def main():
special = None
no_threads = 1
db = connect_to_mysql()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species(cursor)
total = 0
for species in all_species:
print species
switch_to_db(cursor, ensembl_db_name[species])
qry = "select count(1) from usearch_exon"
rows = search_db(cursor, qry)
count = int(rows[0][0])
print "\t usearch exons: ", count
total += count
qry = "select count(1) from sw_exon"
rows = search_db(cursor, qry)
count = int(rows[0][0])
print "\t sw exons: ", count
total += count
print
print 'total: ', total
cursor.close()
db.close()
def main():
db_name = "exolocator_db"
db = connect_to_mysql(user="******", passwd="tooiram")
cursor = db.cursor()
switch_to_db(cursor, db_name)
cfg = ConfigurationReader(user="******", passwd="tooiram", check=False)
inpath = cfg.get_path('afs_dumps')
indir = "%s/exon_map" % inpath
infile = "%s/exon_map.sql" % indir
if (not os.path.exists(infile)):
print "not found: ", infile
sys.exit(1)
print "reading", infile
qry = "drop table exon_map"
rows = search_db(cursor, qry)
# I could not get this to run, though it runs fine directly from the mysql shell:
#qry = "source %s" % infile
#rows = search_db(cursor, qry, verbose=True)
cursor.close()
db.close()
credentials = " -u marioot -ptooiram"
cmd = "mysql %s exolocator_db < %s" % (credentials, infile)
print cmd
ret = commands.getoutput(cmd)
print ret
return True
def main():
special = None
no_threads = 1
db = connect_to_mysql()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
total = 0
for species in all_species:
print species
switch_to_db (cursor, ensembl_db_name[species])
qry = "select count(1) from usearch_exon"
rows = search_db (cursor, qry)
count = int(rows[0][0])
print "\t usearch exons: ", count
total += count
qry = "select count(1) from sw_exon"
rows = search_db (cursor, qry)
count = int(rows[0][0])
print "\t sw exons: ", count
total += count
print
print 'total: ', total
cursor.close()
db.close()
def main ():
db_name = "exolocator_db"
db = connect_to_mysql(user="******", passwd="tooiram")
cursor = db.cursor()
switch_to_db (cursor, db_name)
cfg = ConfigurationReader (user="******", passwd="tooiram", check=False)
inpath = cfg.get_path('afs_dumps')
indir = "%s/exon_map" % inpath
infile = "%s/exon_map.sql" % indir
if (not os.path.exists(infile)):
print "not found: ", infile
sys.exit(1)
print "reading", infile
qry = "drop table exon_map"
rows = search_db(cursor, qry)
# I could not get this to run, though it runs fine directly from the mysql shell:
#qry = "source %s" % infile
#rows = search_db(cursor, qry, verbose=True)
cursor.close()
db.close()
credentials = " -u marioot -ptooiram"
cmd = "mysql %s exolocator_db < %s" % (credentials, infile)
print cmd
ret = commands.getoutput(cmd)
print ret
return True
def collect_orthologues(gene_list, db_info):
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
db_human = connect_to_mysql()
cursor_human = db_human.cursor()
switch_to_db (cursor_human, ensembl_db_name['homo_sapiens'])
ensembl_compara_name = get_compara_name(cursor)
print ensembl_compara_name
db_compara = connect_to_mysql()
cursor_compara = db_compara.cursor()
switch_to_db (cursor_compara, ensembl_compara_name)
ortho_table = {}
ortho_table ['ortholog_one2one'] = 'orthologue'
ortho_table ['apparent_ortholog_one2one'] = 'orthologue'
ortho_table ['possible_ortholog'] = 'unresolved_ortho'
ortho_table ['ortholog_one2many'] = 'unresolved_ortho'
ortho_table ['ortholog_many2many'] = 'unresolved_ortho'
ct = 0
for gene_id in gene_list:
ct += 1
# find stable
stable_id = gene2stable(cursor_human, gene_id=gene_id)
# memebr id refers to entries in compara db
member_id = stable2member(cursor_compara, stable_id)
#print gene_id, stable_id, member_id
if ( not ct%100): print ct , "out of ", len(gene_list)
# in compara table, get everything that homology has to say about
# the possible orthologues
# find all orthologous pairs suggested for this gene
for ortho_type in ['ortholog_one2one','possible_ortholog', 'apparent_ortholog_one2one',
'ortholog_one2many','ortholog_many2many']:
orthos = get_orthologues(cursor_compara, ortho_type, member_id)
if ( orthos):
store_orthologues (cursor_human, ortho_table[ortho_type], cursor, all_species,
ensembl_db_name, gene_id, orthos)
cursor.close()
db.close()
cursor_human.close()
db_human.close()
cursor_compara.close()
db_compara.close()
def map_cleanup (cursor, ensembl_db_name, human_exons):
switch_to_db(cursor,ensembl_db_name['homo_sapiens'])
for exon in human_exons:
