def setup_pfam_mysql(self):
'''
Establishes a connection to Pfam MySQL db using settings from file
'''
if hasattr(self, "pfam_mysql"):
return
self.pfam_mysql = PfamMySQLQuery()
executable_locs = {}
params = {}
for s in ['db_username', 'db_password', 'db_address', 'db_name']:
params[s] = self.settings_repository.get_item_value(('mysql_database_pfam', s))
self.pfam_mysql.setup_caller(executable_locations=executable_locs, params=params)
class SifterAnalysisProject(GitAnalysisProject):
'''
This project controller extends the "GitAnalysisProject"
adding lots of functionality specific to doing a SIFTER analysis
'''
implements(pc.IAnalysisProject)
def __init__(self):
'''
Initializes an empty query container
'''
self.query_collection = SifterQueryCollection()
def parse_hypothetical_pfam_domains(self, pfam_results):
'''
Given pfam_results (from PfamHMMERQuery().parse_results()
Creates domain_architecture entries for queries
'''
# For each HMMER hit, adds a domain entry.
for i in range(len(pfam_results)):
#pprint(vars(pfam_hmm_scan.parsed_results[i]))
if not hasattr(pfam_results[i], "_annotations"):
continue
seq_id = pfam_results[i]._annotations['seqName']
domain_id = pfam_results[i]._annotations['hmmName']
new_domain_region = ProteinInformation(
identifier=domain_id,
content = pfam_results[i],
retrieval_method="PFamQuery",
retrieval_date=datetime.now().strftime("%Y-%m-%d %H:%M"))
q = self.query_collection.get_query(query_id=seq_id)
p = q.protein_collection.get_protein(protein_id=seq_id)
p.domain_architecture.add_domain_region(domain_region=new_domain_region)
def setup_pfam_mysql(self):
'''
Establishes a connection to Pfam MySQL db using settings from file
'''
if hasattr(self, "pfam_mysql"):
return
self.pfam_mysql = PfamMySQLQuery()
executable_locs = {}
params = {}
for s in ['db_username', 'db_password', 'db_address', 'db_name']:
params[s] = self.settings_repository.get_item_value(('mysql_database_pfam', s))
self.pfam_mysql.setup_caller(executable_locations=executable_locs, params=params)
def get_pfam_mysql_for_each_query(self, db_table, db_field, extra_where_str, file_postfix, gzipped=True):
'''
For each query, extracts the Pfam MSA from MySQL.
'''
self.setup_pfam_mysql()
# For each query, loop over each domain and get each
# Pfam MSA from the pre-computed Pfam table.
for q in self.query_collection:
p = q.protein_collection.get_protein(q.query_id)
cache_ids = []
for d in p.domain_architecture:
outpt_fname = os.path.dirname(q.destination) + "/" + d.identifier + file_postfix
if d.identifier in cache_ids:
# If we've already queried this then no need to duplicate.
continue
else:
# Get file from MySQL query, gunzip, and save into query directory.
mysql_aq = "(select auto_pfamA from pfamA where pfamA_id='" + d.identifier + "')"
mysql_q = "select "+db_field+" from "+db_table+" " \
+ "where auto_pfamA = "+mysql_aq+" " \
+ extra_where_str + ";"
print mysql_q
self.pfam_mysql.call({'query': mysql_q})
self.pfam_mysql.parse_results()
f_output = open(outpt_fname, "w")
if gzipped:
f_gzipped = StringIO.StringIO(self.pfam_mysql.parsed_results[0][0][0])
f = gzip.GzipFile(fileobj=f_gzipped, mode='rbU')
f_output.write(f.read())
f.close()
f_gzipped.close()
else:
f_output.write(self.pfam_mysql.parsed_results[0][0][0])
f_output.close()
# Store id into cache to not retrieve multiple times
cache_ids.append(d.identifier)
# Store reference to files created in the query.
if d.meta_data is None:
d.meta_data = {}
d.meta_data[db_field] = os.path.basename(outpt_fname)
def make_phylo_placers_for_queries(self):
'''
Creates phylogenetic placement infrastructure for
queries.
This will use the ThirdPartyWrappers.PplacerWrapper
'''
for q in self.query_collection:
p = q.protein_collection.get_protein(q.query_id)
for d in p.domain_architecture:
# Get file from MySQL query, gunzip, and save into query directory.
#d.identifier
#d.meta_data[db_field]
pass
def make_queries_pfamscan(self, query_file):
'''
Calls HMMER on PFam HMM files
Creates ProteinCollection, adding a scaffold AnnotatedProtein for the
query sequences in query_file.
'''
# Create project infrastructure for queries
self.query_collection.create_query_scaffolds(query_file=query_file,
destination_base_dir=self.file_repository.identity)
self.query_collection.export_to_files()
self.file_repository.save_repository(save_description="Query scaffolds created.")
# Call HMMER on Pfam HMMs for the input fasta file
pfam_hmm_scan = PfamHMMERQuery()
# Get necessary executable locations from project settings
executable_locs = {}
executable_locs['hmmpress'] = self.settings_repository.get_item_value(('executable_locations', 'hmmpress'))
executable_locs['hmmscan'] = self.settings_repository.get_item_value(('executable_locations', 'hmmscan'))
# Get necessary parameters from project settings
params = {}
params['pfam_db_loc'] = self.settings_repository.get_item_value(('local_database_locations', 'pfam_data'))
params['query_sequences_fasta_file'] = query_file
# Make call and do initial parsing of results into format for ProteinInformation retrievers.
pfam_hmm_scan.setup_caller(executable_locations=executable_locs, params=params)
pfam_hmm_scan.call()
pfam_hmm_scan.parse_results()
# Create domain regions for each HMM hit
self.parse_hypothetical_pfam_domains(pfam_results=pfam_hmm_scan.parsed_results)
self.query_collection.export_to_files()
self.file_repository.save_repository(save_description="Hypothetical domain architectures from Pfam HMMs written to query files.")
# Get and gunzip MSAs, trees.
self.get_pfam_mysql_for_each_query(db_table="alignments_and_trees",
db_field="alignment",
extra_where_str="and type='full'",
file_postfix=".sto")
self.get_pfam_mysql_for_each_query(db_table="alignments_and_trees",
db_field="tree",
extra_where_str="and type='full'",
file_postfix=".tree")
self.query_collection.export_to_files()
self.file_repository.save_repository(save_description="Exported alignments and trees into query directories from Pfam MySQL.")
'''
