-
Notifications
You must be signed in to change notification settings - Fork 0
/
mgf_read_match.py
287 lines (211 loc) · 8.19 KB
/
mgf_read_match.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
import numpy as np
import pandas as pd
import timeit
import random
from functools import reduce
import sys
print('Running version:')
print(sys.version)
# Write a function for mgf reading into an object
# In[2]:
def load_mgf(fname):
'''Read the file into one huge list, without pre-defined format'''
FIELDS = ('TITLE=', 'RTINSECONDS=', 'PEPMASS=', 'CHARGE=', 'SCANS=')
def format_precursor(spectrum):
#Cover for a case when there's no precursor intensity
if ' ' in spectrum['PEPMASS']:
spectrum['PEPMASS'] = [
float(x) for x in spectrum['PEPMASS'].split(' ')
]
else:
spectrum['PEPMASS'] = [ float(spectrum['PEPMASS']),]
#Check the polarity, which may or mmay not be given after the digits
# "2+", "3-" etc
polarityMultiplier = 1
if spectrum['CHARGE'][-1] == '-':
polarityMultiplier = -1
if not spectrum['CHARGE'][-1].isnumeric():
spectrum['CHARGE'] = polarityMultiplier * int( spectrum['CHARGE'][:-1] )
else:
spectrum['CHARGE'] = int( spectrum['CHARGE'] )
return True
def ms_data_to_df(spectrum):
spectrum['ms_data'] = pd.DataFrame(
spectrum['ms_data'],
columns = ('m/z', 'Intensity')
)
return True
spectraList = []
with open(fname, 'r') as fh:
state = False
for line in fh:
if line[0].isnumeric() and state == True:
spectrum['ms_data'].append(
[ float(x) for x in line.rstrip().split(' ') ]
)
elif 'BEGIN IONS' in line:
spectrum = {'ms_data': []}
state = True
elif 'END IONS' in line:
#Do not add the spectrum to the list if it doesn't contain fragment masses
if len(spectrum['ms_data']) > 0:
ms_data_to_df(spectrum)
format_precursor(spectrum)
spectraList.append(spectrum)
state = False
else:
for fieldName in FIELDS:
if fieldName in line and state == True:
spectrum[fieldName[:-1] ] = line.rstrip().split(fieldName)[1]
return spectraList
# In[3]:
fname = 'Yeast_1000spectra.mgf'
# In[4]:
#get_ipython().run_cell_magic('timeit', '-r 20', #'load_mgf(fname)')
t = lambda: load_mgf(fname)
print('Timing the load_mgf function:')
num_reps = 200
r = timeit.repeat(t, repeat = num_reps, number=1)
print(f'Mean time {np.mean(r)*1000:.1f} ms, standard deviation {np.std(r)*1000:.1f} ms for {num_reps} repeats')
# In[5]:
res = load_mgf(fname)
#print(len(res))
# In[6]:
res[500]
# In[7]:
res[500]['ms_data']
# Let's now check each spectrum for known mass differences
# Monoisotopic masses of amino acids:
# In[8]:
AA_DELTAS = {
'G': 57.02147, 'A': 71.03712, 'S': 87.03203, 'P': 97.05277, 'V': 99.06842,
'T': 101.04768, 'Ccam': 160.03065, 'Cmes': 148.996912, 'I/L': 113.08407,
'N': 114.04293, 'D': 115.02695, 'Q': 128.05858, 'K': 128.09497, 'E': 129.0426,
'M': 131.04049, 'Mox': 147.0354, 'H': 137.05891, 'F': 147.06842, 'R': 156.10112,
'Y': 163.06333, 'W': 186.07932
}
# In[9]:
#Flatten the values from the dictionary
singleResDeltas = np.array(
list( AA_DELTAS.values() ), dtype = 'float64'
)
#print(singleResDeltas.dtype)
#Add doubly-charged and triply-charged mass Deltas (simply divide by 2 and 3)
singleResDeltas = np.concatenate(
(
singleResDeltas,
singleResDeltas / 2,
singleResDeltas / 3
)
)
#print(singleResDeltas.shape)
#print(singleResDeltas[:5])
# Now take the spectra one-by-one, find pairwise mass differences and match them to the list.<br>
# * Calculate pairwise absolute differences between the
# * Subtract the experimental mass Deltas from the theoretical
# * Calculate relative difference
# * Select the cases with the relative difference lower than threshold (matches)
# * Summarize and report the matches
# In[10]:
def find_matches(spectra, masses_to_match, rel_tolerance = 1e-5, float_arr_type = 'float64'):
