-
Notifications
You must be signed in to change notification settings - Fork 30
/
Copy pathtest_preprocess.py
215 lines (182 loc) · 6.3 KB
/
test_preprocess.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
import anndata
import numpy as np
import pandas as pd
import pytest
from grelu.data.preprocess import (
check_chrom_ends,
filter_blacklist,
filter_cells,
filter_chrom_ends,
filter_coverage,
filter_overlapping,
merge_intervals_by_column,
split,
)
def test_split():
"""
Test dataset splitting by chromosome
"""
intervals = pd.DataFrame(
{
"chrom": ["chr1", "chr2", "chr3", "chr4", "chrX", "chr1"],
"start": [1, 10, 1, 1, 1, 1],
"end": [3, 13, 3, 3, 3, 3],
}
)
# Interval input
train, val, test = split(
intervals, train_chroms="autosomes", val_chroms=["chr3"], test_chroms=["chr4"]
)
assert train.equals(intervals.iloc[[0, 1, 5], :])
assert val.equals(intervals.iloc[[2], :])
assert test.equals(intervals.iloc[[3], :])
# AnnData input
ad = anndata.AnnData(np.random.rand(4, 6), dtype=np.float32)
ad.var = intervals
ad.var.index = ad.var.index.astype(str)
train_ad, val_ad, test_ad = split(
ad, train_chroms="autosomes", val_chroms=["chr3"], test_chroms=["chr4"]
)
assert train_ad.obs.equals(val_ad.obs)
assert train_ad.obs.equals(test_ad.obs)
assert train_ad.var.equals(ad.var.iloc[[0, 1, 5], :])
assert val_ad.var.equals(ad.var.iloc[[2], :])
assert test_ad.var.equals(ad.var.iloc[[3], :])
def test_filter_coverage():
"""
Test filtering intervals by maximum coverage
"""
ad = anndata.AnnData(np.array([[0, 1, 2, 5], [0, 4, 0, 7]]), dtype=np.float32)
# Max, cutoff
ad_filtered = filter_coverage(ad, aggfunc=np.max, cutoff=1, method="cutoff")
assert ad_filtered.obs.equals(ad.obs)
assert np.all(ad_filtered.var_names == ["1", "2", "3"])
# Top 2
ad_filtered = filter_coverage(ad, aggfunc=np.mean, cutoff=2, method="top")
assert ad_filtered.obs.equals(ad.obs)
assert np.all(ad_filtered.var_names == ["1", "3"])
def test_filter_cells():
"""
test filtering cell types by number of cells
"""
ad = anndata.AnnData(np.random.rand(4, 6), dtype=np.float32)
ad.obs = pd.DataFrame(
{"cell_type": ["A", "B", "C", "D"], "n_cells": [500, 1500, 10, 3000]}
)
ad.obs.index = ad.obs.index.astype(str)
ad_filtered = filter_cells(ad, cutoff=1000, count_key="n_cells")
assert ad_filtered.var.equals(ad.var)
assert ad_filtered.obs.equals(ad.obs.iloc[[1, 3], :])
def test_filter_overlapping():
intervals = pd.DataFrame(
{
"chrom": ["chr10", "chr10", "chr10"],
"start": [10, 1000, 45000],
"end": [1010, 2000, 46000],
}
)
ref_intervals = pd.DataFrame(
{
"chrom": ["chr10", "chr10"],
"start": [100, 900],
"end": [200, 970],
}
)
# No window, overlapping
assert filter_overlapping(intervals, ref_intervals).equals(intervals.iloc[[0], :])
# Window, non-overlapping
assert filter_overlapping(intervals, ref_intervals, window=50, invert=True).equals(
intervals.iloc[[2], :]
)
def test_filter_blacklist():
intervals = pd.DataFrame(
{
"chrom": ["chr10", "chr10", "chr10", "chr10", "chr10"],
"start": [10, 1000, 45000, 46000, 48000],
"end": [1010, 2000, 46000, 47000, 49000],
}
)
assert filter_blacklist(intervals, genome="hg38").equals(intervals.iloc[-2:, :])
chrom_end_intervals = pd.DataFrame(
{
"chrom": ["chr1", "chr1", "chr1", "chr1", "chr1"],
"start": [-10, 10, 1000, 248956300, 248956350],
"end": [90, 110, 1100, 248956400, 248956450],
}
)
def test_filter_chrom_ends():
assert filter_chrom_ends(chrom_end_intervals, genome="hg38").equals(
chrom_end_intervals.iloc[[1, 2, 3], :]
)
assert filter_chrom_ends(chrom_end_intervals, genome="hg38", pad=100).equals(
chrom_end_intervals.iloc[[2], :]
)
def test_check_chrom_ends():
with pytest.raises(Exception) as e_info:
check_chrom_ends(chrom_end_intervals, genome="hg38")
assert (
str(e_info.value)
== "Indices of intervals that extend beyond the chromosome ends: 0,4."
)
check_chrom_ends(chrom_end_intervals.iloc[1:2], genome="hg38")
def test_merge_intervals_by_column():
intervals = pd.DataFrame(
{
"chrom": ["chr10", "chr10", "chr10", "chr10", "chr10"],
"start": [10, 1000, 46000, 45000, 48000],
"end": [1010, 2000, 47000, 47500, 49000],
"strand": ["+", "+", "-", "-", "-"],
"gene": ["A", "A", "B", "B", "B"],
}
)
# Test merge
merged = merge_intervals_by_column(intervals, group_col="gene")
assert merged.equals(
pd.DataFrame(
{
"gene": ["A", "B"],
"chrom": ["chr10", "chr10"],
"start": [10, 45000],
"end": [2000, 49000],
"strand": ["+", "-"],
}
)
)
# Test merge without strand
merged = merge_intervals_by_column(
intervals[["chrom", "start", "end", "gene"]], group_col="gene"
)
assert merged.equals(
pd.DataFrame(
{
"gene": ["A", "B"],
"chrom": ["chr10", "chr10"],
"start": [10, 45000],
"end": [2000, 49000],
}
)
)
# Test merge with non-unique chromosomes
with pytest.raises(AssertionError):
intervals = pd.DataFrame(
{
"chrom": ["chr10", "chr10", "chr1", "chr10", "chr10"],
"start": [10, 1000, 45000, 46000, 48000],
"end": [1010, 2000, 46000, 47000, 49000],
"strand": ["+", "+", "-", "-", "-"],
"gene": ["A", "A", "B", "B", "B"],
}
)
merge_intervals_by_column(intervals, group_col="gene")
# Test merge with non-unique strand
with pytest.raises(AssertionError):
intervals = pd.DataFrame(
{
"chrom": ["chr10", "chr10", "chr10", "chr10", "chr10"],
"start": [10, 1000, 45000, 46000, 48000],
"end": [1010, 2000, 46000, 47000, 49000],
"strand": ["+", "+", "+", "-", "-"],
"gene": ["A", "A", "B", "B", "B"],
}
)
merge_intervals_by_column(intervals, group_col="gene")