evaluate.py 18.2 KB
Newer Older
1
2
3
4
#!/usr/bin/env python3


import sys
5
sys.setrecursionlimit(10000)
6
7
8
9
10
11
import argparse
from termcolor import colored
import networkx as nx


def parse_args():
12
    parser = argparse.ArgumentParser(description='Process a d2 graph (complete graph or path) to evaluate its quality.')
13
    parser.add_argument('filename', type=str,
Yoann Dufresne's avatar
Yoann Dufresne committed
14
                        help='The file to evalute')
15
    parser.add_argument('--type', '-t', choices=["d2", "path", "d2-2annotate", "dgraphs"], default="path", required=True,
16
                        help="Define the data type to evaluate. Must be 'd2' or 'path' or 'd2-2annotate' (Rayan's hack).")
Yoann Dufresne's avatar
Yoann Dufresne committed
17
18
    parser.add_argument('--light-print', '-l', action='store_true',
                        help='Print only wrong nodes and paths')
Yoann Dufresne's avatar
Yoann Dufresne committed
19
    parser.add_argument('--barcode_graph', '-b', help="Path to the barcode graph corresponding to the d2_graph to analyse.")
20
    parser.add_argument('--optimization_file', '-o',
Yoann Dufresne's avatar
Yoann Dufresne committed
21
                        help="If the main file is a d2, a file formatted for optimization can be set. This file will be used to compute the coverage of the longest path on the barcode graph.")
22
23
24
25
26
27
28
29
30
31
32
33
34
35

    args = parser.parse_args()
    return args


def load_graph(filename):
    if filename.endswith('.graphml'):
        return nx.read_graphml(filename)
    elif filename.endswith('.gexf'):
        return nx.read_gexf(filename)
    else:
        print("Wrong file format. Require graphml or gefx format", file=sys.stderr)
        exit()

36
37
38
39
40
41
42
43
44
45
46
def save_graph(g, filename):
    if filename.endswith('.graphml'):
        return nx.write_graphml(g,filename)
    elif filename.endswith('.gexf'):
        return nx.write_gexf(g,filename)
    else:
        print("Wrong file format. Require graphml or gefx format", file=sys.stderr)
        exit()



47

Yoann Dufresne's avatar
Yoann Dufresne committed
48
49
# ------------- Path Graph -------------

Yoann Dufresne's avatar
Yoann Dufresne committed
50
51
52
53
def mols_from_node(node_name):
    return [int(idx) for idx in node_name.split(":")[1].split(".")[0].split("_")]


54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
def parse_udg_qualities(graph):
    """ Compute the quality for the best udgs present in the graph.
        All the node names must be under the format :
        {idx}:{mol1_id}_{mol2_id}_...{molx_id}.other_things_here
        :param graph: The networkx graph representinf the deconvolved graph
        :return: A tuple containing two dictionaries. The first one with theoritical frequencies of each node, the second one with observed frequencies.
    """
    dg_per_node = {}

    for node, data in graph.nodes(data=True):
        str_udg = data["udg"]
        central, h1, h2 = str_to_udg_lists(str_udg)

        if central not in dg_per_node:
            dg_per_node[central] = []
        dg_per_node[central].append(data["udg"])

    for node in dg_per_node:
        print(node, dg_per_node[node])
        print(len(dg_per_node))

    return dg_per_node


Yoann Dufresne's avatar
Yoann Dufresne committed
78
def parse_path_graph_frequencies(graph, barcode_graph):
79
80
81
82
83
84
85
86
    """ Compute appearance frequencies from node names.
        All the node names must be under the format :
        {idx}:{mol1_id}_{mol2_id}_...{molx_id}.other_things_here
        :param graph: The networkx graph representing the deconvolved graph
        :param only_wong: If True, don't print correct nodes
        :param file_pointer: Where to print the output. If set to stdout, then pretty print. If set to None, don't print anything.
        :return: A tuple containing two dictionaries. The first one with theoretical frequencies of each node, the second one with observed frequencies.
    """
Yoann Dufresne's avatar
Yoann Dufresne committed
87
    # Compute origin nodes formatted as `{idx}:{mol1_id}_{mol2_id}_...`
Yoann Dufresne's avatar
Yoann Dufresne committed
88
    observed_frequencies = {}
Yoann Dufresne's avatar
Yoann Dufresne committed
89
    real_frequencies = {}
90
    origin_node_names = []
Yoann Dufresne's avatar
Yoann Dufresne committed
91
92
    node_per_barcode = {}

