def dstruc_loadin(self, curs):
'''
'''
sys.stderr.write("Loading Data STructure...\n")
from codense.common import get_go_no2go_id, get_gene_no2gene_id, get_go_no2name, get_gene_id2gene_no, get_gene_no2go_no
self.go_no2go_id = get_go_no2go_id(curs)
self.go_no2go_name = get_go_no2name(curs)
self.gene_no2gene_id = get_gene_no2gene_id(curs)
self.gene_id2gene_no = get_gene_id2gene_no(curs)
self.global_gene_to_go_dict = get_gene_no2go_no(curs)
#04-01-05 the second kind in label_dict
gene_no2no = {}
for gene_no in self.gene_no2gene_id:
gene_no2no[gene_no] = gene_no
self.label_dict = {1:self.gene_no2gene_id,
2: gene_no2no}
curs.execute("select gene_no,go_functions from gene")
if self.type == 3:
curs.execute("select array_upper(recurrence_array,1) from %s limit 1"%self.table)
rows = curs.fetchall()
self.no_of_datasets = int(rows[0][0])
sys.stderr.write("Done\n")
def dstruc_loadin(self, curs):
"""
03-09-05
get the context from mcl_table via linking through mcl_id of p_gene_table
context_dict is set
"""
from codense.common import get_known_genes_dict, get_go_no2go_id, get_go_no2name, get_gene_no2gene_id
self.known_genes_dict = get_known_genes_dict(curs)
self.go_no2go_id = get_go_no2go_id(curs)
self.go_no2go_name = get_go_no2name(curs)
self.gene_no2gene_id = get_gene_no2gene_id(curs)
sys.stderr.write("Setting up gene_prediction_dict...")
# setup self.gene_prediction_dict
curs.execute(
"select p.gene_no, p.go_no, p.is_correct, p.is_correct_l1, p.is_correct_lca, m.vertex_set\
from %s p, %s g, %s m where g.p_gene_id=p.p_gene_id and m.mcl_id=p.mcl_id"
% (self.gene_table, self.table, self.mcl_table)
)
rows = curs.fetchall()
for row in rows:
gene_no = row[0]
if self.type == 2 and gene_no not in self.known_genes_dict:
# I only want the known genes, but this gene is unknown
continue
elif self.type == 3 and gene_no in self.known_genes_dict:
# i only want the unknown genes, but this gene is known
continue
go_no = row[1]
is_correct = row[2]
is_correct_l1 = row[3]
is_correct_lca = row[4]
vertex_set = row[5][1:-1].split(",")
vertex_set = map(int, vertex_set)
item = function_struc()
item.is_correct = is_correct
item.is_correct_l1 = is_correct_l1
item.is_correct_lca = is_correct_lca
# context_dict is a set
item.context_dict = Set(vertex_set)
if gene_no not in self.gene_prediction_dict:
self.gene_prediction_dict[gene_no] = gene_prediction()
self.gene_prediction_dict[gene_no].p_functions_struc_dict[go_no] = item
else:
self.gene_prediction_dict[gene_no].p_functions_struc_dict[go_no] = item
sys.stderr.write("Done\n")
"""
def run(self): """ 09-28-05 12-19-05 use class_list and output_fname_list to ease program writing 12-30-05 fix a bug in indexing darwin_instance_list 2006-09-25 2007-02-08 add context_prediction_csv_format """ tf_darwin_ofname = os.path.join(self.output_dir, '%s.tf.darwin'%self.cluster_bs_table) cluster_darwin_ofname = os.path.join(self.output_dir, '%s.cluster.darwin'%os.path.basename(self.input_fname)) prediction_darwin_ofname = os.path.join(self.output_dir, '%s.prediction.darwin'%os.path.basename(self.input_fname)) pattern_darwin_ofname = os.path.join(self.output_dir, '%s.pattern.darwin'%self.pattern_table) if not os.path.isdir(self.output_dir): os.makedirs(self.output_dir) conn, curs = db_connect(self.hostname, self.dbname, self.schema) tax_id = org2tax_id(self.organism) #gene_no2id = get_gene_no2gene_id(curs) #Watch, if unigene, should use this. gene_id2symbol = get_gene_id2gene_symbol(curs, tax_id) gene_id2symbol = self.replace_prime_in_gene_id2symbol(gene_id2symbol) #01-26-06 #gene_no2symbol = dict_transfer(gene_no2id, gene_id2symbol) #Jasmine wants the gene symbol 09-28-05 #gene_id is integer in gene.gene table and same as gene_no, so just use it. go_no2name = get_go_no2name(curs) #09-28-05 Jasmine wants the go_name, not go_id #2006-09-25 use gene_id2symbol to replace mt_no2tf_name #mt_no2tf_name = get_mt_no2tf_name() mt_no2tf_name = gene_id2symbol class_list = [tf_darwin_format, cluster_darwin_format, prediction_darwin_format, pattern_darwin_format, context_prediction_csv_format] context_prediction_csv_fname = os.path.join(self.output_dir, '%s.context.csv'%self.input_fname) output_fname_list = [tf_darwin_ofname, cluster_darwin_ofname, prediction_darwin_ofname, pattern_darwin_ofname, context_prediction_csv_fname] darwin_instance_list = [] for i in range(len(self.running_bit)): if self.running_bit[i] == '1': darwin_instance_list.append(class_list[i](self.hostname, self.dbname, self.schema, self.pattern_table,\ self.cluster_bs_table, self.input_fname, self.lm_bit, self.acc_cut_off, \ output_fname_list[i], gene_id2symbol, go_no2name, mt_no2tf_name, debug, report)) #2006-09-25 current_pos = len(darwin_instance_list)-1 #12-30-05 darwin_instance_list[current_pos].start() for i in range(len(darwin_instance_list)): darwin_instance_list[i].join()
def output(self, curs, outf, known_gene_no2p_gene_id_src, unknown_gene_no2p_gene_id_src, p_gene_id_src_map):
"""
03-03-05
loop over gene_no2p_gene_id_src and p_gene_id_src_map
03-13-05
add a column, #clusters in the output file
--output_one_gene()
--output_function_group()
"""
#three dictionaries
gene_no2gene_id = get_gene_no2gene_id(curs)
gene_no2direct_go = get_gene_no2direct_go(curs)
go_no2go_id = get_go_no2go_id(curs)
go_no2name = get_go_no2name(curs)
go_no2accuracy, go_no2accuracy_pair = self.get_go_no2accuracy(curs, self.p_gene_table, self.gene_p_table)
sys.stderr.write("Outputing prediction table...")
