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 | class CoolerBedGraph:
"""Generates a bedgraph file from a cooler file created by interactions-store.
Attributes:
cooler (cooler.Cooler): Cooler file to use for bedgraph production
capture_name (str): Name of capture probe being processed.
sparse (bool): Only output bins with interactions.
only_cis (bool): Only output cis interactions.
"""
def __init__(
self,
uri: str,
sparse: bool = True,
only_cis: bool = False,
region_to_limit: str = None,
):
"""
Args:
uri (str): Path to cooler group in hdf5 file.
sparse (bool, optional): Only output non-zero bins. Defaults to True.
"""
self._sparse = sparse
self._only_cis = only_cis
logger.info(f"Reading {uri}")
self._cooler = cooler.Cooler(uri)
self.viewpoint_name = self._cooler.info["metadata"]["viewpoint_name"]
self._viewpoint_bins = self._cooler.info["metadata"]["viewpoint_bins"]
if len(self._viewpoint_bins) > 1:
self.multiple_viewpoint_bins = True
logger.warning(
f"Viewpoint {self.viewpoint_name} has multiple bins! {self._viewpoint_bins}. Proceed with caution!"
)
else:
self.multiple_viewpoint_bins = False
self.viewpoint_chroms = self._cooler.info["metadata"]["viewpoint_chrom"]
self.n_cis_interactions = self._cooler.info["metadata"]["n_cis_interactions"]
logger.info(f"Processing {self.viewpoint_name}")
if only_cis:
pixels = []
bins = []
for chrom in self.viewpoint_chroms:
_bins = self._cooler.bins().fetch(chrom)
viewpoint_chrom_bins = self._bins["name"]
_pixels = (
self._cooler.pixels()
.fetch(self.viewpoint_chroms)
.query(
"(bin1_id in @viewpoint_chrom_bins) and (bin2_id in @viewpoint_chrom_bins)"
)
)
_bins = self._cooler.bins().fetch(chrom)
pixels.append(_pixels)
bins.append(_bins)
self._pixels = pd.concat(pixels)
self._bins = pd.concat(bins)
elif region_to_limit:
self._pixels = self._cooler.pixels().fetch(region_to_limit)
self._bins = self._cooler.bins().fetch(region_to_limit)
else:
self._pixels = self._cooler.pixels()[:]
# TODO: Avoid this if possible as reading all bins into memory
self._bins = self._cooler.bins()[:]
# Ensure name column is present
self._bins = (
self._bins.assign(name=lambda df: df.index)
if "name" not in self._bins.columns
else self._bins
)
self._reporters = None
def _get_reporters(self):
logger.info("Extracting reporters")
concat_ids = pd.concat([self._pixels["bin1_id"], self._pixels["bin2_id"]])
concat_ids_filt = concat_ids.loc[lambda ser: ser.isin(self._viewpoint_bins)]
pixels = self._pixels.loc[concat_ids_filt.index]
df_interactions = pd.DataFrame()
df_interactions["capture"] = np.where(
pixels["bin1_id"].isin(self._viewpoint_bins),
pixels["bin1_id"],
pixels["bin2_id"],
)
df_interactions["reporter"] = np.where(
pixels["bin1_id"].isin(self._viewpoint_bins),
pixels["bin2_id"],
pixels["bin1_id"],
)
df_interactions["count"] = pixels["count"].values
return df_interactions.sort_values(["capture", "reporter"]).reset_index(
drop=True
)
def extract_bedgraph(
self, normalisation: Literal["raw", "n_cis", "region"] = "raw", **norm_kwargs
) -> pd.DataFrame:
logger.info("Generating bedgraph")
df_bdg = (
self._bins.merge(
self.reporters,
left_on="name",
right_on="reporter",
how="inner" if self._sparse else "outer",
)[["chrom", "start", "end", "count"]]
.assign(count=lambda df: df["count"].fillna(0))
.sort_values(["chrom", "start"])
)
# TODO: This is a hack to deal with multiple bins for a viewpoint
if self.multiple_viewpoint_bins:
gr_bdg = pr.PyRanges(
df_bdg.rename(
columns={"chrom": "Chromosome", "start": "Start", "end": "End"}
)
)
df_bdg = (
gr_bdg.cluster()
.df.groupby("Cluster")
.agg(
{
"count": "sum",
"Start": "min",
"End": "max",
"Chromosome": "first",
}
)
.reset_index()
.rename(
columns={"Start": "start", "End": "end", "Chromosome": "chrom"}
)[["chrom", "start", "end", "count"]]
)
if not normalisation == "raw":
logger.info("Normalising bedgraph")
self._normalise_bedgraph(df_bdg, method=normalisation, **norm_kwargs)
return df_bdg
@property
def reporters(self) -> pd.DataFrame:
"""Interactions with capture fragments/bins.
Returns:
pd.DataFrame: DataFrame containing just bins interacting with the capture probe.
"""
if self._reporters is not None:
return self._reporters
else:
self._reporters = self._get_reporters()
return self._reporters
def _normalise_bedgraph(
self, bedgraph, scale_factor=1e6, method: str = "n_cis", region: str = None
) -> pd.DataFrame:
"""Normalises the bedgraph (in place).
Uses the number of cis interactions to normalise the bedgraph counts.
Args:
scale_factor (int, optional): Scaling factor for normalisation. Defaults to 1e6.
Returns:
pd.DataFrame: Normalised bedgraph formatted DataFrame
"""
if method == "raw":
pass
elif method == "n_cis":
self._normalise_by_n_cis(bedgraph, scale_factor)
elif method == "region":
self._normalise_by_regions(bedgraph, scale_factor, region)
def _normalise_by_n_cis(self, bedgraph, scale_factor: float):
bedgraph["count"] = (bedgraph["count"] / self.n_cis_interactions) * scale_factor
def _normalise_by_regions(self, bedgraph, scale_factor: float, regions: str):
if not is_valid_bed(regions):
raise ValueError(
"A valid bed file is required for region based normalisation"
)
df_viewpoint_norm_regions = pd.read_csv(
regions, sep="\t", names=["chrom", "start", "end", "name"]
)
df_viewpoint_norm_regions = df_viewpoint_norm_regions.loc[
lambda df: df["name"].str.contains(self.viewpoint_name)
]
counts_in_regions = []
for region in df_viewpoint_norm_regions.itertuples():
counts_in_regions.append(
bedgraph.query(
"(chrom == @region.chrom) and (start >= @region.start) and (start <= @region.end)"
)
)
df_counts_in_regions = pd.concat(counts_in_regions)
total_counts_in_region = df_counts_in_regions["count"].sum()
bedgraph["count"] = (bedgraph["count"] / total_counts_in_region) * scale_factor
def to_pyranges(
self, normalisation: Literal["raw", "n_cis", "region"] = "raw", **norm_kwargs
):
return pr.PyRanges(
self.extract_bedgraph(
normalisation=normalisation, norm_kwargs=norm_kwargs
).rename(columns={"chrom": "Chromosome", "start": "Start", "end": "End"})
)
|