Yoann Dufresne's avatar
Yoann Dufresne committed
93
    for node, data in graph.nodes(data=True):
94
95
        parsed = parse_dg_name(graph, node)
        origin_name = parsed[0][1]
Yoann Dufresne's avatar
Yoann Dufresne committed
96
97

        if origin_name not in node_per_barcode:
Yoann Dufresne's avatar
Yoann Dufresne committed
98
99
            node_per_barcode[origin_name] = []
        node_per_barcode[origin_name].append(node)
100
101

        # Count frequency
Yoann Dufresne's avatar
Yoann Dufresne committed
102
        if origin_name not in observed_frequencies:
Yoann Dufresne's avatar
Yoann Dufresne committed
103
            observed_frequencies[origin_name] = 0
104
            origin_node_names.append(origin_name)
Yoann Dufresne's avatar
Yoann Dufresne committed
105
        observed_frequencies[origin_name] += 1
106

Yoann Dufresne's avatar
Yoann Dufresne committed
107
108
    # Theoretical frequencies
    real_frequencies = {node_id: len(node_id.split(":")[1].split("_")) for node_id in barcode_graph.nodes()}
109

Yoann Dufresne's avatar
Yoann Dufresne committed
110
    return real_frequencies, observed_frequencies, node_per_barcode
111
112


Yoann Dufresne's avatar
Yoann Dufresne committed
113
114
115
116
117
118
119
120
121
""" This function aims to look for direct molecule neighbors.
    If a node has more than 2 direct neighbors, it's not rightly splitted
"""
def parse_graph_path(graph):
    neighborhood = {}

    for node in graph.nodes():
        molecules = mols_from_node(node)
        neighbors = list(graph.neighbors(node))
122

Yoann Dufresne's avatar
Yoann Dufresne committed
123
124
125
126
127
128
129
130
        neighborhood[node] = []
        for mol in molecules:
            for nei in neighbors:
                nei_mols = mols_from_node(nei)
                if mol-1 in nei_mols:
                    neighborhood[node].append(nei)
                if mol+1 in nei_mols:
                    neighborhood[node].append(nei)
131

Yoann Dufresne's avatar
Yoann Dufresne committed
132
    return neighborhood
133
134


Yoann Dufresne's avatar
Yoann Dufresne committed
135
136
def print_path_summary(frequencies, light_print=False, file_pointer=sys.stdout):
    if file_pointer is None:
Yoann Dufresne's avatar
Yoann Dufresne committed
137
138
        return

Yoann Dufresne's avatar
Yoann Dufresne committed
139
    print("--- Nodes analysis ---", file=file_pointer)
Yoann Dufresne's avatar
Yoann Dufresne committed
140
141
142
    theoretical_frequencies, observed_frequencies, node_per_barcode = frequencies
    for key in theoretical_frequencies:
        obs, the = observed_frequencies[key] if key in observed_frequencies else 0, theoretical_frequencies[key]
Yoann Dufresne's avatar
Yoann Dufresne committed
143
144
145
146
147
        result = f"{key}: {obs}/{the}"

        if file_pointer == sys.stdout:
            result = colored(result, 'green' if obs==the else 'red')

Yoann Dufresne's avatar
Yoann Dufresne committed
148
        if light_print and obs == the:
Yoann Dufresne's avatar
Yoann Dufresne committed
149
150
151
152
153
            continue

        print(result, file=file_pointer)


154
155
    print("--- Global summary ---", file=file_pointer)

