thewall commited on
Commit
bb1bdca
1 Parent(s): a078e4d

Update jolma_subset.py

Browse files

remove deepbindweight
remove protein and aptamer prefix

Files changed (1) hide show
  1. jolma_subset.py +42 -43
jolma_subset.py CHANGED
@@ -1,9 +1,7 @@
1
- import os
2
  import re
3
  import pandas as pd
4
- import numpy as np
5
  import datasets
6
-
7
 
8
  logger = datasets.logging.get_logger(__name__)
9
 
@@ -54,32 +52,19 @@ _URL = "ftp://ftp.sra.ebi.ac.uk/vol1/run/"
54
  # _REVERSE_PRIMER = "CCTATGCGTGCTAGTGTGA"
55
  # _DESIGN_LENGTH = 30
56
 
57
- import datasets
58
 
59
- config = datasets.load_dataset(path="thewall/deepbindweight", split="all")
60
- info = pd.read_excel(config['selex'][0])
61
- protein_info = pd.read_excel(config['tf'][0], index_col=0)
62
- _URLS = {"min_count_10": "",
63
- "min_count_3": ""}
64
- _DESIGN_LENGTH = {"min_count_10": None,
65
- "min_count_3": None}
66
  pattern = re.compile("(\d+)")
67
- for idx, row in info.iterrows():
68
- sra_id = row["SRA ID"]
69
- file = row["file"]
70
- _URLS[sra_id] = "/".join([_URL, sra_id[:6], sra_id, file])
71
- _DESIGN_LENGTH[sra_id] = int(pattern.search(row["Ligand"]).group(0))
72
-
73
  URL = "https://huggingface.co/datasets/thewall/jolma_subset/resolve/main"
74
 
 
75
  class JolmaSubsetConfig(datasets.BuilderConfig):
76
- def __init__(self, length_match=True, design_length=None, filter_N=True,
77
- protein_prefix="1", protein_suffix="2", max_length=1000, max_gene_num=1,
78
- aptamer_prefix="[BOS]", aptamer_suffix="[EOS]", **kwargs):
79
  super(JolmaSubsetConfig, self).__init__(**kwargs)
80
- self.length_match = length_match
81
- self.design_length = design_length
82
- self.filter_N = filter_N
83
  self.data_dir = kwargs.get("data_dir")
84
  self.protein_prefix = protein_prefix
85
  self.protein_suffix = protein_suffix
@@ -90,8 +75,12 @@ class JolmaSubsetConfig(datasets.BuilderConfig):
90
 
91
 
92
  class JolmaSubset(datasets.GeneratorBasedBuilder):
 
 
 
 
93
  BUILDER_CONFIGS = [
94
- JolmaSubsetConfig(name=key, design_length=_DESIGN_LENGTH[key]) for key in ["min_count_3", "min_count_10"]
95
  ]
96
 
97
  DEFAULT_CONFIG_NAME = "min_count_3"
@@ -114,11 +103,29 @@ class JolmaSubset(datasets.GeneratorBasedBuilder):
114
  citation=_CITATION,
115
  )
116
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
117
  def _split_generators(self, dl_manager):
118
- # downloaded_files = dl_manager.download_and_extract(self.config.url)
119
- # logger.info(f"Download from {self.config.url}")
120
  file = dl_manager.download(f"{URL}/{self.config.name}.gz.csv")
121
- # file = os.path.join(filepath, os.listdir(filepath)[0])
122
  return [
123
  datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": file}),
124
  ]
@@ -126,23 +133,24 @@ class JolmaSubset(datasets.GeneratorBasedBuilder):
126
  def _generate_examples(self, filepath):
127
  """This function returns the examples in the raw (text) form."""
128
  logger.info("generating examples from = %s", filepath)
129
- proteins = protein_info["Sequence"]
130
- protein_id = protein_info["Entry"]
131
- gene_num = protein_info["Unique Gene"]
132
  data = pd.read_csv(filepath)
133
  for key, row in data.iterrows():
134
  sra_id = row["identifier"].split(":")[0]
135
- protein_seq = f"{self.config.protein_prefix}{proteins.loc[sra_id]}{self.config.protein_suffix}"
 
 
 
 
136
  aptamer_seq = f'{self.config.aptamer_prefix}{row["seq"]}{self.config.aptamer_suffix}'
137
  if len(protein_seq)>self.config.max_length:
138
  continue
139
- if gene_num.loc[sra_id]>self.config.max_gene_num:
140
  continue
141
- if str(proteins.loc[sra_id])=="nan":
142
  continue
143
  ans = {"id": key,
144
  "protein": protein_seq,
145
- "protein_id": protein_id.loc[sra_id],
146
  "seq": aptamer_seq,
147
  "identifier": row["identifier"],
148
  "count": int(row["count"]),
@@ -150,15 +158,6 @@ class JolmaSubset(datasets.GeneratorBasedBuilder):
150
  yield key, ans
151
 