qry = "delete from exon_map where exon_id = %d " % exon.exon_id
qry += " and exon_known = %d " % exon.is_known
qry += " and cognate_exon_known > 1 "
qry += " and similarity is NULL"
rows = search_db (cursor, qry, verbose=False)
return True
def main():
no_threads = 10
db = connect_to_mysql()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
species = 'homo_sapiens'
switch_to_db (cursor, ensembl_db_name[species])
gene_list = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
cursor.close()
db.close()
parallelize (no_threads, collect_orthologues, gene_list, [local_db, ensembl_db_name])
return True
def all_species_all_genes_loop(species_list, db_info):
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cursor = db.cursor()
#####################################
for species in species_list:
print
print "############################"
print species
sys.stdout.flush()
if not switch_to_db(cursor, ensembl_db_name[species]):
return False
if (species=='homo_sapiens'):
gene_ids = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
else:
gene_ids = get_gene_ids (cursor, biotype='protein_coding')
#for all protein coding genes in a species
#for gene_id in [10093105]:
for gene_id in gene_ids:
# for all exons in the gene
exons = gene2exon_list(cursor, gene_id)
if (not exons):
print 'no exons for gene', gene_id
continue
####################################
pep_seqs(cursor, gene_id, exons)
####################################
if not gene_ids.index(gene_id)%1000:
print "%50s: %5.1f%% " % (species, 100*(float( gene_ids.index(gene_id) +1 )/len(gene_ids)) )
sys.stdout.flush()
print species, "done"
cursor.close()
db.close()
def main():
no_threads = 12
special = None
if len(sys.argv) > 1 and len(sys.argv)<3:
print "usage: %s <set name> <no of processes>" % sys.argv[0]
exit(1) # after usage statement
elif len(sys.argv)>=3:
special = sys.argv[1]
special = special.lower()
if special == 'none': special = None
no_processes = int(sys.argv[2])
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
print "running ", sys.argv[0]
if special:
print "using", special, "set"
if special == 'complement':
gene_list = get_complement_ids(cursor, ensembl_db_name, cfg)
else:
gene_list = get_theme_ids (cursor, ensembl_db_name, cfg, special )
else:
print "using all protein coding genes"
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
gene_list = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
cursor.close()
db.close()
parallelize (no_processes, multiple_exon_alnmt, gene_list, [local_db, ensembl_db_name])
return True
def main():
no_threads = 1
special = ''
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
species = ''
if len(sys.argv) > 1 and len(sys.argv)<3 or len(sys.argv) >= 2 and sys.argv[1]=="-h":
print "usage: %s <set name/species> <number of processes>" % sys.argv[0]
exit(1) # after usage statement
elif len(sys.argv)==3:
special = sys.argv[1].lower()
if special == 'none':
special = None
elif special in all_species:
species = special
no_threads = int(sys.argv[2])
print '======================================='
print sys.argv[0]
if species:
print species, "only"
switch_to_db (cursor, ensembl_db_name[species])
if (species=='homo_sapiens'):
gene_ids = get_gene_ids (cursor, biotype='protein_coding', is_known=1, ref_only=True)
else:
gene_ids = get_gene_ids (cursor, biotype='protein_coding')
parallelize_args = [no_threads, one_species_all_genes_loop, gene_ids, [local_db, ensembl_db_name, species]]
elif special:
print "using", special, "set"
gene_list = get_theme_ids (cursor, ensembl_db_name, cfg, special )
parallelize_args = [no_threads, ortologues_for_given_genes_loop, gene_list, [local_db, ensembl_db_name]]
else:
parallelize_args = [no_threads, all_species_all_genes_loop, all_species, [local_db, ensembl_db_name]]
cursor.close()
db .close()
parallelize (*parallelize_args)
def ortologues_for_given_genes_loop (gene_list, db_info):
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cursor = db.cursor()
#####################################
for gene_id in gene_list:
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
orthologues = get_orthos (cursor, gene_id, 'orthologue') # get_orthos changes the db pointer
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
orthologues += get_orthos (cursor, gene_id, 'unresolved_ortho')
for [ortho_gene_id, ortho_species] in [[gene_id,'homo_sapiens']] + orthologues:
print ">>> ", ortho_species, ortho_gene_id
switch_to_db (cursor, ensembl_db_name[ortho_species])
# for all exons in the gene
exons = gene2exon_list(cursor, ortho_gene_id)