Yoann Dufresne's avatar
Yoann Dufresne committed
156
157
    # --- Frequency usage ---
    # Tags
Yoann Dufresne's avatar
Yoann Dufresne committed
158
159
160
    distinct_theoretical_nodes = len(theoretical_frequencies)
    distinct_observed_nodes = len(observed_frequencies)
    print(f"Distinct barcodes: {distinct_observed_nodes}/{distinct_theoretical_nodes}", file=file_pointer)
Yoann Dufresne's avatar
Yoann Dufresne committed
161
    # molecules
Yoann Dufresne's avatar
Yoann Dufresne committed
162
163
164
    cumulative_theoretical_nodes = sum(theoretical_frequencies.values())
    cumulative_observed_nodes = sum(observed_frequencies.values())
    print(f"Molecules: {cumulative_observed_nodes}/{cumulative_theoretical_nodes}", file=file_pointer)
Yoann Dufresne's avatar
Yoann Dufresne committed
165
166
167
    # Wrong splits
    over_split = 0
    under_split = 0
Yoann Dufresne's avatar
Yoann Dufresne committed
168
169
170
171
172
    for barcode in theoretical_frequencies:
        observed = observed_frequencies[barcode] if barcode in observed_frequencies else 0
        theoretic = theoretical_frequencies[barcode]
        over_split += max(observed-theoretic, 0)
        under_split += max(theoretic-observed, 0)
Yoann Dufresne's avatar
Yoann Dufresne committed
173
    print(f"Under/Over splitting: {under_split} - {over_split}")
174
175


176
177
178
179
def print_dgraphs_summary(frequencies, light_print=False):
    pass


Yoann Dufresne's avatar
Yoann Dufresne committed
180
181
# ------------- D2 Graph -------------

182
183
def str_to_udg_lists(s):
    udg = s.replace("]", "").replace(' [', '[')
184
    return udg.split('[')[1:]
185

186
187
def parse_dg_name(gr, name):
    udg = nx.get_node_attributes(gr, 'udg')[name]
Rayan Chikhi's avatar
Rayan Chikhi committed
188
189
190
191
    res = str_to_udg_lists(udg)
    if len(res) != 3:
        print("parsing problem:",res)
    central, h1, h2 = res
Yoann Dufresne's avatar
Yoann Dufresne committed
192
    
193
194
    idx = name
    score = nx.get_node_attributes(gr, 'score')[name]
Yoann Dufresne's avatar
Yoann Dufresne committed
195
196
197
198
199
200
201
202

    # Parse hands
    h1 = h1.split(', ')
    h2 = h2.split(', ')

    return (idx, central, score), h1, h2


203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
def path_to_jumps(path):
    chuncks = []
    prev_start = -1000
    current_molecule = -1000

    for mol, node in path:
        # If there is a gap
        if mol > current_molecule + 1:
            chuncks.append((prev_start, current_molecule))
            prev_start = mol

        current_molecule = mol

    # Add the last piece
    chuncks.append((prev_start, current_molecule))

    del chuncks[0]
    return chuncks


Yoann Dufresne's avatar
Yoann Dufresne committed
223
def print_d2_summary(connected_components, longest_path, coverage_vars=(0, 0), light_print=False):
Yoann Dufresne's avatar
Yoann Dufresne committed
224
225
226
    print("--- Global summary ---")
    print(f"Number of connected components: {len(connected_components)}")
    print(f"Total number of nodes: {sum([len(x) for x in connected_components])}")
Yoann Dufresne's avatar
Yoann Dufresne committed
227
228
    no_singleton = [x for x in connected_components if len(x) > 1]
    print(f"There are {len(no_singleton)} connex components with at least 2 nodes")
Yoann Dufresne's avatar
Yoann Dufresne committed
229
230
231
    print(f"The 5 largest components: {[len(x) for x in connected_components][:5]}")

