i04dZddlZddlZddlmZddlmZddlmZddlmZm Z  ddl Z d Z n #e $rd Z YnwxYweje d d d d [ GddeZ ddZdS)a Leaflet identification --- :mod:`MDAnalysis.analysis.leaflet` ============================================================== This module implements the *LeafletFinder* algorithm, described in [Michaud-Agrawal2011]_. It can identify the lipids in a bilayer of arbitrary shape and topology, including planar and undulating bilayers under periodic boundary conditions or vesicles. One can use this information to identify * the upper and lower leaflet of a *planar membrane* by comparing the the :meth:`~MDAnalysis.core.groups.AtomGroup.center_of_geometry` of the leaflet groups, or * the outer and inner leaflet of a *vesicle* by comparing histograms of distances from the centre of geometry (or possibly simply the :meth:`~MDAnalysis.core.groups.AtomGroup.radius_of_gyration`). See example scripts in the MDAnalysisCookbook_ on how to use :class:`LeafletFinder`. The function :func:`optimize_cutoff` implements a (slow) heuristic method to find the best cut off for the LeafletFinder algorithm. .. _MDAnalysisCookbook: https://github.com/MDAnalysis/MDAnalysisCookbook/tree/master/examples Algorithm --------- 1. build a graph of all phosphate distances < cutoff 2. identify the largest connected subgraphs 3. analyse first and second largest graph, which correspond to the leaflets For further details see [Michaud-Agrawal2011]_. Classes and Functions --------------------- .. autoclass:: LeafletFinder :members: .. autofunction:: optimize_cutoff N)core) distances) selections)dueDoiTFz10.1002/jcc.21787zLeafletFinder algorithmzMDAnalysis.analysis.leaflet) descriptionpath cite_modulecZeZdZdZddZdZdZdZdd Zd Z dd Z d Z d Z dZ dZdS) LeafletFinderaIdentify atoms in the same leaflet of a lipid bilayer. This class implements the *LeafletFinder* algorithm [Michaud-Agrawal2011]_. Parameters ---------- universe : Universe :class:`~MDAnalysis.core.universe.Universe` object. select : AtomGroup or str A AtomGroup instance or a :meth:`Universe.select_atoms` selection string for atoms that define the lipid head groups, e.g. universe.atoms.PO4 or "name PO4" or "name P*" cutoff : float (optional) head group-defining atoms within a distance of `cutoff` Angstroms are deemed to be in the same leaflet [15.0] pbc : bool (optional) take periodic boundary conditions into account [``False``] sparse : bool (optional) ``None``: use fastest possible routine; ``True``: use slow sparse matrix implementation (for large systems); ``False``: use fast :func:`~MDAnalysis.lib.distances.distance_array` implementation [``None``]. Raises ------ ImportError This class requires the optional package `networkx`. If not found an ImportError is raised. Example ------- The components of the graph are stored in the list :attr:`LeafletFinder.components`; the atoms in each component are numbered consecutively, starting at 0. To obtain the atoms in the input structure use :meth:`LeafletFinder.groups`:: u = mda.Universe(PDB) L = LeafletFinder(u, 'name P*') leaflet0 = L.groups(0) leaflet1 = L.groups(1) The residues can be accessed through the standard MDAnalysis mechanism:: leaflet0.residues provides a :class:`~MDAnalysis.core.groups.ResidueGroup` instance. Similarly, all atoms in the first leaflet are then :: leaflet0.residues.atoms .. versionchanged:: 1.0.0 Changed `selection` keyword to `select` .. versionchanged:: 2.0.0 The universe keyword no longer accepts non-Universe arguments. Please create a :class:`~MDAnalysis.core.universe.Universe` first. .@FNc8tsd}t|||_||_t |jt jjr |j|_n| |j|_||_ ||_ | |dS)NzThe LeafletFinder class requires an installation of networkx. Please install networkx https://networkx.org/documentation/stable/install.html) HAS_NX ImportErroruniverseselectionstring isinstancergroups AtomGroup selection select_atomspbcsparse _init_graph)selfrselectcutoffrrerrmsgs e/srv/www/vhosts/g4struct/public_html/venv/lib/python3.11/site-packages/MDAnalysis/analysis/leaflet.py__init__zLeafletFinder.