if (not exons):
if verbose: print 'no exons for gene', ortho_gene_id
continue
##############################
pep_seqs(cursor, ortho_gene_id, exons)
####################################
if not gene_list.index(gene_id)%1000:
print "%5.1f%% " % (100*(float( gene_list.index(gene_id) +1 )/len(gene_list)) )
sys.stdout.flush()
cursor.close()
db.close()
def collect_paralogues(species_list, db_info):
[local_db, ensembl_db_name] = db_info
db_species = connect_to_mysql()
cursor_species = db_species.cursor()
ensembl_compara_name = get_compara_name(cursor_species)
print ensembl_compara_name
db_compara = connect_to_mysql()
cursor_compara = db_compara.cursor()
switch_to_db (cursor_compara, ensembl_compara_name)
for species in species_list:
switch_to_db (cursor_species, ensembl_db_name[species])
# it looks I cannot demand that the gene is known, because for many species
# most of the genes still have 'predicted' status
gene_list = get_gene_ids (cursor_species, biotype='protein_coding')
ct = 0
for gene_id in gene_list:
ct += 1
# find stable
stable_id = gene2stable(cursor_species, gene_id=gene_id)
# memebr id refers to entries in compara db
member_id = stable2member(cursor_compara, stable_id)
#print gene_id, stable_id, member_id
if (not ct%100):
print species, ct , "out of ", len(gene_list)
# find all paralogue pairs suggested for this gene
ortho_type = 'within_species_paralog'
paralogues = get_orthologues(cursor_compara, ortho_type, member_id)
if not paralogues: continue
store_paralogues (cursor_species, gene_id, paralogues)
print species, 'done'
cursor_species.close()
db_species.close()
cursor_compara.close()
db_compara.close()
def main():
db_name = "exolocator_db"
db = connect_to_mysql(user="******", passwd="tooiram")
cursor = db.cursor()
switch_to_db (cursor, db_name)
cfg = ConfigurationReader (user="******", passwd="tooiram", check=False)
in_path = cfg.get_path('afs_dumps')
in_path += "/para_dump"
if (not os.path.exists(in_path)):
print in_path, "not found"
sys.exit(1) # exit on non-existent outdir
###############
if 1:
qry = "drop table paralog"
search_db (cursor, qry)
qry = "create table paralog (id int(10) primary key auto_increment) "
search_db (cursor, qry)
qry = "alter table paralog ADD gene_id1 varchar(30) "
search_db (cursor, qry)
qry = "alter table paralog ADD gene_id2 varchar(30) "
search_db (cursor, qry)
create_index (cursor, db_name,'gene_id_index', 'paralog', ['gene_id1', 'gene_id2'])
###############
os.chdir(in_path)
filenames = glob.glob("*_para_dump.txt")
###############
for infile in filenames:
print infile
store(cursor, infile)
cursor.close()
db .close()
def main():
db_name = "exolocator_db"
db = connect_to_mysql(user="******", passwd="tooiram")
cursor = db.cursor()
switch_to_db(cursor, db_name)
cfg = ConfigurationReader(user="******", passwd="tooiram", check=False)
in_path = cfg.get_path('afs_dumps')
in_path += "/para_dump"
if (not os.path.exists(in_path)):
print in_path, "not found"
sys.exit(1) # exit on non-existent outdir
###############
if 1:
qry = "drop table paralog"
search_db(cursor, qry)
qry = "create table paralog (id int(10) primary key auto_increment) "
search_db(cursor, qry)
qry = "alter table paralog ADD gene_id1 varchar(30) "
search_db(cursor, qry)
qry = "alter table paralog ADD gene_id2 varchar(30) "
search_db(cursor, qry)
create_index(cursor, db_name, 'gene_id_index', 'paralog',
['gene_id1', 'gene_id2'])
###############
os.chdir(in_path)
filenames = glob.glob("*_para_dump.txt")
###############
for infile in filenames:
print infile
store(cursor, infile)
cursor.close()
db.close()
def one_species_all_genes_loop(gene_ids, db_info):
[local_db, ensembl_db_name, species] = db_info
db = connect_to_mysql()
cursor = db.cursor()
switch_to_db (cursor, ensembl_db_name[species])
#for gene_id in [10092907]:
for gene_id in gene_ids:
# for all exons in the gene
exons = gene2exon_list(cursor, gene_id)
if (not exons):
if verbose: print 'no exons for gene', gene_id
continue
####################################
pep_seqs(cursor, gene_id, exons)
####################################
if not gene_ids.index(gene_id) % 100:
print "\t done with %d out of %d (%5.1f%%) " % (gene_ids.index(gene_id) + 1, len(gene_ids),
100 * (float(gene_ids.index(gene_id) + 1) / len(gene_ids)))
sys.stdout.flush()
cursor.close()
db.close()
def dump_orthos (species_list, db_info):
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
# find db ids adn common names for each species db
[all_species, ensembl_db_name] = get_species (cursor)