    print("--- Largest component analysis ---")
232
233
234
235
236
237
238
239
240
    # Get the list of node idx
    path_dg_idx = [int(x[1].split(" ")[0]) for x in longest_path]
    # print("\n".join(longest_path))
    if not light_print:
        print("Longest path for increasing molecule number:")
        print(path_dg_idx)
    print(f"Size of the longest path: {len(longest_path)}")
    print("Jumps in central nodes:")
    print(path_to_jumps(longest_path))
Yoann Dufresne's avatar
Yoann Dufresne committed
241
242
243
244

    print(f"Number of usable coverage variables: {len(coverage_vars[1])}")
    print(f"Coverage: {len(coverage_vars[0])}/{len(coverage_vars[1])}")
    print(f"Missing coverage variables:\n{coverage_vars[1]-coverage_vars[0]}")
Yoann Dufresne's avatar
Yoann Dufresne committed
245
246
247


def compute_next_nodes(d2_component):
248
249
250
251
252
    # First parse dg names
    dg_names = {}
    for node in d2_component.nodes():
        dg_names[node] = parse_dg_name(d2_component,node)

Yoann Dufresne's avatar
Yoann Dufresne committed
253
254
255
256
    next_nodes = {}

    for node in d2_component.nodes():
        # Parse the current node name
257
        head, h1, h2 = dg_names[node]
Yoann Dufresne's avatar
Yoann Dufresne committed
258
259
260
261
262
263
        next_nodes[node] = {}

        # Get the current molecule idxs
        molecule_idxs = mols_from_node(head[1])
        for mol_idx in molecule_idxs:
            nexts = []
264
265
            for neighbor in d2_component[node]:
                nei_head, _, _ = dg_names[neighbor]
Yoann Dufresne's avatar
Yoann Dufresne committed
266
267
268
269
270
271
272
273
274
275
276
277
278
279
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
                nei_mols = mols_from_node(nei_head[1])
                nei_mols = [x for x in nei_mols if x > mol_idx]
                
                # If there are molecule next
                if len(nei_mols) > 0:
                    next_nei_mol = min(nei_mols)
                    nexts.append((next_nei_mol, neighbor))

            nexts.sort(key=lambda x: x[0])
            next_nodes[node][mol_idx] = nexts
            # print(next_nodes)

    return next_nodes


def compute_longest_increasing_paths(d2_component):
    next_nodes = compute_next_nodes(d2_component)

    # Compute the longest path for each node
    longest_paths = {}
    for idx, start_node in enumerate(next_nodes):
        # print(f"{idx}/{len(next_nodes)}")
        for mol_idx in next_nodes[start_node]:
            recursive_longest_path(start_node, mol_idx , next_nodes, longest_paths)

    # Get the longest path size
    max_len, node_val, mol_idx = 0, None, -1
    for node in longest_paths:
        for mol in longest_paths[node]:
            length, _, _ = longest_paths[node][mol]
            if max_len < length:
                max_len = length
                node_val = node
                mol_idx = mol

    # Backtrack the longest path
    path = backtrack_longest_path(node_val, mol_idx, longest_paths)
    return path


def backtrack_longest_path(node, molecule, longest_paths, path=[]):
    if node == None:
        return path

310
    path.append((molecule, node))
Yoann Dufresne's avatar
Yoann Dufresne committed
311
312
313
314
315
316
317
318
319
320
321
    length, next_node, next_mol = longest_paths[node][molecule]
    return backtrack_longest_path(next_node, next_mol, longest_paths, path)


def recursive_longest_path(current_node, current_molecule, next_nodes, longest_paths):
    # Dynamic programming
    if current_node in longest_paths and current_molecule in longest_paths[current_node]:
        return longest_paths[current_node][current_molecule]

    longest = 0
    longest_next = None
322
    min_mol_idx = float('inf') 
Yoann Dufresne's avatar
Yoann Dufresne committed
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
    # Recursively compute the longest path
    for mol_idx, node in next_nodes[current_node][current_molecule]:
        size, _, _ = recursive_longest_path(node, mol_idx, next_nodes, longest_paths)
        if size + 1 > longest:
            longest = size + 1
            longest_next = node
            min_mol_idx = mol_idx
        # If there is an alternative path with shorter distance
        elif size + 1 == longest and mol_idx < min_mol_idx:
            longest = size + 1
            longest_next = node
            min_mol_idx = mol_idx
    
