segmentmodel.py 58.1 KB
Newer Older
1
import numpy as np
2
import random
3
from arbdmodel import PointParticle, ParticleType, Group, ArbdModel
4
from coords import rotationAboutAxis, quaternion_from_matrix, quaternion_to_matrix
5
6
7
8
from nonbonded import *
from copy import copy, deepcopy
from nbPot import nbDnaScheme

cmaffeo2's avatar
cmaffeo2 committed
9
10
from scipy.special import erf
import scipy.optimize as opt
11
from scipy import interpolate
cmaffeo2's avatar
cmaffeo2 committed
12

cmaffeo2's avatar
cmaffeo2 committed
13
14
15
from CanonicalNucleotideAtoms import canonicalNtFwd, canonicalNtRev, seqComplement
from CanonicalNucleotideAtoms import enmTemplateHC, enmTemplateSQ, enmCorrectionsHC

16
import pdb
17
"""
cmaffeo2's avatar
cmaffeo2 committed
18
TODO:
cmaffeo2's avatar
cmaffeo2 committed
19
 + fix handling of crossovers for atomic representation
cmaffeo2's avatar
cmaffeo2 committed
20
21
 + map to atomic representation
    - add nicks
22
    - handle cylcic dna
cmaffeo2's avatar
cmaffeo2 committed
23
24
25
26
    - shrink ssDNA
    - shrink dsDNA backbone
    - make orientation continuous
    - sequence
27
28
29
 - remove performance bottlenecks
 - test for large systems
 - assign sequence
30
 - document
31
 - rework location class Location 
32
33
34
35
"""

class Location():
    """ Site for connection within an object """
36
    def __init__(self, container, address, type_, on_fwd_strand = True):
37
        ## TODO: remove cyclic references(?)
38
        self.container = container
39
        self.address = address  # represents position along contour length in segments
cmaffeo2's avatar
cmaffeo2 committed
40
        # assert( type_ in ("end3","end5") ) # TODO remove or make conditional
41
        self.on_fwd_strand = on_fwd_strand
42
43
        self.type_ = type_
        self.particle = None
44
        self.connection = None
45

46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
        self.prev_in_strand = None
        self.next_in_strand = None

    def get_connected_location(self):
        if self.connection is None:
            return None
        else:
            return self.connection.other(self)

    def set_connection(self,connection):
        self.connection = connection # TODO weakref? 

    def __repr__(self):
        if self.on_fwd_strand:
            on_fwd = "on_fwd_strand"
        else:
            on_fwd = "on_rev_strand"
        # return "<Location in {} at contour {} {} with connection {}>".format( self.container.name, self.address, self.on_fwd_strand, self.connection )
        # return "<Location {} in {} at contour {} {} with connection {}>".format( self.type_, self.container.name, self.address, on_fwd, self.connection )
        return "<Location {}.{}[{:.2f},{:d}]>".format( self.container.name, self.type_, self.address, self.on_fwd_strand)
        
67
68
69
70
71
72
73
74
75
class Connection():
    """ Abstract base class for connection between two elements """
    def __init__(self, A, B, type_ = None):
        assert( isinstance(A,Location) )
        assert( isinstance(B,Location) )
        self.A = A
        self.B = B
        self.type_ = type_
        
76
77
78
79
80
81
82
83
    def other(self, location):
        if location is self.A:
            return self.B
        elif location is self.B:
            return self.A
        else:
            raise Exception("OutOfBoundsError")
        
84
85
86
# class ConnectableElement(Transformable):
class ConnectableElement():
    """ Abstract base class """
87
88
89
90
    ## TODO: eliminate mutable default arguments
    def __init__(self, connection_locations=[], connections=[]):
        ## TODO decide on names
        self.locations = self.connection_locations = connection_locations
91
92
        self.connections = connections

93
94
95
96
97
98
99
100
101
102
103
104
    def get_locations(self, type_=None, exclude=[]):
        locs = [l for l in self.connection_locations if (type_ is None or l.type_ == type_) and l.type_ not in exclude]
        counter = dict()
        for l in locs:
            if l in counter:
                counter[l] += 1
            else:
                counter[l] = 1
        assert( np.all( [counter[l] == 1 for l in locs] ) )
        return locs                

    def get_connections_and_locations(self, connection_type=None, exclude=[]):
105
106
        """ Returns a list with each entry of the form:
            connection, location_in_self, location_in_other """
107
        type_ = connection_type
108
109
        ret = []
        for c in self.connections:
110
            if (type_ is None or c.type_ == type_) and c.type_ not in exclude:
111
                if   c.A.container is self:
112
                    ret.append( [c, c.A, c.B] )
113
                elif c.B.container is self:
114
115
                    ret.append( [c, c.B, c.A] )
                else:
116
117
                    import pdb
                    pdb.set_trace()
118
119
120
                    raise Exception("Object contains connection that fails to refer to object")
        return ret

121
    def _connect(self, other, connection):
122
123
124
        ## TODO fix circular references        
        A,B = [connection.A, connection.B]
        A.connection = B.connection = connection
125
126
        self.connections.append(connection)
        other.connections.append(connection)
127
128
129
130
131
132
        l = A.container.locations
        if A not in l: l.append(A)
        l = B.container.locations
        if B not in l: l.append(B)
        

