-
Notifications
You must be signed in to change notification settings - Fork 0
/
milab.py
483 lines (395 loc) · 21.2 KB
/
milab.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
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
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
import glob
import subprocess
import re
import os
import pandas as pd
import json
import seaborn as sns
import matplotlib
import matplotlib.pyplot as plt
import itertools
from numpy.random import default_rng
import numpy as np
from functools import reduce
def minnn_extract(r1, r2, minnn_output_base, output_suffix):
cmd = f'minnn -Xmx25G extract \
--pattern \'(R1:N{{*}})\^tggtatcaacgcagagt(UMI:NNNNtNNNNtNNNN)tc(R2:N{{*}})\' -f \
--json-report "{minnn_output_base}{output_suffix}.json" \
--input "{r1}" "{r2}" \
--output "{minnn_output_base}{output_suffix}.mif"'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def minnn_filter(minnn_output_base, filter_name, input_suffix, output_suffix):
cmd = f'minnn -Xmx25G filter \'{filter_name}\' -f \
--json-report "{minnn_output_base}{output_suffix}.json" \
--input "{minnn_output_base}{input_suffix}.mif" \
--output "{minnn_output_base}{output_suffix}.mif"'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def minnn_sort(minnn_output_base, input_suffix, output_suffix):
cmd = f'minnn -Xmx25G sort -f \
--groups UMI \
--input "{minnn_output_base}{input_suffix}.mif" \
--output "{minnn_output_base}{output_suffix}.mif"'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def minnn_correct(minnn_output_base, input_suffix, output_suffix):
cmd = f'minnn -Xmx25G correct -f \
--threads 10 \
--json-report "{minnn_output_base}{output_suffix}.json" \
--groups UMI \
--input "{minnn_output_base}{input_suffix}.mif" \
--output "{minnn_output_base}{output_suffix}.mif"'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def minnn_consensus(minnn_output_base, input_suffix, output_suffix):
cmd = f'minnn -Xmx25G consensus -f \
--threads 10 \
--json-report "{minnn_output_base}{output_suffix}.json" \
--groups UMI \
--input "{minnn_output_base}{input_suffix}.mif" \
--output "{minnn_output_base}{output_suffix}.mif" > "{minnn_output_base}{output_suffix}.log" 2>&1'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def minnn_mif2fastq(minnn_output_base, input_suffix, fastq_output_base):
cmd = f'minnn -Xmx25G mif2fastq -f \
--input "{minnn_output_base}{input_suffix}.mif" \
--group R1="{fastq_output_base}_R1.fastq" \
R2="{fastq_output_base}_R2.fastq"'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def get_sample_name_from_string(filename):
return re.sub("(?:.*/)?(.*?)(\_L00\d)?(?:_R1.*(?:.fastq|.gz))", r"\1", filename)
def minnn_run(r1, r2, minnn_output, fastq_output):
sample_name = get_sample_name_from_string(r1)
print(sample_name)
minnn_output_base = minnn_output + sample_name
fastq_output_base = fastq_output + sample_name
minnn_extract(r1, r2, minnn_output_base, "_extract")
minnn_filter(minnn_output_base, "NoWildcards(UMI)", "_extract", "_extracted_filter")
minnn_sort(minnn_output_base, "_extracted_filter", "_extracted_filtered_sorted")
minnn_correct(minnn_output_base, "_extracted_filtered_sorted", "_corrected")
minnn_sort(minnn_output_base, "_corrected", "_corrected_sorted")
minnn_consensus(minnn_output_base, "_corrected_sorted", "_consensus")
minnn_filter(minnn_output_base, "MinConsensusReads = 2", "_consensus", "_consensus_filtered2")
minnn_mif2fastq(minnn_output_base, "_consensus_filtered2", fastq_output_base)
def mixcr_run(species, material, Five_end, Three_end, adapters, r1, r2, output_path, analyze_param="", align_param="",
assemble_param=""):
samplename = get_sample_name_from_string(r1)
output_base = output_path + samplename
