def fetch_rRNA_genes_from_biomart(org,host='www.ensembl.org'):
'''
'''
try:
if org not in ['hsapiens',
'mmusculus']:
raise ValueError('Organism {0} not supported for biomart lookup'.\
format(org))
s=biomart.BioMart(host=host)
s.new_query()
if org=='hsapiens':
s.add_dataset_to_xml('hsapiens_gene_ensembl')
s.add_attribute_to_xml('hgnc_symbol') # fetch human gene names
elif org=='mmusculus':
s.add_dataset_to_xml('mmusculus_gene_ensembl')
s.add_attribute_to_xml('mgi_symbol') # fetch mouse gene names
else:
raise ValueError('Organism {0} not supported for biomart lookup'.\
format(org))
s.add_filter_to_xml("biotype", "rRNA") # filter query for rRNA genes
s.add_attribute_to_xml('chromosome_name') # fetch chr names
xml=s.get_xml() # fetch xml
res=pd.read_csv(StringIO(s.query(xml)), sep='\t', header=None)
res.columns=['symbol','chromosome_name'] # reformat dataframe
res.set_index('symbol',inplace=True)
res=res[~res.index.duplicated(keep='first')]
return res.index
except:
raise
def biomart_query_all(verbose=False, extra_fields=None, force_download=False):
"""
pulls down all entries from BIOMART for Human: symbol, trasncript, gene, length, type
"""
THE_FILE = pathlib.Path(__file__).parent / 'biomart_all.csv.gz'
if not force_download and os.path.exists(THE_FILE):
return _biomart_df_postprocess(pd.read_csv(THE_FILE, index_col=0))
s = biomart.BioMart(host=HOST)
s.new_query()
s.add_dataset_to_xml('hsapiens_gene_ensembl')
# what we want to get back
# s.add_attribute_to_xml('entrezgene')
fields = [
'hgnc_symbol',
'ensembl_gene_id',
'ensembl_gene_id_version',
'transcript_length',
'ensembl_transcript_id',
'ensembl_transcript_id_version',
'transcript_biotype',
'chromosome_name',
'start_position',
'end_position',
'external_synonym',
]
if extra_fields:
fields.extend(extra_fields)
for f in fields:
s.add_attribute_to_xml(f)
xml = s.get_xml()
if verbose:
print(xml)
res = s.query(xml)
df = pd.read_csv(io.StringIO(res), sep='\t', header=None)
df.columns = fields
df = df.drop_duplicates()
df = _biomart_df_postprocess(df)
df.to_csv(THE_FILE)
return df
def gene_coordinates(host, org, gene, chr_exclude=[]) -> pd.DataFrame:
"""Retrieve gene coordinates for specific organism through BioMart.
Parameters
----------
host : {{'www.ensembl.org', ...}}
A valid BioMart host URL. Can be used to control genome build.
org : {{'hsapiens', 'mmusculus', 'drerio'}}
Organism to query. Currently available are human ('hsapiens'), mouse
('mmusculus') and zebrafish ('drerio').
gene :
The gene symbol (e.g. 'hgnc_symbol' for human) for which to retrieve
coordinates.
chr_exclude :
A list of chromosomes to exclude from query.
Returns
-------
A `pd.DataFrame` containing gene coordinates for the specified gene symbol.
"""
try:
from bioservices import biomart
except ImportError:
raise ImportError(
'You need to install the `bioservices` module.')
from io import StringIO
s = biomart.BioMart(host=host)
# building query
s.new_query()
if org == 'hsapiens':
s.add_dataset_to_xml('hsapiens_gene_ensembl')
s.add_attribute_to_xml('hgnc_symbol')
elif org == 'mmusculus':
s.add_dataset_to_xml('mmusculus_gene_ensembl')
s.add_attribute_to_xml('mgi_symbol')
elif org == 'drerio':
s.add_dataset_to_xml('drerio_gene_ensembl')
s.add_attribute_to_xml('zfin_id_symbol')
else:
logg.debug('organism ', str(org), ' is unavailable', no_indent=True)
return None
s.add_attribute_to_xml('chromosome_name')
s.add_attribute_to_xml('start_position')
s.add_attribute_to_xml('end_position')
xml = s.get_xml()
# parsing gene coordinates
res = pd.read_csv(StringIO(s.query(xml)), sep='\t', header=None)
res.columns = ['symbol', 'chromosome_name', 'start', 'end']
res = res.dropna()
res = res[~res['chromosome_name'].isin(chr_exclude)]
res = res.set_index('symbol')
return res.loc[[gene], :]
def mitochondrial_genes(host, org) -> pd.Index:
"""Mitochondrial gene symbols for specific organism through BioMart.