# in the afa headers use 'trivial' names for the species: cow, dog, pig, ...
trivial_name = translate_to_trivial(cursor, all_species)
out_path = cfg.get_path('afs_dumps')
outfile = "{0}/orthologue_dump.txt".format(out_path)
print outfile
of = erropen (outfile,"w")
species = 'homo_sapiens'
switch_to_db (cursor, ensembl_db_name[species])
qry = "select * from orthologue"
rows = search_db (cursor, qry)
for row in rows:
[pair_id, human_gene_id, cognate_gene_id, genome_db_id, source] = row
species = genome_db_id2species (cursor, genome_db_id)
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
human_stable_id = gene2stable(cursor, human_gene_id)
switch_to_db (cursor, ensembl_db_name[species])
cognate_stable_id = gene2stable(cursor, cognate_gene_id)
print >>of, orthos_tabstring ([human_stable_id, cognate_stable_id, species, trivial_name[species]])
of.close()
cursor.close()
db .close()
def multiple_exon_alnmt(species_list, db_info):
[local_db, ensembl_db_name] = db_info
verbose = False
db = connect_to_mysql()
cfg = ConfigurationReader()
acg = AlignmentCommandGenerator()
cursor = db.cursor()
for species in species_list:
print
print "############################"
print species
switch_to_db (cursor, ensembl_db_name[species])
gene_ids = get_gene_ids (cursor, biotype='protein_coding')
#gene_ids = get_theme_ids(cursor, cfg, 'wnt_pathway')
if not gene_ids:
print "no gene_ids"
continue
gene_ct = 0
tot = 0
ok = 0
no_maps = 0
no_pepseq = 0
no_paralogues = 0
for gene_id in gene_ids:
if verbose: start = time()
gene_ct += 1
if not gene_ct%100: print species, gene_ct, "genes out of", len(gene_ids)
if verbose:
print
print gene_id, gene2stable(cursor, gene_id), get_description (cursor, gene_id)
# get the paralogues - only the representative for the family will have this
paralogues = get_paras (cursor, gene_id)
if not paralogues:
if verbose: print "\t not a template or no paralogues"
continue
if verbose: print "paralogues: ", paralogues
# get _all_ exons
template_exons = gene2exon_list(cursor, gene_id)
if (not template_exons):
if verbose: print 'no exons for ', gene_id
continue
# find all template exons we are tracking in the database
for template_exon in template_exons:
if verbose: print template_exon.exon_id
maps = get_maps(cursor, ensembl_db_name, template_exon.exon_id,
template_exon.is_known, species=species, table='para_exon_map')
if not maps:
no_maps += 1
continue
# output to fasta:
seqname = "{0}:{1}:{2}".format('template', template_exon.exon_id, template_exon.is_known)
exon_seqs_info = get_exon_seqs (cursor, template_exon.exon_id, template_exon.is_known)
if not exon_seqs_info: continue
[exon_seq_id, pepseq, pepseq_transl_start, pepseq_transl_end,
left_flank, right_flank, dna_seq] = exon_seqs_info
if (not pepseq):
if ( template_exon.is_coding and template_exon.covering_exon <0): # this should be a master exon
print "no pep seq for", template_exon.exon_id, "coding ", template_exon.is_coding,
print "canonical: ", template_exon.is_canonical
print "length of dna ", len(dna_seq)
no_pepseq += 1
continue
tot += 1
sequences = {seqname:pepseq}
headers = [seqname]
for map in maps:
exon = map2exon(cursor, ensembl_db_name, map, paralogue=True)
pepseq = get_exon_pepseq (cursor,exon)
if (not pepseq):
continue
seqname = "{0}:{1}:{2}".format('para', map.exon_id_2, map.exon_known_2)
headers.append(seqname)
sequences[seqname] = pepseq
fasta_fnm = "{0}/{1}_{2}_{3}.fa".format( cfg.dir_path['scratch'], species, template_exon.exon_id, template_exon.is_known)
output_fasta (fasta_fnm, headers, sequences)
if (len(headers) <=1 ):
print "single species in the alignment (?)"
no_paralogues += 1
continue
# align
afa_fnm = "{0}/{1}_{2}_{3}.afa".format( cfg.dir_path['scratch'], species, template_exon.exon_id, template_exon.is_known)
mafftcmd = acg.generate_mafft_command (fasta_fnm, afa_fnm)
ret = commands.getoutput(mafftcmd)
# read in the alignment
inf = erropen(afa_fnm, "r")
if not inf:
print gene_id
continue
template_seq_seen = False
for record in SeqIO.parse(inf, "fasta"):
### store the alignment as bitstring
# Generate the bitmap
bs = Bits(bin='0b' + re.sub("[^0]","1", str(record.seq).replace('-','0')))
msa_bitmap = bs.tobytes()
# Retrieve information on the cognate
label, cognate_exon_id, cognate_exon_known = record.id.split(':')
if (label == 'template'):
template_seq_seen = True
# Write the bitmap to the database
#print "updating: ", template_exon.exon_id
store_or_update(cursor, "para_exon_map", {"cognate_exon_id" :cognate_exon_id,
"cognate_exon_known" :cognate_exon_known,
"exon_id" :template_exon.exon_id,
"exon_known" :template_exon.is_known},
{"msa_bitstring":MySQLdb.escape_string(msa_bitmap)})
inf.close()
ok += 1
commands.getoutput("rm "+afa_fnm+" "+fasta_fnm)
if verbose: print " time: %8.3f\n" % (time()-start);
outstr = species + " done \n"
outstr += "tot: %d ok: %d \n" % (tot, ok)
outstr += "no maps %d \n" % no_pepseq
outstr += "no pepseq %d \n" % no_pepseq
outstr += "no paralogues %d \n" % no_paralogues
outstr += "\n"
print outstr
def main():
db = connect_to_mysql()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
for species in all_species:
if not species=='homo_sapiens': continue
print
print species
switch_to_db (cursor, ensembl_db_name[species])
if (species=='homo_sapiens'):
gene_ids = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
else:
gene_ids = get_gene_ids (cursor, biotype='protein_coding')
tot_exons = 0
no_exon_seq = 0
short_dna = 0
pepseq_ok = 0
mismatch = 0
stored_incorrect = 0
translation_fail = 0
#####################################
#for gene_id in [10092907]:
for gene_id in gene_ids:
#for tot in range(1000):
#gene_id = choice(gene_ids)
# get _all_ exons
exons = gene2exon_list(cursor, gene_id)
if (not exons):
print 'no exons for gene', gene_id
sys.exit(1)
for exon in exons:
#####################################
if not exon.is_coding:
print exon.exon_id, " not coding "
continue
if exon.covering_exon >0:
print exon.exon_id, " is covered by ", exon.covering_exon
continue
tot_exons += 1
# exons seqs are its aa translation, left_flank, right_flank, and dna_seq
exon_seqs = get_exon_seqs(cursor, exon.exon_id, exon.is_known)
if (not exon_seqs):
no_exon_seq += 1
print "no exon seqs for ", gene_id, exon.exon_id
#exit(1)
continue
[exon_seq_id, pepseq, pepseq_transl_start,
pepseq_transl_end, left_flank, right_flank, dna_seq] = exon_seqs
if len(dna_seq)<3:
short_dna += 1
print "short_dna:", dna_seq
continue
if (pepseq_transl_start == -10): # ??? what is this shit? adn what happens downstream if the pepseq_transl_start is None?
