-
Notifications
You must be signed in to change notification settings - Fork 0
/
k-means_algorithm.py
328 lines (225 loc) · 11.5 KB
/
k-means_algorithm.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
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
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
import sys
import random
import math
import matplotlib
#matplotlib.use('Agg')
import matplotlib.pyplot as plt
import numpy as np
from Bio.Blast import NCBIWWW
from Bio.Blast import NCBIXML
from Bio import SeqIO
class FastAreader:
'''
Define objects to read FastA files.
instantiation:
thisReader = FastAreader ('testTiny.fa')
usage:
for head, seq in thisReader.readFasta():
print (head,seq)
'''
def __init__(self, fname= 'read1.fa'):
'''contructor: saves attribute fname '''
self.fname = fname
def doOpen(self):
''' Handle file opens, allowing STDIN.'''
if self.fname == '':
return sys.stdin
else:
return open(self.fname)
def readFasta(self):
''' Read an entire FastA record and return the sequence header/sequence'''
header = ''
sequence = ''
with self.doOpen() as fileH:
#header = ''
#sequence = ''
# skip to first fasta header
line = fileH.readline()
while not line.startswith('>'):
line = fileH.readline()
header = line[1:].rstrip()
for line in fileH:
if line.startswith('>'):
yield header, sequence
header = line[1:].rstrip()
sequence = ''
else:
sequence += ''.join(line.rstrip().split()).upper()
yield header, sequence
class k_means:
"""
This class develops genomic categorization.
"""
def __init__(self, fname=''):
'''contructor: saves attribute fname '''
self.fname = fname
self.read_dict = {}
self.k = 9
self.multi_read_cluster_holder_dict = {}
self.centroids_dict = {}
self.multi_read_centroid_change_list = [i for i in range(self.k)]
self.seq_dict = {}
self.sequence_assembly_dict = {}
self.x_coords_dict = {}
self.y_coords_dict = {}
self.random_string = ''
def rand_seq(self, seq):
"""Develops a random string the length of the read sequence for testing and comparison.
Input: original sequence
Output: random sequence of the same length
"""
return ''.join(random.choices(['A','G','T','C'], k=len(seq)))
def pr_matrix(self, head, seq):
"""
Develops a dictionary in which each kmer key is associated with a list of unique (x,y) = (position, frequency) coordinate tuples.
These values are attached to the given kmer and used at y-coordinates for Euclidian distance calculations.
1) The sequence is parsed into 4-mers.
2) The 4-mers are used for keys in the dictionary.
3) The number of times each 4-mer occurs in the sequence is counted and divided by the number of possible 4mers, 256.
Input: FASTA sequence
Output: Dictionary with frequency Pr values attached to kmers.
"""
#seq = self.rand_seq(seq)
self.kmers = [seq[i:i + 4] for i in range(0, len(seq), 4)]
self.seq_dict[head] = seq
self.sequence_assembly_dict[head] = {}
kmer_xy_coordinate_list = []
for i in range(0, len(self.kmers)):
if len(self.kmers[i]) == 4:
kmer_xy_coordinate_list.append((i*4, self.kmers.count(self.kmers[i]) / 256))
self.read_dict[head] = kmer_xy_coordinate_list
def random_centroids(self):
"""Randomly selects k number of initial centroids by position in FASTA sequence.
The initial centroids are attached to the read in the overarching read_dict.
These centroids will be updated and used for cluster identification.
Input: none
Output: dictionary random centroids
"""
for read in self.read_dict.keys():
random_centroids = []
for i in range(0, self.k):
random_centroids.append(self.read_dict[read][random.randint(0, self.k)])
self.centroids_dict[read] = random_centroids
self.multi_read_cluster_holder_dict[read] = {i: [] for i in range(0, self.k)}
def distance_calculation(self):
"""Euclidian distance is calculated between each centroid and all non-self points.
Input: a list of points and centroids.
Output: a list of distances between each centroid and all other points.
"""
for read in self.read_dict.keys():
temp_placement_dict = {i: [] for i in range(0, self.k)}
for point in self.read_dict[read]:
distance_list = []
x2,y2 = point[0], point[1]
for centroid in self.centroids_dict[read]:
x1, y1 = centroid[0], centroid[1]
#distance calculation between centroid and point
distance = math.sqrt(abs(x2 - x1) + abs(y2 - y1))
distance_list.append(distance)
temp_placement_dict[distance_list.index(min(distance_list))].append(point)
for i in range(self.k):
self.multi_read_cluster_holder_dict[read][i] = temp_placement_dict[i]
def centroid_recalculation(self):
"""Recalculates centroids by cluster averaging.