133
134
    # def _find_connections(self, loc):
    #     return [c for c in self.connections if c.A == loc or c.B == loc]
135
136
137
138
139
140
141
142
143
144
145
146
147

class SegmentParticle(PointParticle):
    def __init__(self, type_, position, name="A", segname="A", **kwargs):
        self.contour_position = None
        PointParticle.__init__(self, type_, position, name=name, segname=segname, **kwargs)
        self.intrahelical_neighbors = []
        self.other_neighbors = []

    def get_contour_position(self,seg):
        assert( isinstance(seg,Segment) )
        if seg == self.parent:
            return self.contour_position
        else:
148
            ## TODO replace with something more elegant
149
150
151
152
153
154
155
            for c,A,B in self.parent.get_connections_and_locations():
                if A.particle is self and B.container is seg:
                    nt = np.abs( (self.contour_position - A.address)*A.container.num_nts )
                    if B.address < 0.5:
                        return B.address-nt/seg.num_nts
                    else:
                        return B.address+nt/seg.num_nts
cmaffeo2's avatar
cmaffeo2 committed
156
            ## ERROR
157
158
159
160
161
162
163
164
165
166
            print("")
            for c,A,B in self.parent.get_connections_and_locations():
                print("  ",c.type_)
                print(A,B)
                print(A.particle,self)
                print(B.container,seg)
            print("")
            import pdb
            pdb.set_trace()
            raise Exception("Did not find location for particle {} in Segment {}".format(self,seg))
167
168

## TODO break this class into smaller, better encapsulated pieces
169
170
171
172
173
174
175
176
177
178
179
class Segment(ConnectableElement, Group):

    """ Base class that describes a segment of DNA. When built from
    cadnano models, should not span helices """

    """Define basic particle types"""
    dsDNA_particle = ParticleType("D",
                                  diffusivity = 43.5,
                                  mass = 300,
                                  radius = 3,                 
                              )
cmaffeo2's avatar
cmaffeo2 committed
180
181
182
183
184
    orientation_particle = ParticleType("O",
                                        diffusivity = 100,
                                        mass = 300,
                                        radius = 1,
                                    )
185

cmaffeo2's avatar
cmaffeo2 committed
186
    # orientation_bond = HarmonicBond(10,2)
187
    orientation_bond = HarmonicBond(30,1.5, rRange = (0,500) )
188
189
190
191
192
193
194
195
196
197
198
199
200

    ssDNA_particle = ParticleType("S",
                                  diffusivity = 43.5,
                                  mass = 150,
                                  radius = 3,                 
                              )

    def __init__(self, name, num_nts, 
                 start_position = np.array((0,0,0)),
                 end_position = None, 
                 segment_model = None):

        Group.__init__(self, name, children=[])
201
        ConnectableElement.__init__(self, connection_locations=[], connections=[])
202

cmaffeo2's avatar
cmaffeo2 committed
203
204
205
206
207
        self.start_orientation = None
        self.twist_per_nt = 0

        self.beads = [c for c in self.children] # self.beads will not contain orientation beads

208
209
210
        self._bead_model_generation = 0    # TODO: remove?
        self.segment_model = segment_model # TODO: remove?

cmaffeo2's avatar
cmaffeo2 committed
211
        self.num_nts = int(num_nts)
212
213
214
215
216
        if end_position is None:
            end_position = np.array((0,0,self.distance_per_nt*num_nts)) + start_position
        self.start_position = start_position
        self.end_position = end_position

217
218
219
220
        ## Set up interpolation for positions
        a = np.array([self.start_position,self.end_position]).T
        tck, u = interpolate.splprep( a, u=[0,1], s=0, k=1)
        self.position_spline_params = tck
221
222
        
        self.sequence = None
223

224
225
226
227
228
    def clear_all(self):
        Group.clear_all(self)  # TODO: use super?
        self.beads = []
        for c,loc,other in self.get_connections_and_locations():
            loc.particle = None
229
230
231
232
233
234
235
236

    def contour_to_position(self,s):
        p = interpolate.splev( s, self.position_spline_params )
        if len(p) > 1: p = np.array(p).T
        return p

    def contour_to_tangent(self,s):
        t = interpolate.splev( s, self.position_spline_params, der=1 )
237
238
        t = (t / np.linalg.norm(t,axis=0))
        return t.T
239
240
241
242
243
244
245
        

    def contour_to_orientation(self,s):
        if self.start_orientation is not None:
            # axis = self.start_orientation.dot( np.array((0,0,1)) )
            if self.quaternion_spline_params is None:
                axis = self.contour_to_tangent(s)
246
                orientation = rotationAboutAxis( axis, s*self.twist_per_nt*self.num_nts, normalizeAxis=True )
247
248
249
250
251
                ## TODO: ensure this is correct
                # orientation = self.start_orientation.dot(orientation) # .dot( self.start_orientation )
                orientation = orientation.dot( self.start_orientation )
            else:
                q = interpolate.splev(s, self.quaternion_spline_params)
cmaffeo2's avatar
cmaffeo2 committed
252
                q = q/np.linalg.norm(q)
253
                orientation = quaternion_to_matrix(q)
cmaffeo2's avatar
cmaffeo2 committed
254
                
255
256
257
258
        else:
            orientation = None
        return orientation

cmaffeo2's avatar
cmaffeo2 committed
259
260
261
262
    def get_contour_sorted_connections_and_locations(self):
        sort_fn = lambda c: c[1].address
        cl = self.get_connections_and_locations()
        return sorted(cl, key=sort_fn)
263
264
265
266
267
268
269
270
271
272
    
    def randomize_unset_sequence(self):
        bases = list(seqComplement.keys())
        if self.sequence is None:
            self.sequence = [random.choice(bases) for i in range(self.num_nts)]
        else:
            assert(len(self.sequence) == self.num_nts) # TODO move
            for i in range(len(self.sequence)):
                if self.sequence[i] is None:
                    self.sequence[i] = random.choice(bases)
273

cmaffeo2's avatar
cmaffeo2 committed
274
275
276
    def _get_num_beads(self, max_basepairs_per_bead, max_nucleotides_per_bead ):
        raise NotImplementedError