writer = csv.writer(outf, delimiter='\t')
#first output the known genes
for (gene_no, p_gene_id_src_list) in known_gene_no2p_gene_id_src.iteritems():
self.output_one_gene(curs, writer, gene_no, gene_no2gene_id, gene_no2direct_go)
row = ['go_no', 'go_id', 'go_name', 'is_correct', 'is_correct_L1', 'is_correct_lca', 'p_value_list', '#clusters', 'mcl_id_list', \
'e_acc', 'e_acc_pair', 'cluster_context']
writer.writerow(row)
for p_gene_id_src in p_gene_id_src_list:
self.output_function_group(curs, writer, p_gene_id_src_map[p_gene_id_src], gene_no2gene_id,\
go_no2go_id, go_no2name, go_no2accuracy, go_no2accuracy_pair)
writer.writerow([])
#second output the unknown genes
for (gene_no, p_gene_id_src_list) in unknown_gene_no2p_gene_id_src.iteritems():
self.output_one_gene(curs, writer, gene_no, gene_no2gene_id, gene_no2direct_go)
row = ['go_no', 'go_id', 'go_name', 'is_correct', 'is_correct_L1', 'is_correct_lca', 'p_value_list', '#clusters', 'mcl_id_list', \
'e_acc', 'e_acc_pair', 'cluster_context']
writer.writerow(row)
for p_gene_id_src in p_gene_id_src_list:
self.output_function_group(curs, writer, p_gene_id_src_map[p_gene_id_src], gene_no2gene_id,\
go_no2go_id, go_no2name, go_no2accuracy, go_no2accuracy_pair)
writer.writerow([])
del writer
sys.stderr.write("Done\n")
def output1(self, curs, outf, known_gene_no2p_gene_id_src, unknown_gene_no2p_gene_id_src, p_gene_id_src_map):
"""
03-15-05
copied from output()
"""
#three dictionaries
gene_no2gene_id = get_gene_no2gene_id(curs)
gene_no2direct_go = get_gene_no2direct_go(curs)
go_no2go_id = get_go_no2go_id(curs)
go_no2name = get_go_no2name(curs)
go_no2accuracy, go_no2accuracy_pair = self.get_go_no2accuracy(curs, self.p_gene_table, self.gene_p_table)
from codense.common import get_prediction_pair2lca_list
prediction_pair2lca_list = get_prediction_pair2lca_list(curs,p_gene_table=self.p_gene_table)
sys.stderr.write("Outputing prediction table...")
writer = csv.writer(outf, delimiter='\t')
#first output the known genes
for (gene_no, p_gene_id_src_list) in known_gene_no2p_gene_id_src.iteritems():
self.output_one_gene1(curs, writer, gene_no, gene_no2gene_id, gene_no2direct_go)
row = ['go_id', 'go_name', 'is_correct_lca', 'lca_list', 'p_value_list', '#clusters',\
'e_acc']
writer.writerow(row)
for p_gene_id_src in p_gene_id_src_list:
#NOTE: the arguments passed to this function is different between known and unknown genes.
self.output_function_group1(curs, writer, p_gene_id_src_map[p_gene_id_src], gene_no2gene_id,\
go_no2go_id, go_no2name, go_no2accuracy, go_no2accuracy_pair, prediction_pair2lca_list, gene_no)
writer.writerow([])
#second output the unknown genes
for (gene_no, p_gene_id_src_list) in unknown_gene_no2p_gene_id_src.iteritems():
self.output_one_gene1(curs, writer, gene_no, gene_no2gene_id, gene_no2direct_go)
row = ['go_id', 'go_name', 'p_value_list', '#clusters', 'e_acc']
writer.writerow(row)
for p_gene_id_src in p_gene_id_src_list:
self.output_function_group1(curs, writer, p_gene_id_src_map[p_gene_id_src], gene_no2gene_id,\
go_no2go_id, go_no2name, go_no2accuracy, go_no2accuracy_pair)
writer.writerow([])
del writer
sys.stderr.write("Done\n")
