.autgrp(G)
  >>> grpsize1 * 10**grpsize2
  6.0

.. _graph_isomorphism:

Graph isomorphism
-----------------

In this example we use Python NetworkX to compare molecules from the
Refmac monomer library with the PDB's Chemical Component Dictionary (CCD).
The same could be done with other graph analysis libraries,
such as Boost Graph Library, igraph, etc.

The program below takes compares specified monomer cif files with
corresponding CCD entries. Hydrogens and bond types are ignored.
It takes less than half a minute to go through the 25,000 monomer
files distributed with CCP4 (as of Oct 2018),
so we do not try to optimize the program.

.. literalinclude:: ../examples/ccd_gi.py
   :language: python
   :lines: 3-

If we run it on monomers that start with M we get:

.. code-block:: console

  $ examples/ccd_gi.py $CLIBD_MON/m/*.cif
  M10 is isomorphic
         O9 -> O4
         O4 -> O9
  MK8 is isomorphic
         O2 -> OXT
  MMR differs
        missing: O12 O4
  2 of 821 monomers not found in CCD

So in M10 the two atoms marked green are swapped:

.. image:: img/M10-isomorphism.png
    :align: center
    :scale: 100

(The image was generated in NGL and compressed with Compress-Or-Die.)

.. _substructure_matching:

Substructure matching
---------------------

Now, a little script to illustrate subgraph isomorphism.
The script takes a (three-letter-)code of a molecule that is to be used
as a pattern and finds CCD entries that contain such a substructure.
As in the previous example, hydrogens and bond types are ignored.

.. literalinclude:: ../examples/ccd_subgraph.py
   :language: python
   :lines: 3-

Let us check which entries have HEM as a substructure:

.. code-block:: console

  $ examples/ccd_subgraph.py HEM
  1FH 	 +6 nodes, +7 edges
  2FH 	 +6 nodes, +7 edges
  4HE 	 +7 nodes, +8 edges
  522 	 +2 nodes, +2 edges
  6CO 	 +6 nodes, +7 edges
  6CQ 	 +7 nodes, +8 edges
  89R 	 +3 nodes, +3 edges
  CLN 	 +1 nodes, +2 edges
  DDH 	 +2 nodes, +2 edges
  FEC 	 +6 nodes, +6 edges
  HAS 	 +22 nodes, +22 edges
  HCO 	 +1 nodes, +1 edges
  HDM 	 +2 nodes, +2 edges
  HEA 	 +17 nodes, +17 edges
  HEB 	 +0 nodes, +0 edges
  HEC 	 +0 nodes, +0 edges
  HEM 	 +0 nodes, +0 edges
  HEO 	 +16 nodes, +16 edges
  HEV 	 +2 nodes, +2 edges
  HP5 	 +2 nodes, +2 edges
  ISW 	 +0 nodes, +0 edges
  MH0 	 +0 nodes, +0 edges
  MHM 	 +0 nodes, +0 edges
  N7H 	 +3 nodes, +3 edges
  NTE 	 +3 nodes, +3 edges
  OBV 	 +14 nodes, +14 edges
  SRM 	 +20 nodes, +20 edges
  UFE 	 +18 nodes, +18 edges

.. _maximum_common_subgraph:

Maximum common subgraph
-----------------------

In this example, we use McGregor's algorithm, implemented in the Boost Graph
Library, to identify the largest common induced subgraphs. To ensure connectivity,
we set the option `only_connected_subgraphs=true`.

To illustrate this example, we compare the ligands AUD and LSA:

.. image:: img/aud_lsa.png
    :align: center
    :scale: 100

The complete code is in :file:`examples/with_bgl.cpp`. This file also contains
examples of using the BGL's VF2 implementation for checking graph
and subgraph isomorphisms.

.. literalinclude:: ../examples/with_bgl.cpp
   :start-after: Example 4
   :end-before: minimal program


Torsion angles
==============

This section presents functions dedicated to the calculation of the dihedral
angles φ (phi), ψ (psi), and ω (omega) of the protein backbone.
These functions are built upon the more general `calculate_dihedral` function,
introduced in :ref:`the section about coordinates <coordinates>`,
which takes four points in space as arguments.