cmd = f'mixcr analyze amplicon -s {species} --starting-material {material} --5-end {Five_end} --3-end {Three_end} -f\
-j {output_base} --adapters {adapters} {analyze_param}'
if align_param != "":
cmd += f'--align "{align_param}'
if assemble_param != "":
cmd += f'--align "{assemble_param}'
cmd += f'--report {output_base}.report {r1} {r2} {output_base}'
process = subprocess.Popen([cmd], stdout=True, stderr=True, shell=True)
process.wait()
def parse_anchor_points(data):
anchor_points_regex = "^^(?:-?[0-9]*:){8}(?:-?[0-9]*):(?P<CDR3Begin>-?[0-9]*):(?P<V3Deletion>-?[0-9]*):(?P<VEnd>-?[0-9]*):(?P<DStart>-?[0-9]*):(?P<D5Deletion>-?[0-9]*):(?P<D3Deletion>-?[0-9]*):(?P<DEnd>-?[0-9]*):(?P<JStart>-?[0-9]*):(?P<J5Deletion>-?[0-9]*):(?P<CDR3End>-?[0-9]*):(?:-?[0-9]*:){2}(?:-?[0-9]*)$"
data = pd.concat([data, data.refPoints.str.extract(anchor_points_regex, expand=True).apply(pd.to_numeric)], axis=1)
return data
def read_mixcr_table(filename):
result = pd.read_table(filename)
if len(result) == 0:
return None
result = parse_anchor_points(result)
result = result[["cloneCount", "cloneFraction", "nSeqCDR3", "aaSeqCDR3", "allVHitsWithScore",
"allDHitsWithScore", "allJHitsWithScore", "allCHitsWithScore", "VEnd", "DStart", "DEnd", "JStart"]]
result = result.rename(columns={"cloneCount": "count",
"cloneFraction": "frequency",
"nSeqCDR3": "CDR3nt",
"aaSeqCDR3": "CDR3aa",
"allVHitsWithScore": "V",
"allDHitsWithScore": "D",
"allJHitsWithScore": "J",
"allCHitsWithScore": "C"})
result["V"] = result.V.str.replace("\*.*", "")
result["D"] = result.D.fillna("N/A").str.replace("\*.*", "")
result["J"] = result.J.str.replace("\*.*", "")
result["C"] = result.C.str.replace("\*.*", "")
result["N"] = result.apply(lambda row: count_n(row.VEnd, row.DStart, row.DEnd, row.JStart), axis=1)
result["CDR3length"] = result.CDR3nt.str.len()
result["count"] = result["count"].astype(int)
return result
def only_productive(data):
result = data.copy()
result = result.loc[~result["CDR3aa"].str.contains("[\*,_]", regex=True)]
result["frequency"] = result["count"] / result["count"].sum()
return result
def get_key(name, corr_dict, default):
for pattern, value in corr_dict.items():
if pattern in name:
return value
return default
def get_feature(data, feature, weighted=True):
if weighted and data.frequency.sum() != 0:
return (data[feature] * data.frequency).sum() / data.frequency.sum()
elif data.frequency.sum() != 0:
return data[feature].mean()
else:
return 0
def count_n(VEnd, DBegin, DEnd, JBegin):
if pd.notna(DBegin):
if VEnd != DBegin:
VD = DBegin - VEnd
else:
VD = 0
if JBegin != DEnd:
N = JBegin - DEnd
else:
DJ = 0
N = VD + DJ
else:
if JBegin != VEnd:
N = JBegin - VEnd
else:
return 0
if N <= 0:
return 0
else:
return N
def basic_analysis(mixcr_path, chain_dict, material_dict, full_clonesets_export_path, functional_clonesets_export_path,
minnn_path=""):
general_samples_dict = {}
# Create folders if needed
create_folder(full_clonesets_export_path)
create_folder(functional_clonesets_export_path)
# Create list of samples based on .clns filees in mixcr folder
samples = []
for filename in glob.glob(mixcr_path + "*.vdjca"):
samples.append(re.sub("(.*/)(.*)(\.vdjca)", r"\2", filename))
# get chain based on chainDict
for sample in samples:
if "Undetermined" in sample:
continue
print(sample)
chain = get_key(sample, chain_dict, "ALL")
# load sample file
data = read_mixcr_table(mixcr_path + sample + ".clonotypes." + chain + ".txt")
if data is None:
continue
general_samples_dict[sample] = {}
full_sample_path = full_clonesets_export_path + sample + ".txt"
functional_sample_path = functional_clonesets_export_path + sample + ".txt"
# save pretty file to the folder
data.to_csv(full_sample_path, sep="\t", index=False)
data_productive = only_productive(data)