translation_fail += 1
print "pepseq_transl_start:", pepseq_transl_start
continue
mitochondrial = is_mitochondrial(cursor, gene_id)
dnaseq = Seq (dna_seq[pepseq_transl_start:pepseq_transl_end], generic_dna)
if (mitochondrial):
pepseq2 = dnaseq.translate(table="Vertebrate Mitochondrial").tostring()
else:
pepseq2 = dnaseq.translate().tostring()
if True:
print exon.exon_id
print "pep stored:", pepseq
print "dna transl:", pepseq2
print "dna begin:", dna_seq[:12]
print "start:" , pepseq_transl_start,
print "end:", pepseq_transl_end
print
if (not pepseq == pepseq2):
stored_incorrect += 1
else:
pepseq_ok += 1
print "total coding exons ", tot_exons
print "no exon seq info ", no_exon_seq
print "short dna ", short_dna
print "transl failure ", translation_fail
print "stored pepseq does not correspond to the translation of stored dna: ", stored_incorrect
print "pepseq ok ", pepseq_ok
cursor.close()
db .close()
def main():
db = connect_to_mysql()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species(cursor)
for species in all_species:
if not species == 'homo_sapiens': continue
print
print species
switch_to_db(cursor, ensembl_db_name[species])
if (species == 'homo_sapiens'):
gene_ids = get_gene_ids(cursor,
biotype='protein_coding',
is_known=1)
else:
gene_ids = get_gene_ids(cursor, biotype='protein_coding')
tot_exons = 0
no_exon_seq = 0
short_dna = 0
pepseq_ok = 0
mismatch = 0
stored_incorrect = 0
translation_fail = 0
#####################################
#for gene_id in [10092907]:
for gene_id in gene_ids:
#for tot in range(1000):
#gene_id = choice(gene_ids)
# get _all_ exons
exons = gene2exon_list(cursor, gene_id)
if (not exons):
print 'no exons for gene', gene_id
sys.exit(1)
for exon in exons:
#####################################
if not exon.is_coding:
print exon.exon_id, " not coding "
continue
if exon.covering_exon > 0:
print exon.exon_id, " is covered by ", exon.covering_exon
continue
tot_exons += 1
# exons seqs are its aa translation, left_flank, right_flank, and dna_seq
exon_seqs = get_exon_seqs(cursor, exon.exon_id, exon.is_known)
if (not exon_seqs):
no_exon_seq += 1
print "no exon seqs for ", gene_id, exon.exon_id
#exit(1)
continue
[
exon_seq_id, pepseq, pepseq_transl_start,
pepseq_transl_end, left_flank, right_flank, dna_seq
] = exon_seqs
if len(dna_seq) < 3:
short_dna += 1
print "short_dna:", dna_seq
continue
if (
pepseq_transl_start == -10
): # ??? what is this shit? adn what happens downstream if the pepseq_transl_start is None?
translation_fail += 1
print "pepseq_transl_start:", pepseq_transl_start
continue
mitochondrial = is_mitochondrial(cursor, gene_id)
dnaseq = Seq(dna_seq[pepseq_transl_start:pepseq_transl_end],
generic_dna)
if (mitochondrial):
pepseq2 = dnaseq.translate(
table="Vertebrate Mitochondrial").tostring()
else:
pepseq2 = dnaseq.translate().tostring()
if True:
print exon.exon_id
print "pep stored:", pepseq
print "dna transl:", pepseq2
print "dna begin:", dna_seq[:12]
print "start:", pepseq_transl_start,
print "end:", pepseq_transl_end
print
if (not pepseq == pepseq2):
stored_incorrect += 1
else:
pepseq_ok += 1
print "total coding exons ", tot_exons
print "no exon seq info ", no_exon_seq
print "short dna ", short_dna
print "transl failure ", translation_fail
print "stored pepseq does not correspond to the translation of stored dna: ", stored_incorrect
print "pepseq ok ", pepseq_ok
cursor.close()
db.close()
def main():
no_threads = 1
special = None
if len(sys.argv) > 1 and len(sys.argv)<3:
print "usage: %s <set name> <number of threads> " % sys.argv[0]
exit(1)
elif len(sys.argv)==3:
special = sys.argv[1]
special = special.lower()
if special == 'none': special = None
no_threads = int(sys.argv[2])
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
# find db ids adn common names for each species db
[all_species, ensembl_db_name] = get_species (cursor)
species = 'homo_sapiens'
switch_to_db (cursor, ensembl_db_name[species])
if special:
print "using", special, "set"
gene_list = get_theme_ids (cursor, ensembl_db_name, cfg, special )
else:
print "using all protein coding genes"
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
gene_list = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
incomplete = 0
genes_checked = 0
#for gene_id in gene_list:
#for gene_id in [743609]:
for sampling_count in range(1000):
gene_id = choice(gene_list)
genes_checked += 1
with_map = 0
tot = 0
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
print gene2stable(cursor, gene_id), get_description (cursor, gene_id)
# find all exons we are tracking in the database
human_exons = gene2exon_list(cursor, gene_id)
human_exons.sort(key=lambda exon: exon.start_in_gene)
has_a_map = False
for human_exon in human_exons:
if (not human_exon.is_canonical or not human_exon.is_coding): continue
if verbose:
print
print "\t human", human_exon.exon_id, human_exon.is_known
print "\t ", get_exon_pepseq(cursor, human_exon, ensembl_db_name['homo_sapiens'])
print "\t checking maps ..."