Input: Current cluster information
Output: List of adjusted centroids as component to self.read_dict[read][1]
"""
for read in self.multi_read_cluster_holder_dict.keys():
temp_centroid_list = []
centroid_change_list = []
for i in range(self.k):
if len(self.multi_read_cluster_holder_dict[read][i]) > 0:
new_centroid_x = sum([tup[0] for tup in self.multi_read_cluster_holder_dict[read][i]]) / len(self.multi_read_cluster_holder_dict[read][i])
new_centroid_y = sum([tup[1] for tup in self.multi_read_cluster_holder_dict[read][i]]) / len(self.multi_read_cluster_holder_dict[read][i])
temp_centroid_list.append((new_centroid_x, new_centroid_y))
centroid_change_list.append(abs(new_centroid_x - self.centroids_dict[read][i][0]) + abs(new_centroid_y - self.centroids_dict[read][i][1]))
self.multi_read_centroid_change_list = centroid_change_list
self.centroids_dict[read] = temp_centroid_list
def sequence_assembly(self):
"""Assembling the sequences represented by the predicted clusters.
Input: Finished Clusters
Output: Separated genomic sequences
"""
for read in self.multi_read_cluster_holder_dict.keys():
print(read)
self.x_coords_dict[read] = {i: [] for i in range(0, self.k)}
self.y_coords_dict[read] = {i: [] for i in range(0, self.k)}
for i in self.multi_read_cluster_holder_dict[read].keys():
cluster_sequence = ""
self.sequence_assembly_dict[read][i] = ""
x_coords_list = []
y_coords_list = []
for point in self.multi_read_cluster_holder_dict[read][i]:
cluster_sequence += self.seq_dict[read][point[0]: point[0] + 4]
x_coords_list.append(point[0])
y_coords_list.append(point[1])
self.sequence_assembly_dict[read][i] = cluster_sequence
self.x_coords_dict[read][i] = x_coords_list
self.y_coords_dict[read][i] = y_coords_list
if cluster_sequence != "":
print("cluster: ", i, " cluster size: ", "(", len(self.multi_read_cluster_holder_dict[read][i]), "/", len(self.read_dict[read]), ")", " portion of read: ",
len(self.multi_read_cluster_holder_dict[read][i]) / len(self.read_dict[read]))
print(self.sequence_assembly_dict[read][i])
self.blast(cluster_sequence)
print("_____________________________________________________________________________\n")
print("_____________________________________________________________________________\n")
def cluster_graphing(self):
""" The purpose of this function is to develop graphs of each computational cluster.
Each subcluster will be represented in a different color, with all centroids marked similarly.
Input: Dictionaries of x and y coordinates per cluster per read, dictionaries of centroids
Output: Graphed clusters, each cluster in a distinct color in respect to each read.
"""
for read in self.multi_read_cluster_holder_dict.keys():
for i in self.multi_read_cluster_holder_dict[read].keys():
xpoints_for_plotting = np.array(self.x_coords_dict[read][i])
ypoints_for_plotting = np.array(self.y_coords_dict[read][i])
color_list = ['tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:cyan', 'tab:purple', 'tab:brown',
'tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:cyan', 'tab:purple', 'tab:brown']
plt.scatter(xpoints_for_plotting, ypoints_for_plotting, color=color_list[i], marker='o')
centroids_x_list = [i for i, j in self.centroids_dict[read]]
centroids_y_list = [j for i, j in self.centroids_dict[read]]
centroid_x_points_for_plotting = np.array(centroids_x_list)
centroid_y_points_for_plotting = np.array(centroids_y_list)
plt.scatter(centroid_x_points_for_plotting, centroid_y_points_for_plotting, color='k', marker='*')
plt.title(label=read)
plt.xlabel('kmer position')
plt.ylabel('kmer frequency')
plt.savefig(f"{read}.png")
plt.show()
def blast(self, cluster_sequence):
"""Develops a BLAST request for each clustered sequence via BioPython.
Design taken from Biopython Manual.
"""
E_VALUE_THRESH = 0.0000000000000000001
result_handle = NCBIWWW.qblast("blastn", "nt", sequence=cluster_sequence, expect=1)
blast_record = NCBIXML.read(result_handle)
count = 0
for alignment in blast_record.alignments:
for hsp in alignment.hsps:
if hsp.expect < E_VALUE_THRESH:
count += 1
print(f"sequence: {alignment.title}, length: {alignment.length}nt, e-value: {hsp.expect}")
#print(hsp.query[0:75] + "...")
#print(hsp.match[0:75] + "...")
#print(hsp.sbjct[0:75] + "...")
print(f"There are {count} similar sequences in Blast output\n")
def driver(self):
"""Drivers iterative processes"""
self.random_centroids()
self.distance_calculation()
iteration = 0
while sum(self.multi_read_centroid_change_list) > 0:
self.centroid_recalculation()
self.distance_calculation()
iteration += 1
print("total iterations of k-means clustering: ", iteration)
self.sequence_assembly()
self.cluster_graphing()
def main():
class_access = k_means()
myReader = FastAreader()
for head, seq in myReader.readFasta():
class_access.pr_matrix(head, seq)
class_access.driver()
if __name__ == '__main__':
main()