data_productive.to_csv(functional_sample_path, sep="\t", index=False)
general_samples_dict[sample]["fullSamplePath"] = full_sample_path
general_samples_dict[sample]["functionalSamplePath"] = functional_sample_path
general_samples_dict[sample]["material"] = get_key(sample, material_dict, "RNA")
general_samples_dict[sample]["productiveClonesNmbr"] = len(data_productive)
general_samples_dict[sample]["productiveReadsNmbr"] = data_productive["count"].sum()
general_samples_dict[sample]["meanCDR3"] = get_feature(data_productive, "CDR3length", weighted=True)
general_samples_dict[sample]["meanN"] = get_feature(data_productive, "N", weighted=True)
general_samples_dict[sample]["chain"] = chain
if minnn_path != "":
general_samples_dict[sample]["extract_report"] = json.load(open(minnn_path + sample + "_extract.json"))
general_samples_dict[sample]["extracted_filter"] = json.load(
open(minnn_path + sample + "_extracted_filter.json"))
general_samples_dict[sample]["consensus_report"] = json.load(open(minnn_path + sample + "_consensus.json"))
general_samples_dict[sample]["consensus_filter_report"] = json.load(
open(minnn_path + sample + "_consensus_filtered2.json"))
general_samples_dict[sample]["align_report"] = json.load(open(mixcr_path + sample + ".align.jsonl"))
general_samples_dict[sample]["assemble_report"] = json.load(open(mixcr_path + sample + ".assemble.jsonl"))
return general_samples_dict
def get_report(samples_dict, output_path, minnn=True):
# Generating report
report = pd.DataFrame()
for sample, metadata in samples_dict.items():
single = {"Sample_id": sample, "Starting Material": metadata["material"], "Chain": metadata["chain"]}
if minnn:
single["Total reads"] = metadata["extract_report"]["totalReads"]
single["Reads matched pattern"] = metadata["extract_report"]["matchedReads"]
single["Reads passed 'NoWildcards' filter"] = metadata["extracted_filter"]["matchedReads"]
single["Reads used in consensus"] = single["Reads passed 'NoWildcards' filter"] - \
metadata["consensus_report"]["notUsedReadsCount"]
single["Total consensuses"] = metadata["consensus_report"]["consensusReads"]
single["Number of consensuses with overseq more then 2"] = metadata["align_report"]["totalReadsProcessed"]
else:
single["Total reads"] = metadata["align_report"]["totalReadsProcessed"]
single["Aligned consensuses"] = metadata["align_report"]["aligned"]
single["Number of consensuses in clonotypes"] = metadata["assemble_report"]["readsInClones"]
single["Number of clonotypes"] = metadata["assemble_report"]["clones"]
single["Mean weighted CDR3 length"] = metadata["meanCDR3"]
single["Mean weighted insert size"] = metadata["meanN"]
single["Number of productive clonotypes"] = metadata["productiveClonesNmbr"]
single["Fraction of productive clonotypes"] = single["Number of productive clonotypes"] / single["Number of clonotypes"]
single["Number of consensuses in productive clonotypes"] = metadata["productiveReadsNmbr"]
report = report.append(single, ignore_index=True)
columns = ["Sample_id", "Starting Material", "Chain", "Total reads"]
if minnn:
columns.extend(("Reads matched pattern", "Reads passed 'NoWildcards' filter", "Reads used in consensus",
"Total consensuses", "Number of consensuses with overseq more then 2"))
columns.extend(("Aligned consensuses", "Number of consensuses in clonotypes", "Number of clonotypes",
"Number of consensuses in productive clonotypes", "Number of productive clonotypes",
"Fraction of productive clonotypes", "Mean weighted CDR3 length", "Mean weighted insert size"))
report = report[columns]
if not minnn:
report.columns = report.columns.str.replace("consensuses", "reads")
report.sort_values("Sample_id").to_csv(output_path + "report.txt", sep="\t", index=False)
return report