maps = get_maps(cursor, ensembl_db_name, human_exon.exon_id, human_exon.is_known)
tot += 1
if maps:
has_a_map = True
with_map += 1
#print "ok"
else:
print"no maps for exon", human_exon.exon_id
continue
if verbose:
for map in maps:
species = map.species_2
exon = map2exon(cursor, ensembl_db_name, map)
unaligned_sequence = get_exon_pepseq(cursor, exon, ensembl_db_name[species])
if ( map.similarity):
print "\t", species, map.source, map.exon_id_2, map.exon_known_2
print "\tmaps to ", map.exon_id_1, map.exon_known_1
print "\tsim", map.similarity,
print "\tsource", map.source
print "\t", unaligned_sequence
if not map.bitmap:
print "\t bitmap not assigned"
else:
bs = Bits(bytes=map.bitmap)
reconst_pepseq = ''
if (not bs.count(1) == len(unaligned_sequence)):
print "\talnd seq mismatch"
else:
usi = iter(unaligned_sequence)
for c in bs.bin:
if c == '0': reconst_pepseq += '-'
else: reconst_pepseq += next(usi)
print "\tbinary : ", bs.bin
print "\talnd seq: ", reconst_pepseq
print
if not tot== with_map:
print "#### gene id: %d total exons: %d with map: %d ( = %d%%) " % \
(gene_id, tot, with_map, int(float(with_map)/tot*100) )
incomplete += 1
print "genes checked: %d, incomplete: %d" % (genes_checked, incomplete)
cursor.close()
db.close()
print tot, with_map
def main():
special = None
no_threads = 1
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
[all_species, ensembl_db_name] = get_species (cursor)
if special:
print "using", special, "set"
gene_list = get_theme_ids (cursor, ensembl_db_name, cfg, special)
else:
print "using all protein coding genes"
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
gene_list = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
# loop over all genes
sw_count = 0
tot_count = 0
for human_gene_id in gene_list:
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
human_stable = gene2stable (cursor, human_gene_id)
human_description = get_description(cursor, human_gene_id)
tot_count += 1
#print human_gene_id, human_stable, human_description
human_exons = [e for e in gene2exon_list(cursor, human_gene_id, verbose=True)
if e.covering_exon < 0 and e.is_canonical and e.is_known]
if not human_exons:
#print "\t\t", human_stable, "no exons found"
continue
human_exons.sort(key=lambda exon: exon.start_in_gene)
# loop over all exons in this gene
maps_for_exon = {}
for he in human_exons:
he.stable_id = exon2stable (cursor, he.exon_id, ensembl_db_name['homo_sapiens'])
he.pepseq = get_exon_pepseq (cursor, he, ensembl_db_name['homo_sapiens'])
# maps cleanup: get rid of maps that have "none" as similarity
maps_for_exon[he] = get_maps(cursor, ensembl_db_name, he.exon_id, he.is_known) # exon data
if not maps_for_exon[he]: continue
#maps_for_exon[he] = filter (lambda m: m.source == 'sw_sharp' or m.source == 'usearch',
# maps_for_exon[he])
maps_for_exon[he] = filter (lambda m: m.source == 'usearch',
maps_for_exon[he])
if not maps_for_exon[he]:
#print "\t\t", human_stable, "no maps found"
continue
sw_count += len(maps_for_exon[he])
#break
print "tot count: ", tot_count
print "sw count: ", sw_count
#print "tot count: ", tot_count
#print "sw count: ", sw_count
cursor.close()
db.close()
def dump_exons(species_list, db_info):
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
out_path = "{0}/exons".format(cfg.get_path('afs_dumps'))
if not os.path.exists(out_path):
print out_path, "not found"
exit(1) # exit on failed output dir check
for species in species_list:
#if (not species=='homo_sapiens'):
# continue
outfile = "{0}/{1}_exon_dump.txt".format(out_path, species)
of = erropen(outfile, "w")
if not of: continue
switch_to_db(cursor, ensembl_db_name[species])
if (species == 'homo_sapiens'):
gene_ids = get_gene_ids(cursor,
biotype='protein_coding',
is_known=1,
ref_only=True)
else:
gene_ids = get_gene_ids(cursor, biotype='protein_coding')
source = get_analysis_dict(cursor)
ct = 0
for gene_id in gene_ids:
ct += 1
if (not ct % 1000):
print species, ct, len(gene_ids)
# get _all_ exons
exons = gene2exon_list(cursor, gene_id)
if (not exons):
print 'no exons for ', gene_id
continue
for exon in exons:
if exon.covering_exon > 0: continue
# exons seqs are its aa translation, left_flank, right_flank, and dna_seq
exon_seqs = get_exon_seqs(cursor, exon.exon_id, exon.is_known)
if (not exon_seqs):
continue
# human readable string describing the source of annotation for this exon
if exon.is_known == 2:
analysis = 'sw_sharp'
elif exon.is_known == 3:
analysis = 'usearch'
else:
analysis = source[exon.analysis_id]
# the first field return by get_exon_seqs is the exon_seq_id, so get rid of it
gene_stable_id = gene2stable(cursor, gene_id)