277
    def _generate_one_bead(self, contour_position, nts):
278
279
        raise NotImplementedError

cmaffeo2's avatar
cmaffeo2 committed
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
    def _generate_atomic_nucleotide(self, contour_position, is_fwd, seq):
        """ Seq should include modifications like 5T, T3 Tsinglet; direction matters too """

        # print("Generating nucleotide at {}".format(contour_position))
        
        pos = self.contour_to_position(contour_position)
        if self.local_twist:
            orientation = self.contour_to_orientation(contour_position)
            ## TODO: move this code (?)
            if orientation is None:
                axis = self.contour_to_tangent(contour_position)
                angleVec = np.array([1,0,0])
                if axis.dot(angleVec) > 0.9: angleVec = np.array([0,1,0])
                angleVec = angleVec - angleVec.dot(axis)*axis
                angleVec = angleVec/np.linalg.norm(angleVec)
                y = np.cross(axis,angleVec)
                orientation = np.array([angleVec,y,axis]).T
                ## TODO: improve placement of ssDNA
                # rot = rotationAboutAxis( axis, contour_position*self.twist_per_nt*self.num_nts, normalizeAxis=True )
                # orientation = rot.dot(orientation)
            else:
                orientation = orientation
                            
        else:
            raise NotImplementedError

        # key = self.sequence
        # if self.ntAt5prime is None and self.ntAt3prime is not None: key = "5"+key
        # if self.ntAt5prime is not None and self.ntAt3prime is None: key = key+"3"
        # if self.ntAt5prime is None and self.ntAt3prime is None: key = key+"singlet"

        key = seq

        ## TODO
        if not is_fwd:
            nt_dict = canonicalNtFwd
        else:
            nt_dict = canonicalNtRev
        atoms = nt_dict[ key ].duplicate() # TODO: clone
                        
        # print( atoms.orientation, orientation )
        # atoms.orientation = atoms.orientation.dot( orientation )
        atoms.orientation = orientation.dot(atoms.orientation)
        atoms.position = pos

        ## TODO: scale positions
        return atoms
        
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
    def add_location(self, nt, type_, on_fwd_strand=True):
        ## Create location if needed, add to segment
        c = nt/(self.num_nts-1)
        assert(c >= 0 and c <= 1)
        assert(c*(self.num_nts-1) == round(c*(self.num_nts-1)))
        # TODO? loc = self.Location( address=c, type_=type_, on_fwd_strand=is_fwd )
        loc = Location( self, address=c, type_=type_, on_fwd_strand=on_fwd_strand )
        print("Appending location ",self.name, loc.type_,c, self.num_nts)
        self.locations.append(loc)

    ## TODO? Replace with abstract strand-based model?
    def add_5prime(self, nt, on_fwd_strand=True):
        address = nt/(self.num_nts-1)
        locations = [l for l in self.get_locations("5prime") if
                     l.address == address and
                     l.on_fwd_strand == on_fwd_strand] 
        if len(locations) == 0:
            self.add_location(nt,"5prime",on_fwd_strand)
        else:
            pdb.set_trace()

    def add_3prime(self, nt, on_fwd_strand=True):
        address = nt/(self.num_nts-1) # TODO: put into a fn call
        locations = [l for l in self.get_locations("3prime") if
                     l.address == address and
                     l.on_fwd_strand == on_fwd_strand] 
        if len(locations) == 0:
            self.add_location(nt,"3prime",on_fwd_strand)

cmaffeo2's avatar
cmaffeo2 committed
357
358
359
360
    def get_5prime_locations(self):
        """ Returns tuple of contour_position and direction of strand
        True represents a strand whose 5-to-3 direction increases with contour
        """
361
362
        return [l for l in self.get_locations("5prime")]
        raise NotImplementedError # TODO
cmaffeo2's avatar
cmaffeo2 committed
363

364
    def iterate_connections_and_locations(self, reverse=False):
cmaffeo2's avatar
cmaffeo2 committed
365
366
        ## connections to other segments
        cl = self.get_contour_sorted_connections_and_locations()
367
        if reverse:
cmaffeo2's avatar
cmaffeo2 committed
368
            cl = cl[::-1]
369
370
371
            
        for c in cl:
            yield c
cmaffeo2's avatar
cmaffeo2 committed
372

373
374
375
376
377
378
379
380
381
382
383
    ## TODO rename
    def get_end_of_strand(self, contour_pos, is_fwd):
        """ Walks through locations, checking for crossovers """

        ## Iterate through locations
        # for l in self.locations:
        def loc_iter():
            locations = sorted(self.locations, key=lambda l:l.address)
            if is_fwd:
                for l in locations:
                    yield l
cmaffeo2's avatar
cmaffeo2 committed
384
            else:
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
                for l in locations[::-1]:
                    yield l
            
        for l in loc_iter():
            # if l.particle is None:
            #     pos = l.address
            # else:
            #     pos = l.particle.get_contour_position()          
            pos = l.address

            ## DEBUG
            # if self.name == "1-0" and is_fwd == False:
            #     import pdb
            #     pdb.set_trace()