# Returns file1 count, file2 count, file12 count, file21 count, file1 div, file2 div, div12,
# freq1, freq2, freq12, freq21
def intersect_pair(data1, data2, by):
if by == "nt":
on_column = "CDR3nt"
elif by == "aa":
on_column = "CDR3aa"
else:
return ("Wrong intersect parameter")
merge = pd.merge(data1, data2, on=[on_column, "V", "J", "C"], suffixes=('_1', '_2'), how="inner")
return round(data1["count"].sum()), \
round(data2["count"].sum()), \
round(merge.count_1.sum()), \
round(merge.count_2.sum()), \
round(len(data1[on_column].unique())), \
round(len(data2[on_column].unique())), \
round(len(merge[on_column].unique())), \
data1["frequency"].sum(), \
data2["frequency"].sum(), \
merge.frequency_1.sum(), \
merge.frequency_2.sum()
def intersect(samples_dict, chain, by="nt", output_path="", functional=False):
intersect_table = pd.DataFrame()
for pair in list(itertools.combinations(samples_dict, 2)):
row = {}
if samples_dict[pair[0]]["chain"] != chain or samples_dict[pair[1]]["chain"] != chain:
continue
sample1 = get_sample_data(samples_dict, pair[0], functional)
sample2 = get_sample_data(samples_dict, pair[1], functional)
row["Sample_id1"] = pair[0]
row["Sample_id2"] = pair[1]
row["count1"], row["count2"], row["count12"], row["count21"], row["div1"], row["div2"], row["div12"], \
row["freq1"], row["freq2"], row["freq12"], row["freq21"] = intersect_pair(sample1, sample2, by=by)
intersect_table = intersect_table.append(row, ignore_index=True)
if len(intersect_table) != 0:
intersect_table = intersect_table[
["Sample_id1", "Sample_id2", "count1", "count2", "count12", "count21", "div1", "div2",
"div12", "freq1", "freq2", "freq12", "freq21"]]
intersect_table = intersect_table.sort_values(["Sample_id1", "Sample_id2"])
if output_path != "":
intersect_table.to_csv(output_path + "intersect" + chain + ".txt", sep="\t", index=False)
return intersect_table
def vj_usage(samples_dict, segment, chain, functional=False, weighted=True, output_path="", plot=True):
create_folder(output_path)
sample = {}
for sample_id, metadata in samples_dict.items():
if metadata["chain"] != chain:
continue
data = get_sample_data(samples_dict, sample_id, functional)
if weighted:
sample[sample_id] = data.groupby(segment, axis=0)["count"].sum().apply(lambda x: x / data["count"].sum())
else:
sample[sample_id] = data.groupby(segment, axis=0)["count"].count().apply(
lambda x: x / data["count"].count())
result = pd.DataFrame(sample).fillna(0)
if plot:
plot = sns.clustermap(result, z_score=1, cmap="coolwarm", xticklabels=True, yticklabels=True)
if output_path != "":
if plot:
plot.savefig(output_path + segment + ".usage.pdf")
result.to_csv(output_path + segment + ".usage.txt", sep="\t")
return result
def get_sample_data(samples_dict, sample_id, functional=False):
if functional:
return pd.read_table(samples_dict[sample_id]["functionalSamplePath"])
return pd.read_table(samples_dict[sample_id]["fullSamplePath"])
def plot_samples_intersect(samples_dict, chain, functional=False, equalby=None, figzise=(80, 80),
output_path="", ylim=(-0.2, 40000), xlim=(-0.2, 40000)):
if equalby is None:
equalby = ['CDR3nt', 'V', 'J', "C"]
valid_chain_dict = {}
for sampleId, metadata in samples_dict.items():
if metadata["chain"] != chain:
continue
valid_chain_dict[sampleId] = metadata
nmb_of_samples = len(valid_chain_dict)
samples_pair_list = list(itertools.combinations_with_replacement(sorted(valid_chain_dict), 2))
fig, axes = plt.subplots(nrows=nmb_of_samples, ncols=nmb_of_samples, figsize=figzise, sharex=True, sharey=True)
fig.tight_layout(pad=5.0)
k = -1
for i in reversed(range(nmb_of_samples)):
for j in reversed(range(nmb_of_samples)):
if i < j:
axes[i, j].axis('off')
else:
k += 1
sample1 = get_sample_data(valid_chain_dict, samples_pair_list[k][1], functional)