if (exon.is_known == 1):
exon_stable_id = exon2stable(cursor, exon.exon_id)
elif (exon.is_known == 2):
exon_stable_id = 'sw_sharp_' + str(exon.exon_id)
elif (exon.is_known == 3):
exon_stable_id = 'usearch_' + str(exon.exon_id)
else:
exon_stable_id = "anon"
print >> of, exon_tabstring(exon, gene_stable_id,
exon_stable_id, species, analysis,
exon_seqs[1:])
of.close()
print species, "done"
cursor.close()
db.close()
def multiple_exon_alnmt(gene_list, db_info):
print "process pid: %d, length of gene list: %d" % ( get_process_id(), len(gene_list))
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cfg = ConfigurationReader()
acg = AlignmentCommandGenerator()
cursor = db.cursor()
# find db ids adn common names for each species db
[all_species, ensembl_db_name] = get_species (cursor)
species = 'homo_sapiens'
switch_to_db (cursor, ensembl_db_name[species])
gene_ids = get_gene_ids (cursor, biotype='protein_coding', is_known=1)
# for each human gene
gene_ct = 0
tot = 0
ok = 0
no_maps = 0
no_pepseq = 0
no_orthologues = 0
min_similarity = cfg.get_value('min_accptbl_exon_sim')
#gene_list.reverse()
for gene_id in gene_list:
start = time()
gene_ct += 1
if not gene_ct%10: print gene_ct, "genes out of", len(gene_list)
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
print gene_ct, len(gene_ids), gene_id, gene2stable(cursor, gene_id), get_description (cursor, gene_id)
human_exons = filter (lambda e: e.is_known==1 and e.is_coding and e.covering_exon<0, gene2exon_list(cursor, gene_id))
human_exons.sort(key=lambda exon: exon.start_in_gene)
##################################################################
for human_exon in human_exons:
tot += 1
# find all orthologous exons the human exon maps to
maps = get_maps(cursor, ensembl_db_name, human_exon.exon_id, human_exon.is_known)
if verbose:
print "\texon no.", tot, " id", human_exon.exon_id,
if not maps:
print " no maps"
print human_exon
print
if not maps:
no_maps += 1
continue
# human sequence to fasta:
seqname = "{0}:{1}:{2}".format('homo_sapiens', human_exon.exon_id, human_exon.is_known)
switch_to_db (cursor, ensembl_db_name['homo_sapiens'])
[exon_seq_id, pepseq, pepseq_transl_start, pepseq_transl_end,
left_flank, right_flank, dna_seq] = get_exon_seqs (cursor, human_exon.exon_id, human_exon.is_known)
if (not pepseq):
if verbose and human_exon.is_coding and human_exon.covering_exon <0: # this should be a master exon
print "no pep seq for", human_exon.exon_id, "coding ", human_exon.is_coding,
print "canonical: ", human_exon.is_canonical
print "length of dna ", len(dna_seq)
no_pepseq += 1
continue
# collect seq from all maps, and output them in fasta format
hassw = False
headers = []
sequences = {}
exons_per_species = {}
for map in maps:
switch_to_db (cursor, ensembl_db_name[map.species_2])
if map.similarity < min_similarity: continue
exon = map2exon(cursor, ensembl_db_name, map)
pepseq = get_exon_pepseq (cursor,exon)
if (not pepseq):
continue
if map.source == 'sw_sharp':
exon_known_code = 2
hassw = True
elif map.source == 'usearch':
exon_known_code = 3
hassw = True
else:
exon_known_code = map.exon_known_2
seqname = "{0}:{1}:{2}".format(map.species_2, map.exon_id_2, exon_known_code)
headers.append(seqname)
sequences[seqname] = pepseq
# for split exon concatenation (see below)
if not map.species_2 in exons_per_species.keys():
exons_per_species[map.species_2] = []
exons_per_species[map.species_2].append ([ map.exon_id_2, exon_known_code]);
if (len(headers) <=1 ):
if verbose: print "single species in the alignment"
no_orthologues += 1
continue
# concatenate exons from the same gene - the alignment program might go wrong otherwise
concatenated = concatenate_exons (cursor, ensembl_db_name, sequences, exons_per_species)
fasta_fnm = "{0}/{1}.fa".format( cfg.dir_path['scratch'], human_exon.exon_id)
output_fasta (fasta_fnm, sequences.keys(), sequences)
# align
afa_fnm = "{0}/{1}.afa".format( cfg.dir_path['scratch'], human_exon.exon_id)
mafftcmd = acg.generate_mafft_command (fasta_fnm, afa_fnm)
ret = commands.getoutput(mafftcmd)
if (verbose): print 'almt to', afa_fnm
# read in the alignment
inf = erropen(afa_fnm, "r")
aligned_seqs = {}
for record in SeqIO.parse(inf, "fasta"):
aligned_seqs[record.id] = str(record.seq)
inf.close()
# split back the concatenated exons
if concatenated: split_concatenated_exons (aligned_seqs, concatenated)
human_seq_seen = False
for seq_name, sequence in aligned_seqs.iteritems():
# if this is one of the concatenated seqs, split them back to two
### store the alignment as bitstring
# Generate the bitmap
bs = Bits(bin='0b' + re.sub("[^0]","1", sequence.replace('-','0')))
