contact_hist_refcmp_selcmp_pair

mdtools.structure.contact_hist_refcmp_selcmp_pair(cm, natms_per_refcmp=1, natms_per_selcmp=1, minlength=0, dtype=int)[source]

Bin the number of “bonds” (Atom- Atom contacts) between pairs of reference and selection compounds.

A compound is usually a chemically meaningful subgroup of an AtomGroup. This can e.g. be a Segment, Residue, fragment or a single Atom. Refer to the MDAnalysis’ user guide for an explanation of these terms. Note that in any case, only Atoms belonging to the original AtomGroup are taken into account, even if the compound might comprise additional Atoms that are not contained in the original AtomGroup.

Parameters:

cm (array_like) – (Boolean) contact matrix of shape (m, n) as e.g. generated by mdtools.structure.contact_matrix(), where m is the number of reference Atoms and n is the number of selection Atoms. Alternatively, cm can already be a compound contact count matrix as e.g. generated by mdtools.structure.cmp_contact_count_matrix(). In this case, you probably want to set natms_per_refcmp and natms_per_selcmp to 1, to keep cm unaltered.
natms_per_refcmp (int or array_like, optional) – Number of Atoms per reference compound. Can be a single integer or an array of integers. If natms_per_refcmp is a single integer, all reference compounds are assumed to contain the same number of Atoms. In this case, natms_per_refcmp must be an integer divisor of cm.shape[0]. If natms_per_refcmp is an array of integers, it must contain the number of reference Atoms for each single reference compound. In this case, sum(natms_per_refcmp) must be equal to cm.shape[0].
natms_per_selcmp (int or array_like, optional) – Same for selection compounds (natms_per_selcmp is checked against cm.shape[1]).
minlength (int, optional) – A minimum number of bins for the output array. The output array will have at least this number of elements, though it will be longer if necessary. See numpy.bincount() for further information.
dtype (dtype, optional) – Data type of the output array.

Returns:

hist_refcmp_selcmp_pair (numpy.ndarray) – Histogram of the number of “bonds” (Atom- Atom contacts) between pairs of reference and selection compounds (refcmp-selcmp pairs). A refcmp-selcmp pair is defined as a reference and selection compound that are connected with each other via at least one “bond”.

See also

mdtools.structure.contact_matrix(): Construct a boolean contact matrix for two MDAnalysis AtomGroups
mdtools.structure.contact_hists(): Bin the number of contacts between reference and selection compounds into histograms.
mdtools.structure.contact_hist_refcmp_diff_selcmp(): Bin the number of contacts that reference compounds establish to different selection compounds into a histogram.
mdtools.structure.contact_hist_refcmp_same_selcmp(): Bin the number of contacts that reference compounds establish to the same selection compound into a histogram.
mdtools.structure.contact_hist_refcmp_selcmp_tot(): Bin the total number of contacts that reference compounds establish to selection compounds into a histogram.
numpy.bincount(): Count the number of occurrences of each value in an array of non-negative ints

Notes

Atoms belonging to the same compound must form a contiguous set in the input contact matrix, otherwise the result will be wrong.

About the output array:

hist_refcmp_selcmp_pair

The first element is meaningless (a refcmp-selcmp pair with zero “bonds” is not a pair) and therefore set to zero. The second element is the number of refcmp-selcmp pairs connected via exactly one “bond”, the third element is the number of refcmp-selcmp pairs connected via exactly two “bonds”, and so on.

The sum of all histogram elements might exceed the number of reference compounds, because a single reference compound can be connected to different selection compounds via different numbers of “bonds”. Even each histogram element on its own might exceed the number of reference compounds, because a single reference compound can be connected to different selection compounds via the same number of “bonds”.

Hence, this histogram should be normalized by the number of refcmp-selcmp pairs and not by the number of reference compounds. Then it is e.g. possible to say that 100 % of the refcmp-selcmp connections are monodentate while at the same time 50 % of the refcmp-selcmp connections are bidentate.

This behavior is complementary to the histogram returned by mdtools.structure.contact_hist_refcmp_same_selcmp()

If both natms_per_refcmp and natms_per_selcmp are 1 and cm is a true boolean contact matrix, hist_refcmp_selcmp_pair is equal to [0, y], where y is the number of refatm-selatm pairs.

Examples

>>> cm = np.tril(np.ones((5,4), dtype=bool), -1)
>>> cm[3, 0] = 0
>>> cm
array([[False, False, False, False],
       [ True, False, False, False],
       [ True,  True, False, False],
       [False,  True,  True, False],
       [ True,  True,  True,  True]])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(cm)
array([0, 9])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(cm=cm, minlength=4)
array([0, 9, 0, 0])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(cm=cm, minlength=1)
array([0, 9])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(cm=cm, dtype=np.uint32)
array([0, 9], dtype=uint32)
>>> hist = mdt.strc.contact_hist_refcmp_selcmp_pair(cm)
>>> hist[1] == np.count_nonzero(cm)
True

>>> mdt.strc.cmp_contact_count_matrix(
...     cm=cm, natms_per_refcmp=[2, 2, 1]
... )
array([[1, 0, 0, 0],
       [1, 2, 1, 0],
       [1, 1, 1, 1]])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(
...     cm=cm, natms_per_refcmp=[2, 2, 1]
... )
array([0, 7, 1])

>>> mdt.strc.cmp_contact_count_matrix(cm=cm, natms_per_selcmp=2)
array([[0, 0],
       [1, 0],
       [2, 0],
       [1, 1],
       [2, 2]])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(
...     cm=cm, natms_per_selcmp=2
... )
array([0, 3, 3])

>>> mdt.strc.cmp_contact_count_matrix(
...     cm=cm, natms_per_refcmp=[2, 2, 1], natms_per_selcmp=2
... )
array([[1, 0],
       [3, 1],
       [2, 2]])
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(
...     cm=cm, natms_per_refcmp=[2, 2, 1], natms_per_selcmp=2
... )
array([0, 2, 2, 1])

Edge cases:

>>> cm = np.array([], dtype=bool).reshape(0, 4)
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(
...     cm, natms_per_refcmp=[]
... )
array([], dtype=int64)
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(cm, natms_per_refcmp=1)
array([], dtype=int64)
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(
...     cm, natms_per_refcmp=[], minlength=2, dtype=np.uint32
... )
array([0, 0], dtype=uint32)

>>> cm = np.array([], dtype=bool).reshape(6, 0)
>>> mdt.strc.contact_hist_refcmp_selcmp_pair(
...     cm, natms_per_refcmp=3, natms_per_selcmp=[]
... )
array([], dtype=int64)