`calculate_omega()` calculates the ω angle, which is usually around 180°:

.. doctest::

  >>> from math import degrees
  >>> chain = st_virus[0]['A']
  >>> degrees(gemmi.calculate_omega(chain[0], chain[1]))
  159.9092215006572
  >>> for res in chain[:5]:
  ...     next_res = chain.next_residue(res)
  ...     if next_res:
  ...         omega = gemmi.calculate_omega(res, next_res)
  ...         print(res.name, degrees(omega))
  ...
  ALA 159.9092215006572
  ALA -165.26874513591127
  ALA -165.85686681169696
  THR -172.99968385093572
  SER 176.74223937657652

The φ and ψ angles are often used together, so they are calculated
in the same function `calculate_phi_psi()`:

.. doctest::

  >>> for res in chain[:5]:
  ...     prev_res = chain.previous_residue(res)
  ...     next_res = chain.next_residue(res)
  ...     phi, psi = gemmi.calculate_phi_psi(prev_res, res, next_res)
  ...     print('%s %8.2f %8.2f' % (res.name, degrees(phi), degrees(psi)))
  ...
  ALA      nan   106.94
  ALA  -116.64    84.57
  ALA   -45.57   127.40
  THR   -62.01   147.45
  SER   -92.85   161.53

In C++ these functions can be found in `gemmi/calculate.hpp`.

The torsion angles φ and ψ can be visualized on the Ramachandran plot.
Let us plot angles from all PDB entries with a resolution higher than 1.5Å.
Usually, glycine, proline and the residue preceding proline (pre-proline)
are plotted separately. Here, we will exclude pre-proline and make
a separate plot for each amino acid. So first, we calculate angles
and save φ,ψ pairs in a set of files -- one file per residue.

.. literalinclude:: ../examples/rama_gather.py
   :language: python
   :start-at: import sys

The script above works with coordinate files in any of the formats
supported by gemmi (PDB, mmCIF, mmJSON). As of 2019, processing
a :ref:`local copy of the PDB archive <pdb_dir>`
in the PDB format takes about 20 minutes.

In the second step, we plot the data points with Matplotlib.
We use a script that can be found in :file:`examples/rama_plot.py`.
Six of the resulting plots are shown here (click to enlarge):

.. image:: img/ramachandran-per-aa.png
    :align: center
    :scale: 60

.. _topology:

Topology
========

In gemmi, topology contains restraints *applied* to a model,
as well as information about the provenance of these restraints,
and related utilities. (The restraints include bond-distance
restraints and therefore also bonding information).

Applied restraints differ from *template* restraints in a monomer library,
although both are referred to as restraints.
The library (template) restraints specify, for instance,
how to restrain the angle CD2-CE2-NE1 in any TRP residue.
In contrast, topology (concrete, applied to a model) restraints specify
how to restrain the angle between specific atoms in the model (say, atoms #721-#723-#722).
The latter seems like a trivial application of the former, and it often is.
But in general, if the aim is to support all atomic models present in the PDB,
including those with unusual arrangements of alternative conformations,
the process of determining the topology is quite convoluted.

The typical macromolecular refinement workflow begins by reading
a coordinate file and a monomer library. The template monomer restraints
from the library (for bond distances, angles, etc.) are applied to monomers,
and link definitions are matched to explicit and implicit links between
monomers. Link definitions contain additional restraints and modifications
to restraints within the linked monomers.

Currently, our topology works only with the monomer library from CCP4.
This library was introduced in the
`early 2000s <https://doi.org/10.1107/S0907444904023510>`_.
Monomer libraries distributed with other popular macromolecular refinement
programs, PHENIX and BUSTER, are organized somewhat differently
(geostd from PHENIX is actually quite similar).

The restraints that we use are also similar to what is used
in molecular dynamics (bond, angle, dihedral and improper dihedral restraints).
Although the MD potentials have been deemed inadequate for refinement,
and the restraints in experimental structural biology have been
improved independently of the restraints in MD,
they haven't diverged too much and with a little work one could be
substituted for the other.

When preparing a topology, macromolecular programs (in particular, Refmac)
may also add hydrogens or shift existing hydrogens to the riding positions.
And reorder atoms. In Python, we have one function that does it all:

.. code-block:: python

    gemmi.prepare_topology(st: gemmi.Structure,
                           monlib: gemmi.MonLib,
                           model_index: int = 0,
                           h_change: gemmi.HydrogenChange = HydrogenChange.NoChange,
                           reorder: bool = False,
                           warnings: object = None,
                           ignore_unknown_links: bool = False) -> gemmi.Topo

where

* `monlib` is an instance of an undocumented MonLib class.
  For now, here is an example of how to read the CCP4 monomer library
  (Refmac dictionary):