Parameters
----------
host : {{'www.ensembl.org', ...}}
A valid BioMart host URL.
org : {{'hsapiens', 'mmusculus', 'drerio'}}
Organism to query. Currently available are human ('hsapiens'), mouse
('mmusculus') and zebrafish ('drerio').
Returns
-------
A :class:`pandas.Index` containing mitochondrial gene symbols.
"""
try:
from bioservices import biomart
except ImportError:
raise ImportError(
'You need to install the `bioservices` module.')
from io import StringIO
s = biomart.BioMart(host=host)
# building query
s.new_query()
if org == 'hsapiens':
s.add_dataset_to_xml('hsapiens_gene_ensembl')
s.add_attribute_to_xml('hgnc_symbol')
elif org == 'mmusculus':
s.add_dataset_to_xml('mmusculus_gene_ensembl')
s.add_attribute_to_xml('mgi_symbol')
elif org == 'drerio':
s.add_dataset_to_xml('drerio_gene_ensembl')
s.add_attribute_to_xml('zfin_id_symbol')
else:
logg.debug('organism ', str(org), ' is unavailable', no_indent=True)
return None
s.add_attribute_to_xml('chromosome_name')
xml = s.get_xml()
# parsing mitochondrial gene symbols
res = pd.read_csv(StringIO(s.query(xml)), sep='\t', header=None)
res.columns = ['symbol', 'chromosome_name']
res = res.dropna()
res = res[res['chromosome_name'] == 'MT']
res = res.set_index('symbol')
res = res[~res.index.duplicated(keep='first')]
return res.index
def biomart_query_new(attributes: list,
filter: str,
filter_values: list,
batchsize=10000,
verbose=False):
"""
query biomart similar to the R-function getBM()
:param attributes: which attributes to retrieve from biomart
:param filter: field which to filter on
:param filter_values: return only entries where filter \in filter_values
"""
s = biomart.BioMart(host=HOST)
batch_results = []
n_batches = int(np.ceil(len(filter_values) / batchsize))
for id_batch in tqdm.tqdm(batch(filter_values, batchsize=batchsize),
total=n_batches):
s.new_query()
s.add_dataset_to_xml('hsapiens_gene_ensembl')
# what we want to get back
for a in attributes:
s.add_attribute_to_xml(a)
# the query should be comma separated
# better make sure theres no whitespace
query = ",".join([_.strip() for _ in id_batch])
s.add_filter_to_xml(filter, query)
xml = s.get_xml()
if verbose:
print(xml)
res = s.query(xml)
df = pd.read_csv(io.StringIO(res), sep='\t', header=None)
df.columns = attributes
df = df.drop_duplicates()
batch_results.append(df)
df = pd.concat(batch_results, axis=0).drop_duplicates()
return df
def __init__(self, host='www.ensembl.org'):
self.s = biomart.BioMart(host=host)
def print_attributes():
"""
thats all the atribures that we can get from biomart for a gene/transcript
"""
s = biomart.BioMart(host=HOST)
return list(s.attributes('hsapiens_gene_ensembl').keys())
import sys
from Bio import SeqIO
# Load annotation
gene_table_file = sys.argv[1]
gene_table = pd.read_csv(gene_table_file)
gene_table.set_index('acc', inplace=True)
# Define column data types
gene_table['chromosome'] = gene_table['chromosome'].astype(str)
for col in ['start', 'end', 'length', 'strand']:
gene_table[col] = gene_table[col].astype(int)
# Number of genes to be analyzed at a time
chunk_size = int(sys.argv[2])
# Setup server
b = biomart.BioMart()
b.host = 'www.ensembl.org'
datasets = b.datasets("ENSEMBL_MART_ENSEMBL")
# Set sequence file name
seq_file = 'all_seqs.fa'
fileO = open(seq_file, 'a')
# Record downloaded sequences
downloaded = [
gene.name.split("|")[2] for gene in SeqIO.parse(seq_file, 'fasta')
]
# Filter table so that only non downloaded genes are processed
gene_table = gene_table.loc[~(gene_table.index.isin(downloaded))]
# Group table by species and chromosome