    # Save the result
    if not current_node in longest_paths:
        longest_paths[current_node] = {}
    longest_paths[current_node][current_molecule] = (longest, longest_next, min_mol_idx)
    return longest_paths[current_node][current_molecule]


Yoann Dufresne's avatar
Yoann Dufresne committed
343
344
345
346
def compute_covered_variables(graph, path):
    path_nodes = set()
    for mol, node_name in path:
        path_nodes.add(node_name)
347

Yoann Dufresne's avatar
Yoann Dufresne committed
348
349
350
351
352
353
354
    used_vars = set()
    total_vars = set()
    for node, data in graph.nodes(data=True):
        vars = data["barcode_edges"].split(" ")
        total_vars.update(vars)
        if node in path_nodes:
            used_vars.update(vars)
355

Yoann Dufresne's avatar
Yoann Dufresne committed
356
    return used_vars, total_vars
357

358
359
360
361
362
363
364
365
366
367
368
369
# returns True iff there exist x in mol1 such that there exists y in mol2 and |x-y| <= some_value
def nearby_udg_molecules(mols1, mols2):
    for x in mols1:
        for y in mols2:
            if abs(x-y) <= 5:
                return True
    return False

def verify_graph_edges(d2_component):
    udg_molecules_dict=dict()
    for node in d2_component.nodes():
        # Parse the current node name
370
        head, c1, c2 = parse_dg_name(d2_component,node)
371
372
373

        # Construct the molecule(s) that this udg really 'reflects'
        # i.e. the udg has a central node and two cliques
rchikhi's avatar
rchikhi committed
374
375
        # that central node is the result of merging of several molecules
        # ideally, only one of those molecules is connected to the molecules of the cliques
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
        # (there could be more than one though; in that case the udg is 'ambiguous')
        # udg_molecules aims to reflect the molecule(s) underlying this udg 
        udg_molecules = set()

        # Get the current molecule idxs of central node
        molecule_idxs = mols_from_node(head[1])
        #print("mol idxs", molecule_idxs)

        # Examine molecule idx's of cliques to see which are close to the central node
        # rationale: c1/c2 contain nearby molecule id's
        for mol_idx in molecule_idxs:
            nexts = []
            for c in [c1,c2]:
                for c_node in c:
                    nei_mols = mols_from_node(c_node.split()[0])
391
                    nei_mols = [x for x in nei_mols if x > mol_idx]  # fixme: also look at the <= molecules for more robustness
392
393
394
395
396
397
398
399
                    
                    # If there are molecule next
                    if len(nei_mols) > 0:
                        next_nei_mol = min(nei_mols)
                        nexts.append((next_nei_mol))

            nexts.sort(key=lambda x: x)
            quality = sum([1.0/x if mol_idx+x in nexts else 0 for x in range(1,6)]) / sum([1.0/x for x in range(1,6)])
rchikhi's avatar
rchikhi committed
400
            if quality > 0.6: # eyeballed but still arbitrary
401
402
403
404
405
406
407
408
                udg_molecules.add(mol_idx)
            #print("mol",mol_idx,molecule_idxs,"quality",quality,"nexts",nexts)
       
        udg_molecules_dict[head[0]]=udg_molecules

    # Then: annotate edges as to whether they're real (their udg_molecule(s) are nearby) or fake
    for n1, n2 , data in d2_component.edges(data=True):
        # Parse the current node name
409
        head, c1, c2 = parse_dg_name(d2_component,n1)
410
411
        node_udg_molecules = udg_molecules_dict[head[0]]
        
412
        n_head, n_c1, n_c2 = parse_dg_name(d2_component,n2)
413
414
415
416
417
418
419
        neighbor_udg_molecules = udg_molecules_dict[n_head[0]]
       
        if nearby_udg_molecules(node_udg_molecules, neighbor_udg_molecules):
            color = 'green'
        else:
            color = 'red'
        data['color'] = color
420
        #print("edge",node_udg_molecules,neighbor_udg_molecules,color)
421
422
423
424
    