  .. code-block:: python

    monlib_path = os.environ['CCP4'] + '/lib/data/monomers'
    resnames = st[0].get_all_residue_names()
    monlib = gemmi.read_monomer_lib(monlib_path, resnames)

* `h_change` can be one of the following:

  * HydrogenChange.NoChange -- no change,
  * HydrogenChange.Shift -- shift existing hydrogens to their ideal (riding) positions,
  * HydrogenChange.Remove -- remove all H and D atoms,
  * HydrogenChange.ReAdd -- discard and re-create hydrogens in ideal positions
    (if the hydrogen position is not uniquely determined, its occupancy is set to zero),
  * HydrogenChange.ReAddButWater -- the same as above, but doesn't add H in waters,
  * HydrogenChange.ReAddKnown -- the same as above, but doesn't add any H atoms whose
    positions are not uniquely determined,

* `reorder` -- changes the order of atoms within each residue
  to match the order in the corresponding monomer cif file,

* `warnings` --  by default, an exception is raised when a chemical component
  is missing in the monomer library, or when a link is missing,
  or when the hydrogen adding procedure encounters an unexpected configuration.
  You can set warnings=sys.stderr to only print a warning to stderr
  and continue. sys.stderr can be replaced with any object that has
  the methods `write(str)` and `flush()`.


TBC

.. _pdb_dir:

Local copy of the PDB archive
=============================

Some of the examples in this documentation work with a local copy
of the Protein Data Bank archive. This subsection describes
the assumed setup.

Like in BioJava, we assume that the `$PDB_DIR` environment variable
points to a directory that contains `structures/divided/mmCIF` -- the same
arrangement as on the
`PDB's FTP <ftp://ftp.wwpdb.org/pub/pdb/data/structures/>`_ server.

.. code-block:: console

    $ cd $PDB_DIR
    $ du -sh structures/*/*  # as of Jun 2017
    34G    structures/divided/mmCIF
    25G    structures/divided/pdb
    101G   structures/divided/structure_factors
    2.6G   structures/obsolete/mmCIF

A traditional way to keep an up-to-date local archive is to rsync it
once a week:

.. code-block:: shell

    #!/bin/sh -x
    set -u  # PDB_DIR must be defined
    rsync_subdir() {
      mkdir -p "$PDB_DIR/$1"
      # Using PDBe (UK) here, can be replaced with RCSB (USA) or PDBj (Japan),
      # see https://www.wwpdb.org/download/downloads
      rsync -rlpt -v -z --delete \
	  rsync.ebi.ac.uk::pub/databases/pdb/data/$1/ "$PDB_DIR/$1/"
    }
    rsync_subdir structures/divided/mmCIF
    #rsync_subdir structures/obsolete/mmCIF
    #rsync_subdir structures/divided/pdb
    #rsync_subdir structures/divided/structure_factors

Gemmi has a helper function for using the local archive copy.
It takes a PDB code (case insensitive) and a symbol denoting what file
is requested: P for PDB, M for mmCIF, S for SF-mmCIF.

.. doctest::

  >>> os.environ['PDB_DIR'] = '/copy'
  >>> gemmi.expand_if_pdb_code('1ABC', 'P') # PDB file
  '/copy/structures/divided/pdb/ab/pdb1abc.ent.gz'
  >>> gemmi.expand_if_pdb_code('1abc', 'M') # mmCIF file
  '/copy/structures/divided/mmCIF/ab/1abc.cif.gz'
  >>> gemmi.expand_if_pdb_code('1abc', 'S') # SF-mmCIF file
  '/copy/structures/divided/structure_factors/ab/r1abcsf.ent.gz'

If the first argument is not in the PDB code format (4 characters for now)
the function returns the first argument.

.. doctest::

  >>> arg = 'file.cif'
  >>> gemmi.is_pdb_code(arg)
  False
  >>> gemmi.expand_if_pdb_code(arg, 'M')
  'file.cif'

Multiprocessing
===============

(Python-specific)

Most of the gemmi objects cannot be pickled. Therefore, they cannot be
passed between processes when using the multiprocessing module.
Currently, the only picklable classes (with protocol >= 2) are:
UnitCell and SpaceGroup.

Usually, it is possible to organize multiprocessing in such a way that
gemmi objects are not passed between processes. The example script below
traverses subdirectories and asynchronously analyzes coordinate files,
using 4 worker processes in parallel.

.. literalinclude:: ../examples/multiproc.py
   :language: python
   :lines: 4-