            ## Skip locations encountered before our strand
cmaffeo2's avatar
cmaffeo2 committed
401
402
403
404
            if is_fwd:
                if pos <= contour_pos: continue
            elif pos >= contour_pos: continue

405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
            ## Stop if we found the 3prime end
            if l.on_fwd_strand == is_fwd and l.type_ == "3prime":
                return 1*is_fwd, None, None, None

            ## Check location connections
            c = l.connection
            if c is None: continue
            B = c.other(l)            

            ## Found a location on the same strand?
            if l.on_fwd_strand == is_fwd:
                # print("from {}, connection {} to {}".format(contour_pos,l,B))
                return pos, B.container, B.address, B.on_fwd_strand
                
            ## Stop at other strand crossovers so basepairs line up
            elif c.type_ == "crossover":
                # print("pausing at {}".format(l))
                return pos, l.container, pos, is_fwd

        ## Made it to the end of the segment without finding a connection
        return 1*is_fwd, None, None, None
cmaffeo2's avatar
cmaffeo2 committed
426
        
427
428
429
430
431
432
433
    def get_nearest_bead(self, contour_position):
        if len(self.beads) < 1: return None
        cs = np.array([b.contour_position for b in self.beads]) # TODO: cache
        # TODO: include connected beads?
        i = np.argmin((cs - contour_position)**2)

        return self.beads[i]
434
435
436

    def get_all_consecutive_beads(self, number):
        assert(number >= 1)
cmaffeo2's avatar
cmaffeo2 committed
437
        ## Assume that consecutive beads in self.beads are bonded
438
        ret = []
cmaffeo2's avatar
cmaffeo2 committed
439
440
        for i in range(len(self.beads)-number+1):
            tmp = [self.beads[i+j] for j in range(0,number)]
441
            ret.append( tmp )
442
        return ret   
443

444
445
446
    def _add_bead(self,b,set_contour=False):
        if set_contour:
            b.contour_position = b.get_contour_position(self)
447
        
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
        # assert(b.parent is None)
        if b.parent is not None:
            b.parent.children.remove(b)
        b.parent = self
        self.children.append(b)
        self.beads.append(b) # don't add orientation bead
        if "orientation_bead" in b.__dict__: # TODO: think of a cleaner approach
            o = b.orientation_bead
            o.contour_position = b.contour_position
            if o.parent is not None:
                o.parent.children.remove(o)
            o.parent = self
            self.children.append(o)
            self.add_bond(b,o, Segment.orientation_bond, exclude=True)

    def _rebuild_children(self, new_children):
        # print("_rebuild_children on %s" % self.name)
        old_children = self.children
        old_beads = self.beads
        self.children = []
        self.beads = []
469
470
471
472
473
474
475
476
477

        if True:
            print("WARNING: DEBUG")
            tmp = []
            for c in new_children:
                if c not in tmp:
                    tmp.append(c)
            new_children = tmp

478
479
480
481
482
        for b in new_children:
            self.beads.append(b)
            self.children.append(b)
            if "orientation_bead" in b.__dict__: # TODO: think of a cleaner approach
                self.children.append(b.orientation_bead)
483
484
485
486
487
            
        # tmp = [c for c in self.children if c not in old_children]
        # assert(len(tmp) == 0)
        # tmp = [c for c in old_children if c not in self.children]
        # assert(len(tmp) == 0)
488
489
        assert(len(old_children) == len(self.children))
        assert(len(old_beads) == len(self.beads))
490

491

cmaffeo2's avatar
cmaffeo2 committed
492
    def _generate_beads(self, bead_model, max_basepairs_per_bead, max_nucleotides_per_bead):
493
494
        
        """ Generate beads (positions, types, etcl) and bonds, angles, dihedrals, exclusions """
cmaffeo2's avatar
cmaffeo2 committed
495
        ## TODO: decide whether to remove bead_model argument
496
        ##       (currently unused)
cmaffeo2's avatar
cmaffeo2 committed
497

498
        ## First find points between-which beads must be generated
cmaffeo2's avatar
cmaffeo2 committed
499
500
        conn_locs = self.get_contour_sorted_connections_and_locations()
        locs = [A for c,A,B in conn_locs]
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
        existing_beads = [l.particle for l in locs if l.particle is not None]
        for b in existing_beads:
            assert(b.parent is not None)

        ## Add ends if they don't exist yet
        ## TODO: what about 1 nt segments?
        if len(existing_beads) == 0 or existing_beads[0].get_contour_position(self) > 0:
            b = self._generate_one_bead(0, 0)
            existing_beads = [b] + existing_beads
        if existing_beads[-1].get_contour_position(self) < 1:
            b = self._generate_one_bead(1, 0)
            existing_beads.append(b)
        assert(len(existing_beads) > 1)

        ## Walk through existing_beads, add beads between
        tmp_children = []       # build list of children in nice order
517
        last = None
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
        for I in range(len(existing_beads)-1):
            eb1,eb2 = [existing_beads[i] for i in (I,I+1)]

            # print(" %s working on %d to %d" % (self.name, eb1.position[2], eb2.position[2]))
            e_ds = eb2.get_contour_position(self) - eb1.get_contour_position(self)
            num_beads = self._get_num_beads( e_ds, max_basepairs_per_bead, max_nucleotides_per_bead )
            ds = e_ds / (num_beads+1)
            nts = ds*self.num_nts
            eb1.num_nts += 0.5*nts
            eb2.num_nts += 0.5*nts