sample2 = get_sample_data(valid_chain_dict, samples_pair_list[k][0], functional)
sample1id = samples_pair_list[k][1]
sample2id = samples_pair_list[k][0]
merge = sample1[['count', 'CDR3nt', "CDR3aa", 'V', 'J', 'C']].merge(
sample2[['count', "CDR3aa", 'CDR3nt', 'V', 'J', 'C']],
how='outer', on=equalby,
suffixes=['_1' + sample1id, '_2' + sample2id]).fillna(0)
sns.set(style="white", color_codes=True)
plot = sns.regplot(ax=axes[i][j], x='count_1' + sample1id, y='count_2' + sample2id,
data=merge, fit_reg=False, scatter_kws={"s": 100})
plot.set(ylim=ylim, xlim=xlim, xscale="symlog", yscale="symlog")
xlabel = sample1id
ylabel = sample2id
if j != 0:
ylabel = ""
if i != nmb_of_samples - 1:
xlabel = ""
axes[i][j].set_xticks([0, 1, 100, 10000])
axes[i][j].set_xticklabels(['0', '1', '100', '10000'], fontsize=35)
axes[i][j].set_yticks([0, 1, 100, 10000])
axes[i][j].set_yticklabels(['0', '1', '100', '10000'], fontsize=35)
axes[i][j].set_xlabel(xlabel, fontsize=40, rotation=45, verticalalignment='top',
horizontalalignment='right')
axes[i][j].set_ylabel(ylabel, fontsize=40, rotation=45, verticalalignment='top',
horizontalalignment='right', y=1.0)
if output_path != "":
create_folder(output_path)
fig.savefig(output_path + ".intersect.pdf", bbox_inches="tight")
fig.savefig(output_path + ".intersect.png", bbox_inches="tight")
def create_folder(path):
if path != "":
output_folder = re.sub("(.*/)(.*)", r"\1", path)
if not os.path.exists(output_folder):
os.makedirs(output_folder)
def filter_samples_dict(samples_dict, patterns: list):
new_dict = {}
for sample_id, metadata in sorted(samples_dict.items()):
if all(c in sample_id for c in patterns):
new_dict[sample_id] = metadata
return new_dict
def downsample(samples_dict, output_folder, x, functional=True):
if output_folder is None:
return "Please specify output folder"
create_folder(output_folder)
updated_samples_dict = {}
if x < 0:
return "X must be a positive number"
for sample, metadata in samples_dict.items():
data = get_sample_data(samples_dict, sample, functional)
rng = default_rng()
downsample_nmbr = x if x < data["count"].sum() else data["count"].sum()
data["count"] = rng.multivariate_hypergeometric(data["count"].astype(np.int64), downsample_nmbr)
data = data.loc[data["count"] != 0]
output_path = output_folder + sample + "_" + str(downsample_nmbr) + ".txt"
data.to_csv(output_path, sep="\t", index=False)
updated_samples_dict[sample] = metadata
updated_samples_dict[sample]["downsamplePath"] = output_path
return updated_samples_dict
def merge_clonesets(samples_list, names_list, on="CDR3nt", value="count", how="outer", function="sum"):
"""Merge clonesets. Function takes a list of samples generated by basic_analysis and a list of labels for each sample.
Indices in both list should correspond. By default merge_clonesets will merge clonesets on CDR3 nucleotide sequence
and return a table with counts for each CDR3 sequence (rows) for every sample (column). Any column can be used as a
merge on (ex. V , J, C, CDR3aa). Value can be set to any numeric column (ex. count, frequency, N, CDR3length) to
demonstrate the value for a particular feature the clonesets were merged on.
Function should be one of: sum, count or mean and will be applied to the value series in each group."""
if len(samples_list) != len(names_list):
return "List of samples and list of names must be the same length"
if not any(function == func for func in ['sum', 'count', 'mean']):
return "Function must be one of: sum, count or mean"
samples = samples_list.copy()
samples = [c.groupby(on).agg(function)[value] for c in samples]
df_final = reduce(lambda left, right: pd.merge(left, right, left_index=True, right_index=True, how=how),
samples).fillna(0)
df_final.columns = names_list
return df_final.reset_index()