152
 
153
- def filter_fn(self, example):
154
- seq = example["seq"]
155
- if self.config.length_match and len(seq)!=self.config.design_length:
156
- return False
157
- if self.config.filter_N and "N" in seq:
158
- return False
159
- return True
160
-
161
-
162
  if __name__=="__main__":
163
  from datasets import load_dataset
164
  dataset = load_dataset("jolma_subset.py", split="all")
 
 
1
  import re
2
  import pandas as pd
 
3
  import datasets
4
+ from functools import cached_property, cache
5
 
6
  logger = datasets.logging.get_logger(__name__)
7
 
 
52
  # _REVERSE_PRIMER = "CCTATGCGTGCTAGTGTGA"
53
  # _DESIGN_LENGTH = 30
54
 
 
55
 
56
+ _DOWNLODE_MANAGER = datasets.DownloadManager()
57
+ _RESOURCE_URL = "https://huggingface.co/datasets/thewall/DeepBindWeight/resolve/main"
58
+ SELEX_INFO_FILE = _DOWNLODE_MANAGER.download(f"{_RESOURCE_URL}/ERP001824-deepbind.xlsx")
59
+ PROTEIN_INFO_FILE = _DOWNLODE_MANAGER.download(f"{_RESOURCE_URL}/ERP001824-UniprotKB.xlsx")
 
 
 
60
  pattern = re.compile("(\d+)")
 
 
 
 
 
 
61
  URL = "https://huggingface.co/datasets/thewall/jolma_subset/resolve/main"
62
 
63
+
64
  class JolmaSubsetConfig(datasets.BuilderConfig):
65
+ def __init__(self, protein_prefix="", protein_suffix="", max_length=1000, max_gene_num=1,
66
+ aptamer_prefix="", aptamer_suffix="", **kwargs):
 
67
  super(JolmaSubsetConfig, self).__init__(**kwargs)
 
 
 
68
  self.data_dir = kwargs.get("data_dir")
69
  self.protein_prefix = protein_prefix
70
  self.protein_suffix = protein_suffix
 
75
 
76
 
77
  class JolmaSubset(datasets.GeneratorBasedBuilder):
78
+
79
+ SELEX_INFO = pd.read_excel(SELEX_INFO_FILE, index_col=0)
80
+ PROTEIN_INFO = pd.read_excel(PROTEIN_INFO_FILE, index_col=0)
81
+
82
  BUILDER_CONFIGS = [
83
+ JolmaSubsetConfig(name=key) for key in ["min_count_3", "min_count_10"]
84
  ]
85
 
86
  DEFAULT_CONFIG_NAME = "min_count_3"
 
103
  citation=_CITATION,
104
  )
105
 
106
+ @cached_property
107
+ def selex_info(self):
108
+ return self.SELEX_INFO.loc[self.config.name]
109
+
110
+ @cached_property
111
+ def protein_info(self):
112
+ return self.PROTEIN_INFO.loc[self.config.name]
113
+
114
+ def design_length(self):
115
+ return int(pattern.search(self.protein_info["Ligand"]).group(0))
116
+
117
+ def get_selex_info(self, sra_id):
118
+ return self.SELEX_INFO.loc[sra_id]
119
+
120
+ def get_protein_info(self, sra_id):
121
+ return self.PROTEIN_INFO.loc[sra_id]
122
+
123
+ @cache
124
+ def get_design_length(self, sra_id):
125
+ return int(pattern.search(self.get_protein_info(sra_id)["Ligand"]).group(0))
126
+
127
  def _split_generators(self, dl_manager):
 
 
128
  file = dl_manager.download(f"{URL}/{self.config.name}.gz.csv")
 
129
  return [
130
  datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": file}),
131
  ]
 
133
  def _generate_examples(self, filepath):
134
  """This function returns the examples in the raw (text) form."""
135
  logger.info("generating examples from = %s", filepath)
 
 
 
136
  data = pd.read_csv(filepath)
137
  for key, row in data.iterrows():
138
  sra_id = row["identifier"].split(":")[0]
139
+ protein_info = self.get_protein_info(sra_id)
140
+ proteins = protein_info["Sequence"]
141
+ gene_num = protein_info["Unique Gene"]
142
+ protein_id = protein_info["Entry"]
143
+ protein_seq = f"{self.config.protein_prefix}{proteins}{self.config.protein_suffix}"
144
  aptamer_seq = f'{self.config.aptamer_prefix}{row["seq"]}{self.config.aptamer_suffix}'
145
  if len(protein_seq)>self.config.max_length:
146
  continue
147
+ if gene_num>self.config.max_gene_num:
148
  continue
149
+ if str(proteins)=="nan" or len(str(proteins))==0:
150
  continue
151
  ans = {"id": key,
152
  "protein": protein_seq,
153
+ "protein_id": protein_id,
154
  "seq": aptamer_seq,
155
  "identifier": row["identifier"],
156
  "count": int(row["count"]),
 
158
  yield key, ans
159
 
160
 
 
 
 
 
 
 
 
 
 
161
  if __name__=="__main__":
162
  from datasets import load_dataset
163
  dataset = load_dataset("jolma_subset.py", split="all")