            ## Add beads
            if eb1.parent == self:
                tmp_children.append(eb1)

            s0 = eb1.get_contour_position(self)
            if last is not None:
                last.intrahelical_neighbors.append(eb1)
                eb1.intrahelical_neighbors.append(last)
            last = eb1
            for j in range(num_beads):
                s = ds*(j+1) + s0
                b = self._generate_one_bead(s,nts)

                last.intrahelical_neighbors.append(b)
                b.intrahelical_neighbors.append(last)
                last = b
                tmp_children.append(b)

        last.intrahelical_neighbors.append(eb2)
        eb2.intrahelical_neighbors.append(last)

        if eb2.parent == self:
            tmp_children.append(eb2)
        self._rebuild_children(tmp_children)
553
554
555
556
557
558
559
560
561
562
563
564
565

    def _regenerate_beads(self, max_nts_per_bead=4, ):
        ...
    

class DoubleStrandedSegment(Segment):

    """ Class that describes a segment of ssDNA. When built from
    cadnano models, should not span helices """

    def __init__(self, name, num_nts, start_position = np.array((0,0,0)),
                 end_position = None, 
                 segment_model = None,
cmaffeo2's avatar
cmaffeo2 committed
566
567
                 local_twist = False,
                 num_turns = None,
568
                 start_orientation = None):
cmaffeo2's avatar
cmaffeo2 committed
569
570
571
        
        self.helical_rise = 10.44
        self.distance_per_nt = 3.4
572
573
574
575
576
        Segment.__init__(self, name, num_nts, 
                         start_position,
                         end_position, 
                         segment_model)

cmaffeo2's avatar
cmaffeo2 committed
577
578
579
580
581
582
583
584
585
        self.local_twist = local_twist
        if num_turns is None:
            num_turns = float(num_nts) / self.helical_rise
        self.twist_per_nt = float(360 * num_turns) / num_nts

        if start_orientation is None:
            start_orientation = np.array(((1,0,0),(0,1,0),(0,0,1)))
        self.start_orientation = start_orientation

586
587
        self.nicks = []

588
        self.start = self.start5 = Location( self, address=0, type_= "end5" )
589
        self.start3 = Location( self, address=0, type_ = "end3", on_fwd_strand=False )
590

591
592
593
594
        self.end = self.end3 = Location( self, address=1, type_ = "end3" )
        self.end5 = Location( self, address=1, type_= "end5", on_fwd_strand=False )
        for l in (self.start5,self.start3,self.end3,self.end5):
            self.locations.append(l)
595

596
597
598
599
600
601
602
603
604
        ## Set up interpolation for azimuthal angles 
        a = np.array([self.start_position,self.end_position]).T
        tck, u = interpolate.splprep( a, u=[0,1], s=0, k=1)
        self.position_spline_params = tck
        
        ## TODO: initialize sensible spline for orientation
        self.quaternion_spline_params = None


605
    ## Convenience methods
606
    ## TODO: add errors if unrealistic connections are made
607
    ## TODO: make connections automatically between unconnected strands
608
    def connect_start5(self, end3, type_="intrahelical", force_connection=False):
609
610
        if isinstance(end3, SingleStrandedSegment):
            end3 = end3.end3
611
612
        self._connect_ends( self.start5, end3, type_, force_connection = force_connection )
    def connect_start3(self, end5, type_="intrahelical", force_connection=False):
613
        if isinstance(end5, SingleStrandedSegment):
614
            end5 = end5.start5
615
616
        self._connect_ends( self.start3, end5, type_, force_connection = force_connection )
    def connect_end3(self, end5, type_="intrahelical", force_connection=False):
617
        if isinstance(end5, SingleStrandedSegment):
618
            end5 = end5.start5
619
620
        self._connect_ends( self.end3, end5, type_, force_connection = force_connection )
    def connect_end5(self, end3, type_="intrahelical", force_connection=False):
621
622
        if isinstance(end3, SingleStrandedSegment):
            end3 = end3.end3
623
        self._connect_ends( self.end5, end3, type_, force_connection = force_connection )
624

625
    def add_crossover(self, nt, other, other_nt, strands_fwd=[True,False]):
cmaffeo2's avatar
cmaffeo2 committed
626
627
628
629
630
        """ Add a crossover between two helices """
        ## Validate other, nt, other_nt
        ##   TODO

        ## Create locations, connections and add to segments
631
        c = nt/(self.num_nts-1)
cmaffeo2's avatar
cmaffeo2 committed
632
        assert(c >= 0 and c <= 1)
633
634
635
636
637
638
639
640
641
642
643
644
645
646
        
        def get_loc(seg, address, on_fwd_strand):
            loc = None
            for l in seg.locations:
                if l.address == address and l.on_fwd_strand == on_fwd_strand:
                    assert(loc is None)
                    loc = l
            if loc is None:
                loc = Location( seg, address=address, type_="crossover", on_fwd_strand=on_fwd_strand )
            return loc

        loc = get_loc(self, c, strands_fwd[0])

        c = other_nt/(other.num_nts-1)
cmaffeo2's avatar
cmaffeo2 committed
647
        assert(c >= 0 and c <= 1)
648
        other_loc = get_loc( other, c, strands_fwd[1] )
cmaffeo2's avatar
cmaffeo2 committed
649
        self._connect(other, Connection( loc, other_loc, type_="crossover" ))
cmaffeo2's avatar
cmaffeo2 committed
650

651
    ## Real work
652
    def _connect_ends(self, end1, end2, type_, force_connection):
653
        ## TODO remove self?
654
655
656
657
658
        ## validate the input
        for end in (end1, end2):
            assert( isinstance(end, Location) )
            assert( end.type_ in ("end3","end5") )
        assert( end1.type_ != end2.type_ )
659
        ## Create and add connection
660
        end1.container._connect( end2.container, Connection( end1, end2, type_=type_ ) )
661