__init__s &I  f%% %  % d*DK,A B B I!1DNN%2243GHHDN       cx||_||_||_dSN)r _get_graphgraph_get_components componentsrrs r!rzLeafletFinder._init_graphs1 __&& ..00r#cF|jr|jjjj}nd}|jj}|jdurJ tj ||j d|}n#t$rtj dtdwxYw|jdurtj ||j d |}nh tj ||j d|}nI#t$r<tj d tdtj ||j d |}YnwxYwtj|S) zBuild graph from adjacency matrix at the given cutoff. Automatically select between high and low memory usage versions of contact_matrix.NFnumpy)r returntypeboxz4N x N matrix too big, use sparse=True or sparse=Noner)category stacklevelTrzcN x N matrix too big - switching to sparse matrix method (works fine, but is currently rather slow))rr trajectoryts dimensionsr positionsrrcontact_matrixr ValueErrorwarningswarn UserWarningNXGraph)rr.coordadjs r!r&zLeafletFinder._get_graphs~ 8 -*-8CCC( ;%   .$+'s    J(    [D *dkhCCC  .$+'s    (   .$+( x}}sA(A=(CAD  D cHdtj|jDS)zEReturn connected components (as sorted numpy arrays), sorted by size.cPg|]#}tjt|$S)npsortlist).0 components r! z1LeafletFinder._get_components..s8    GDOO $ $   r#)r:connected_componentsr'rs r!r(zLeafletFinder._get_componentss0  4TZ@@    r#cB||j}||dS)z5Update components, possibly with a different *cutoff*N)rrr*s r!updatezLeafletFinder.updates( >[F      r#cXtdt|jDS)z/Dict of component index with size of component.c3>K|]\}}|t|fVdSr%)len)rDidxrEs r! z&LeafletFinder.sizes..sE  "Cc)nn%      r#)dict enumerater)rHs r!sizeszLeafletFinder.sizess:  &/&@&@      r#cr|!t|S||S)aReturn a :class:`MDAnalysis.core.groups.AtomGroup` for *component_index*. If no argument is supplied, then a list of all leaflet groups is returned. See Also -------- :meth:`LeafletFinder.group` :meth:`LeafletFinder.groups_iter` )rC groups_itergrouprcomponent_indexs r!rzLeafletFinder.groupss6  "((**++ +::o.. .r#cJd|j|D}|j|S)zIReturn a :class:`MDAnalysis.core.groups.AtomGroup` for *component_index*.cg|]}|Sr@r@)rDis r!rFz'LeafletFinder.group..s???1???r#)r)r)rrWindicess r!rUzLeafletFinder.groups,@?doo>???~g&&r#c#Ktt|jD]}||VdS)z(Iterator over all leaflet :meth:`groups`N)rangerMr)rUrVs r!rTzLeafletFinder.groups_itersL$S%9%9:: . .O**_-- - - - - . .r#c tj||dd}||f|dddjd it |d|5}t |D]4\}}d|dz}||| 5 ddddS#1swxYwYdS) a Write selections for the leaflets to *filename*. The format is typically determined by the extension of *filename* (e.g. "vmd", "pml", or "ndx" for VMD, PyMol, or Gromacs). See :class:`MDAnalysis.selections.base.SelectionWriter` for all options. formatNmodewz?leaflets based on select={selectionstring!r} cutoff={cutoff:f} )r`preamblez leaflet_{0:d}r)namer@)r get_writerpopr_varsrQrTwrite)rfilenamekwargsswwriterrZagrcs r!write_selectionzLeafletFinder.write_selectionsD "8VZZ$-G-G H H R  FC((^W^t**     ,"4#3#3#5#566 , ,2&--q1u66 Rd ++++ , , , , , , , , , , , , , , , , , , ,s"ACC C c~d|j|j|jjt |jS)NzI)r_rrrn_atomsrMr)rHs r!__repr__zLeafletFinder.__repr__2s:Zaa  K N "     r#)rFNr%)__name__ __module__ __qualname____doc__r"rr&r(rJrRrrUrTrmrpr@r#r!rrcs99v!!!!(111...`   !!!!     / / / /''' ... ,,,,     r#r$@4@?皙?c `|ddg}tj|||D]}t||fd|i|} | } t | dkr:t | d} t | d} tj| | z | | zz } | |kr||t | ftj |d}~| ddg |dS) aDFind cutoff that minimizes number of disconnected groups. Applies heuristics to find best groups: 1. at least two groups (assumes that there are at least 2 leaflets) 2. reject any solutions for which: .. math:: \frac{|N_0 - N_1|}{|N_0 + N_1|} > \mathrm{max_imbalance} with :math:`N_i` being the number of lipids in group :math:`i`. This heuristic picks groups with balanced numbers of lipids. Parameters ---------- universe : Universe :class:`MDAnalysis.Universe` instance select : AtomGroup or str AtomGroup or selection string as used for :class:`LeafletFinder` dmin : float (optional) dmax : float (optional) step : float (optional) scan cutoffs from `dmin` to `dmax` at stepsize `step` (in Angstroms) max_imbalance : float (optional) tuning parameter for the balancing heuristic [0.2] kwargs : other keyword arguments other arguments for :class:`LeafletFinder` Returns ------- (cutoff, N) optimum cutoff and number of groups found .. Note:: This function can die in various ways if really no appropriate number of groups can be found; it ought to be made more robust. .. versionchanged:: 1.0.0 Changed `selection` keyword to `select` rNrrrzcutoff,N)namesN)order) rerAarangerrRrMfloatabsappendrec fromrecordsrB)rrdmindmaxstep max_imbalanceri_sizesrLFrRn0n1 imbalanceresultss r!optimize_cutoffr;s%h JJx F)D$--11 8V E EF Ef E E  u::>>  58__ 58__F27OOrBw/ } $ $  vs288::/0000f  z ::G LLXL''' 1:r#)rurvrwrx)rtr7r,rArrrrr networkxr:rrciteobjectrrr@r#r!rsS2--\FF FFF  C) &  U U U U U FU U U v   HHHHHHs -77