resDict = {
'Spectrum_idx': np.array([], dtype='uint32'),
'Exp_idx': np.array([], dtype='uint32'),
'Library_idx': np.array([], dtype='uint32'),
'Rel_error': np.array([], dtype=float_arr_type)
}
#Calculate the minimal value in the list for matching
#and offset it by the matching tolerance
minTheoVal = masses_to_match.min() * (1 - rel_tolerance)
for idx, s in enumerate(spectra):
#Calculate pairwise differences betweeen experimental values
expDeltas = np.subtract.outer(
s['ms_data']['m/z'].to_numpy(), s['ms_data']['m/z'].to_numpy()
)
# Disregard relative deltas that are smaller than the lowest theoretical value
expDeltas = expDeltas[ expDeltas > minTheoVal]
# Calculate relative differences between experimental and theoretical values
relDeltasArr = np.divide(
#Absolute values of the differences between masses
np.abs(
np.subtract.outer(
masses_to_match, expDeltas
)
),
#Means between the masses
(np.add.outer(masses_to_match, expDeltas) / 2)
)
matchingInds = np.where(
pd.DataFrame(
relDeltasArr
).le(rel_tolerance) == True
)
numMatches = matchingInds[0].shape[0]
if numMatches > 0:
resDict['Spectrum_idx'] = np.append(
resDict['Spectrum_idx'],
np.array( [idx, ] * numMatches, dtype='uint32' )
)
resDict['Library_idx'] = np.append(
resDict['Library_idx'], matchingInds[0]
)
resDict['Exp_idx'] = np.append(
resDict['Exp_idx'], matchingInds[1]
)
resDict['Rel_error'] = np.append(
resDict['Rel_error'],
relDeltasArr[ matchingInds[0], matchingInds[1] ]
)
resDF = pd.DataFrame(resDict)
return resDF
# In[11]:
print('Running version:')
print(sys.version)
#get_ipython().run_cell_magic('timeit', '-r 5', #'find_matches(res, singleResDeltas, 1e-5)')
t = lambda: find_matches(res, singleResDeltas, 1e-5)
print('Timing the find_matches function:')
num_reps = 50
r = timeit.repeat(t, repeat=num_reps, number=1)
print(f'Mean time {np.mean(r):.2f} s, standard deviation {np.std(r)*1000:.1f} ms for {num_reps} repeats')
# In[12]:
matches = find_matches(res, singleResDeltas, rel_tolerance = 1e-5)
#print(matches)
# In[13]:
#print(matches[ matches['Spectrum_idx'] == 1 ])
# In[ ]:
# We could also see how quick ar loops and string manipulatios.
# Let's create random sequences of equal length and caluclate thier masses using the good old for loops
aa_curated = [
'G', 'A', 'S', 'P', 'V', 'T', 'N', 'D', 'Q', 'K', 'E', 'M', 'H', 'F', 'R', 'Y', 'W'
]
# Create a list of random peptide sequences
g = lambda: ''.join([ random.choice(aa_curated) for _ in range(20) ])
sequences_list = [ g() for _ in range(10000) ]
print(len(sequences_list))
# Create a function with for loops
def calculate_masses_loop():
masses = []
for i in sequences_list:
mass = 18.010565
for j in i:
mass += AA_DELTAS[j]
masses.append(mass)
return masses
print('Running version:')
print(sys.version)
print('Timing the calculate_masses_loop function:')
num_reps = 200
r = timeit.repeat(calculate_masses_loop, repeat=num_reps, number=1)
print(f'Mean time {np.mean(r)*1000:.2f} ms, standard deviation {np.std(r)*1000:.2f} ms for {num_reps} repeats')
# Redo the function with reduce and list comprehensions
def calculate_masses_reduce():
def find_mass(seq):
return 18 + reduce(
(lambda x, y: x + y),
[ AA_DELTAS[x] for x in seq ]
)
return [ find_mass(x) for x in sequences_list ]
print('Timing the calculate_masses_reduce function:')
num_reps = 200
r = timeit.repeat(calculate_masses_reduce, repeat=num_reps, number=1)
print(f'Mean time {np.mean(r)*1000:.2f} ms, standard deviation {np.std(r)*1000:.2f} ms for {num_reps} repeats')