    # also, annotate nodes by their putative molecule found
    for n, data in d2_component.nodes(data=True):
        # Parse the current node name
425
        head, c1, c2 = parse_dg_name(d2_component,n)
426
427
428
        node_udg_molecules = udg_molecules_dict[head[0]]
        data['udg_molecule']= '_'.join(list(map(str,node_udg_molecules)))
        
429
430
431
432
433
434
435
436
    # aggressive: delete nodes which have either no found udg_molecule, or two udg_molecules
    # turns out it's not a good strategy as the nodes with two udg_molecules are important to connect portions of graph
    # but what if we magically keep those where the two adjacent molecules are close together
    if True:
        d2_component = d2_component.copy()
        nodes_to_remove = []
        for n, data in d2_component.nodes(data=True):
            # Parse the current node name
437
            head, c1, c2 = parse_dg_name(d2_component,n)
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
            if "_" in data['udg_molecule'] or data['udg_molecule'] == '':
                if "_" in data['udg_molecule']:
                    m1, m2 = list(map(int,data['udg_molecule'].split("_")))
                    if abs(m2-m1) < 30: continue # don't remove that kind of nodes
                nodes_to_remove += [n]
        d2_component.remove_nodes_from(nodes_to_remove)
        print("removed",len(nodes_to_remove),"bad nodes")

    # aggressive: delete red edges
    if True:
        d2_component = d2_component.copy()
        edges_to_remove = []
        for n1, n2, data in d2_component.edges(data=True):
            if data['color'] == 'red':
                edges_to_remove += [(n1,n2)]
        d2_component.remove_edges_from(edges_to_remove)
        print("removed",len(edges_to_remove),"bad edges")


    return d2_component
458

459
460
461
462
def main():
    args = parse_args()
    graph = load_graph(args.filename)

Yoann Dufresne's avatar
Yoann Dufresne committed
463
    if args.type == "path":
Yoann Dufresne's avatar
Yoann Dufresne committed
464
465
        barcode_graph = load_graph(args.barcode_graph)
        frequencies = parse_path_graph_frequencies(graph, barcode_graph)
Yoann Dufresne's avatar
Yoann Dufresne committed
466

Yoann Dufresne's avatar
Yoann Dufresne committed
467
        print_path_summary(frequencies, light_print=args.light_print)
468
469
470
    elif args.type == "dgraphs":
        udg_per_node = parse_udg_qualities(graph)
        # print(udg_per_node)
Yoann Dufresne's avatar
Yoann Dufresne committed
471
472
473
474
475
476
    elif args.type == "d2":
        components = list(nx.connected_components(graph))
        components.sort(key=lambda x: -len(x))

        component = graph.subgraph(components[0])
        longest_path = compute_longest_increasing_paths(component)
Yoann Dufresne's avatar
Yoann Dufresne committed
477
478
        vars = compute_covered_variables(graph, longest_path)
        print_d2_summary(components, longest_path, coverage_vars=vars, light_print=args.light_print)
479
480
481
482
483
484
485
486
487
    
    # added by Rayan
    # example:
    # python evaluate.py --type d2-2annotate ~/Dropbox/cnrs/projects/10x-barcodes/yoann_to_cedric_ilp/d2_graph.gexf
    elif args.type == "d2-2annotate":
        components = list(nx.connected_components(graph))
        components.sort(key=lambda x: -len(x))
        component = graph.subgraph(components[0])
 
488
        component = verify_graph_edges(component)
489

Yoann Dufresne's avatar
Yoann Dufresne committed
490
491
492
493
        extension = args.filename.split('.')[-1]
        base_filename = '.'.join(args.filename.split('.')[:-1])
        save_graph(component, base_filename+".verified."+extension)

494
495
496

if __name__ == "__main__":
    main()