662
663
    def _get_num_beads(self, contour, max_basepairs_per_bead, max_nucleotides_per_bead):
        return int(contour*self.num_nts // max_basepairs_per_bead)
cmaffeo2's avatar
cmaffeo2 committed
664

665
666
667
668
    def get_3prime_locations(self):
        ## TODO test
        return self.get_locations("3prime")

cmaffeo2's avatar
cmaffeo2 committed
669
    def get_5prime_locations(self):
670
671
        ## TODO test
        return self.get_locations("5prime")
cmaffeo2's avatar
cmaffeo2 committed
672
673
674
675
676
677
678
679
680
681
682
683
684
        locs = []
        
        ## Add ends of segment if they aren't connected
        cl = self.get_connections_and_locations()
        connlocs = [A for c,A,B in cl]
        if self.start5 not in connlocs:
            locs.append((0,True)) # TODO
        if self.end5 not in connlocs:
            locs.append((1,False)) # TODO
        
        ## TODO Add nicks
        return locs
        
685
686
    def _generate_one_bead(self, contour_position, nts):
        pos = self.contour_to_position(contour_position)
cmaffeo2's avatar
cmaffeo2 committed
687
        if self.local_twist:
688
            orientation = self.contour_to_orientation(contour_position)
cmaffeo2's avatar
cmaffeo2 committed
689
            opos = pos + orientation.dot( np.array((Segment.orientation_bond.r0,0,0)) )
690
691
692
693
694
695
            o = SegmentParticle( Segment.orientation_particle, opos, nts,
                                 num_nts=nts, parent=self )
            bead = SegmentParticle( Segment.dsDNA_particle, pos, nts,
                                    num_nts=nts, parent=self, 
                                    orientation_bead=o,
                                    contour_position=contour_position )
cmaffeo2's avatar
cmaffeo2 committed
696
697

        else:
698
699
700
701
            bead = SegmentParticle( Segment.dsDNA_particle, pos, nts,
                                    num_nts=nts, parent=self,
                                    contour_position=contour_position )
        self._add_bead(bead)
cmaffeo2's avatar
cmaffeo2 committed
702
        return bead
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
        

class SingleStrandedSegment(Segment):

    """ Class that describes a segment of ssDNA. When built from
    cadnano models, should not span helices """

    def __init__(self, name, num_nts, start_position = np.array((0,0,0)),
                 end_position = None, 
                 segment_model = None):

        self.distance_per_nt = 5
        Segment.__init__(self, name, num_nts, 
                         start_position,
                         end_position, 
                         segment_model)

720
        self.start = self.start5 = Location( self, address=0, type_= "end5" ) # TODO change type_?
721
        self.end = self.end3 = Location( self, address=1, type_ = "end3" )
722
723
        for l in (self.start5,self.end3):
            self.locations.append(l)
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741

    def connect_3end(self, end5, force_connection=False):
        self._connect_end( end5,  _5_to_3 = False, force_connection = force_connection )

    def connect_5end(self, end3, force_connection=False):
        self._connect_end( end3,  _5_to_3 = True, force_connection = force_connection )

    def _connect_end(self, other, _5_to_3, force_connection):
        assert( isinstance(other, Location) )
        if _5_to_3 == True:
            my_end = self.end5
            assert( other.type_ == "end3" )
        else:
            my_end = self.end3
            assert( other.type_ == "end5" )

        self._connect( other.container, Connection( my_end, other, type_="intrahelical" ) )

742
743
    def _get_num_beads(self, contour, max_basepairs_per_bead, max_nucleotides_per_bead):
        return int(contour*self.num_nts // max_nucleotides_per_bead)
cmaffeo2's avatar
cmaffeo2 committed
744

cmaffeo2's avatar
cmaffeo2 committed
745
746
747
748
749
750
751
752
753
754
755
    def get_5prime_locations(self):
        locs = []
        
        ## Add ends of segment if they aren't connected
        cl = self.get_connections_and_locations()
        connLocs = [A for c,A,B in cl]
        if self.start not in connLocs:
            locs.append((0,True)) # TODO

        return locs

756
757
    def _generate_one_bead(self, contour_position, nts):
        pos = self.contour_to_position(contour_position)
758
759
760
761
762
        b = SegmentParticle( Segment.ssDNA_particle, pos, nts,
                             num_nts=nts, parent=self,
                             contour_position=contour_position )
        self._add_bead(b)
        return b
763
764

    
cmaffeo2's avatar
cmaffeo2 committed
765
766
767
768
769
770
771
class StrandInSegment(Group):
    """ Class that holds atomic model, maps to segment """
    
    def __init__(self, segment, start, end, is_fwd):
        """ start/end should be provided expressed as contour_length, is_fwd tuples """
        Group.__init__(self)
        self.num_nts = 0
772
        # self.sequence = []
cmaffeo2's avatar
cmaffeo2 committed
773
774
775
776
777
778
        self.segment = segment
        self.start = start
        self.end = end
        self.is_fwd = is_fwd

        ## TODO: get sequence (from segment?)        
779
780
        nts = np.abs(end-start)*(segment.num_nts-1)+1
        self.num_nts = int(round(nts))
cmaffeo2's avatar
cmaffeo2 committed
781
782
        assert( np.abs(self.num_nts-nts) < 0.001 )

783
784
785
786
787
788
789
790
791
792
793
794
795
        # print(" Creating {}-nt StrandInSegment in {} from {} to {} {}".format(self.num_nts, segment.name, start, end, is_fwd))

    def get_sequence(self):
        """ return 5-to-3 """
        seg = self.segment
        nt0 = self.start*(seg.num_nts-1)          # TODO: should it ever be .end?
        assert( nt0 == round(nt0) )
        nt0 = int(nt0)
        assert( (self.end-self.start) >= 0 or not self.is_fwd )
        if self.is_fwd:
            return [seqComplement[seg.sequence[nt]] for nt in range(nt0,nt0+self.num_nts)]
        else:
            return [seg.sequence[nt] for nt in range(nt0,nt0-self.num_nts,-1)]
cmaffeo2's avatar
cmaffeo2 committed
796
797
798
799
800
801
802
803

class Strand(Group):
    """ Class that holds atomic model, maps to segments """
    
    def __init__(self):
        Group.__init__(self)
        self.num_nts = 0
        self.children = self.strand_segments = []
804
805
        
    ## TODO disambiguate names of functions
cmaffeo2's avatar
cmaffeo2 committed
806
807
808
809
810
811
    def add_dna(self, segment, start, end, is_fwd):
        """ start/end should be provided expressed as contour_length, is_fwd tuples """
        s = StrandInSegment( segment, start, end, is_fwd )
        self.strand_segments.append( s )
        self.num_nts += s.num_nts

812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
    def set_sequence(self):
        ...

    # def get_sequence(self):
    #     sequence = []
    #     for ss in self.strand_segments:
    #         sequence.extend( ss.get_sequence() )

    #     assert( len(sequence) >= self.num_nts )
    #     ret = ["5"+sequence[0]] +\
    #           sequence[1:-1] +\
    #           [sequence[-1]+"3"]
    #     assert( len(ret) == self.num_nts )
    #     return ret

cmaffeo2's avatar
cmaffeo2 committed
827
828
829
830
831
    def generate_atomic_model(self):
        last = None
        resid = 1
        for s in self.strand_segments:
            seg = s.segment
832
833
            contour = np.linspace(s.start,s.end,s.num_nts)
            for c,seq in zip(contour,s.get_sequence()):
cmaffeo2's avatar
cmaffeo2 committed
834
                if s.is_fwd:
835
                    nt = seg._generate_atomic_nucleotide( c, s.is_fwd, seq ) # TODO add sequence,termini
cmaffeo2's avatar
cmaffeo2 committed
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
                else:
                    nt = seg._generate_atomic_nucleotide( c, s.is_fwd, "A" )

                s.children.append(nt)
                ## Join last basepairs
                if last is not None:
                    o3,c3,c4,c2,h3 = [last.atoms_by_name[n] 
                                      for n in ("O3'","C3'","C4'","C2'","H3'")]
                    p,o5,o1,o2,c5 = [nt.atoms_by_name[n] 
                                     for n in ("P","O5'","O1P","O2P","C5'")]
                    self.add_bond( o3, p, None )
                    self.add_angle( c3, o3, p, None )
                    for x in (o5,o1,o2):
                        self.add_angle( o3, p, x, None )
                        self.add_dihedral(c3, o3, p, x, None )
                    for x in (c4,c2,h3):
                        self.add_dihedral(x, c3, o3, p, None )
                    self.add_dihedral(o3, p, o5, c5, None)
                nt.__dict__['resid'] = resid
                resid += 1
                last = nt
857
858

class SegmentModel(ArbdModel):
cmaffeo2's avatar
cmaffeo2 committed
859
860
861
    def __init__(self, segments=[], local_twist=True,
                 max_basepairs_per_bead=7,
                 max_nucleotides_per_bead=4,
862
863
864
                 dimensions=(1000,1000,1000), temperature=291,
                 timestep=50e-6, cutoff=50, 
                 decompPeriod=10000, pairlistDistance=None, 
865
866
867
                 nonbondedResolution=0,DEBUG=0):
        self.DEBUG = DEBUG
        if DEBUG > 0: print("Building ARBD Model")
868
869
870
871
872
        ArbdModel.__init__(self,segments,
                           dimensions, temperature, timestep, cutoff, 
                           decompPeriod, pairlistDistance=None,
                           nonbondedResolution=0)

cmaffeo2's avatar
cmaffeo2 committed
873
874
        # self.max_basepairs_per_bead = max_basepairs_per_bead     # dsDNA
        # self.max_nucleotides_per_bead = max_nucleotides_per_bead # ssDNA
875
876
877
878
        self.children = self.segments = segments

        self._bonded_potential = dict() # cache bonded potentials

879
        self._generate_bead_model( max_basepairs_per_bead, max_nucleotides_per_bead, local_twist)
880
881
882

        self.useNonbondedScheme( nbDnaScheme )

883

cmaffeo2's avatar
cmaffeo2 committed
884
    def get_connections(self,type_=None,exclude=[]):
885
886
887
888
        """ Find all connections in model, without double-counting """
        added=set()
        ret=[]
        for s in self.segments:
cmaffeo2's avatar
cmaffeo2 committed
889
            items = [e for e in s.get_connections_and_locations(type_,exclude=exclude) if e[0] not in added]
890
891
892
            added.update([e[0] for e in items])
            ret.extend( items )
        return ret
893
    
894
    def _recursively_get_beads_within_bonds(self,b1,bonds,done=[]):
895
        ret = []
896
897
898
899
900
901
902
903
904
        done = list(done)
        done.append(b1)
        if bonds == 0:
            return [[]]

        for b2 in b1.intrahelical_neighbors:
            if b2 in done: continue
            for tmp in self._recursively_get_beads_within_bonds(b2, bonds-1, done):
                ret.append( [b2]+tmp )
905
906
        return ret

907
908
909
910
911
    def _get_intrahelical_beads(self,num=2):
        ## TODO: add check that this is not called before adding intrahelical_neighbors in _generate_bead_model

        assert(num >= 2)

912
913
        ret = []
        for s in self.segments:
914
915
916
917
918
            for b1 in s.beads:
                for bead_list in self._recursively_get_beads_within_bonds(b1, num-1):
                    assert(len(bead_list) == num-1)
                    if b1.idx < bead_list[-1].idx: # avoid double-counting
                        ret.append([b1]+bead_list)
919
920
        return ret

921

922
923
    def _get_intrahelical_angle_beads(self):
        return self._get_intrahelical_beads(num=3)
924

cmaffeo2's avatar
cmaffeo2 committed
925
    def _get_potential(self, type_, kSpring, d, max_potential = None):
926
927
928
        key = (type_,kSpring,d)
        if key not in self._bonded_potential:
            if type_ == "bond":
929
                self._bonded_potential[key] = HarmonicBond(kSpring,d, rRange=(0,500), max_potential=max_potential)
930
            elif type_ == "angle":
cmaffeo2's avatar
cmaffeo2 committed
931
932
                self._bonded_potential[key] = HarmonicAngle(kSpring,d, max_potential=max_potential)
                # , resolution = 1, maxForce=0.1)
933
            elif type_ == "dihedral":
cmaffeo2's avatar
cmaffeo2 committed
934
                self._bonded_potential[key] = HarmonicDihedral(kSpring,d, max_potential=max_potential)
935
936
937
938
939
940
941
            else:
                raise Exception("Unhandled potential type '%s'" % type_)
        return self._bonded_potential[key]
    def get_bond_potential(self, kSpring, d):
        return self._get_potential("bond", kSpring, d)
    def get_angle_potential(self, kSpring, d):
        return self._get_potential("angle", kSpring, d)
cmaffeo2's avatar
cmaffeo2 committed
942
943
944
945
    def get_dihedral_potential(self, kSpring, d, max_potential=None):
        while d > 180: d-=360
        while d < -180: d+=360
        return self._get_potential("dihedral", kSpring, d, max_potential)
946
947


cmaffeo2's avatar
cmaffeo2 committed
948
949
950
    def _getParent(self, *beads ):
        ## TODO: test
        if np.all( [b1.parent == b2.parent 
951
                    for b1,b2 in zip(beads[:-1],beads[1:])] ):
cmaffeo2's avatar
cmaffeo2 committed
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
            return beads[0].parent
        else:
            return self

    def _get_twist_spring_constant(self, sep):
        """ sep in nt """
        kT = 0.58622522         # kcal/mol
        twist_persistence_length = 90  # set semi-arbitrarily as there is a large spread in literature
        ## <cos(q)> = exp(-s/Lp) = integrate( cos[x] exp(-A x^2), {x, 0, pi} ) / integrate( exp(-A x^2), {x, 0, pi} )
        ##   Assume A is small
        ## int[B_] :=  Normal[Integrate[ Series[Cos[x] Exp[-B x^2], {B, 0, 1}], {x, 0, \[Pi]}]/
        ##             Integrate[Series[Exp[-B x^2], {B, 0, 1}], {x, 0, \[Pi]}]]

        ## Actually, without assumptions I get fitFun below
        ## From http://www.annualreviews.org/doi/pdf/10.1146/annurev.bb.17.060188.001405
        ##   units "3e-19 erg cm/ 295 k K" "nm" =~ 73
        Lp = twist_persistence_length/0.34 

        fitFun = lambda x: np.real(erf( (4*np.pi*x + 1j)/(2*np.sqrt(x)) )) * np.exp(-1/(4*x)) / erf(2*np.sqrt(x)*np.pi) - np.exp(-sep/Lp)
        k = opt.leastsq( fitFun, x0=np.exp(-sep/Lp) )
        return k[0][0] * 2*kT*0.00030461742
973
974
975
976
977
978

    """ Mapping between different resolution models """
    def _clear_beads(self):
        for s in self.segments:
            s.clear_all()
        self.clear_all(keep_children=True)
979
980
981
982
983
984
985
986
987
988
        assert( len([b for b in self]) == 0 )
        locParticles = []
        # for c,A,B in self.get_connections():
        for s in self.segments:
            for c,A,B in s.get_connections_and_locations():
                for l in (A,B):
                    if l.particle is not None:
                        locParticles.append(A.particle)
        assert( len(locParticles) == 0 )
        assert( len([b for s in self.segments for b in s.beads]) == 0 )
989
990
991
992

    def _update_segment_positions(self, bead_coordinates):
        """ Set new function for each segments functions
        contour_to_position and contour_to_orientation """
cmaffeo2's avatar
cmaffeo2 committed
993

994
        for s in self.segments:
995
996
997
            ## TODO: may need to make spline on helices continuously differentiable
            # beads = s.beads # [b for b in s.beads]
            beads = list(set(s.beads + [A.particle for c,A,B in s.get_connections_and_locations("intrahelical")]))
998
999
            ## Skip beads that are None (some locations were not assigned a particle to avoid double-counting) 
            beads = [b for b in beads if b is not None]
1000
            contours = [b.get_contour_position(s) for b in beads]
For faster browsing, not all history is shown. View entire blame