# The returned value of tobytes() will be padded at the end
# with between zero and seven 0 bits to make it byte aligned.
# I will end up with something that looks like extra alignment gaps, that I'll have to return
msa_bitmap = bs.tobytes()
# Retrieve information on the cognate
cognate_species, cognate_exon_id, cognate_exon_known = seq_name.split(':')
if cognate_exon_known == '2':
source = 'sw_sharp'
elif cognate_exon_known == '3':
source = 'usearch'
else:
source = 'ensembl'
if (cognate_species == 'homo_sapiens'):
human_seq_seen = True
cognate_genome_db_id = species2genome_db_id(cursor, cognate_species) # moves the cursor
switch_to_db(cursor, ensembl_db_name['homo_sapiens']) # so move it back to h**o sapiens
# Write the bitmap to the database
#if (cognate_species == 'homo_sapiens'):
if verbose: # and (source=='sw_sharp' or source=='usearch'):
print "storing"
print human_exon.exon_id, human_exon.is_known
print cognate_species, cognate_genome_db_id, cognate_exon_id, cognate_exon_known, source
print sequence
if not msa_bitmap:
print "no msa_bitmap"
continue
store_or_update(cursor, "exon_map", {"cognate_genome_db_id":cognate_genome_db_id,
"cognate_exon_id":cognate_exon_id ,"cognate_exon_known" :cognate_exon_known,
"source": source, "exon_id" :human_exon.exon_id, "exon_known":human_exon.is_known},
{"msa_bitstring":MySQLdb.escape_string(msa_bitmap)})
ok += 1
commands.getoutput("rm "+afa_fnm+" "+fasta_fnm)
if verbose: print " time: %8.3f\n" % (time()-start);
print "tot: ", tot, "ok: ", ok
print "no maps ", no_pepseq
print "no pepseq ", no_pepseq
print "no orthologues ", no_orthologues
print
def concatenate_exons (cursor, ensembl_db_name, sequences, exons_per_species):
concatenated = {}
# are there multiple candidates from the same s pecies?
for species, exon_labels in exons_per_species.iteritems():
if len(exon_labels) < 2: continue
# if there are multiple candidates from the same species - are they from the same gene?
exons_per_gene = {}
for [exon_id, exon_known_code] in exon_labels:
gene_id = exon_id2gene_id (cursor, ensembl_db_name[species], exon_id, exon_known_code)
if not gene_id in exons_per_gene.keys():
exons_per_gene[gene_id] = []
exons_per_gene[gene_id].append ([exon_id, exon_known_code])
# if yes - do they overlap in the gene? ... I so need to do this whole crap differently
# the whole idea was no to be doing this here, but the alignment progs (mafft) can screw up here
# big time
switch_to_db(cursor, ensembl_db_name[species])
for gene_id, exons_from_gene in exons_per_gene.iteritems():
if ( len(exons_from_gene) < 2): continue
# how robust should I be here? how many fragments should I worry about?
# how about some combinatorial pearls, like 3 exons non overlapping, but 4 overlapping 1 or more ...?
# and then exons could be overlapping when the translation regions are not ...
# for now, I'll only offer this patch for the cases when the pieces are non-overlapping
exons = []
for [exon_id, exon_known_code] in exons_from_gene:
exon = get_exon (cursor, exon_id, exon_known_code)
if not exon: continue
exons.append(exon)
# sort by translation start
exons.sort(key=lambda exon: exon.start_in_gene)
# is transl_Start < transl_end of the previous exon
overlap = False
exon_prev = exons[0]
for exon in exons[1:]:
if exon_prev.pepseq_transl_end is None:
end_prev = exon_prev.end_in_gene
else:
end_prev = exon_prev.start_in_gene + exon_prev.pepseq_transl_end
if exon.pepseq_transl_start is None:
start_this = exon.start_in_gene
else:
start_this = exon.start_in_gene + exon.pepseq_transl_start
#print species, "prev end:", end_prev, "this start:", start_this
if end_prev > start_this:
# yes => overlap
overlap = True
break
exon_prev = exon
# if they overlap, do nothing - ther are already both in the fasta set
if overlap: continue
# if they do not not overlap, concatenate them, and mark them as concatenated
new_name = species + "_concat_" + str (len(concatenated) )
concatenated[new_name] = []
concat_seq = ""
for exon in exons: # note 'exons' are sorted, 'exons_from_gene' are not
old_name = "{0}:{1}:{2}".format(species, exon.exon_id, exon.is_known)
if not sequences.has_key(old_name):
#print "no key ", old_name, "in the original set (?) "
continue
concatenated[new_name].append(old_name)
if concat_seq: concat_seq += "Z"
concat_seq += sequences[old_name]
# also remove the original seqs from the alignment
del sequences[old_name]
sequences[new_name] = concat_seq
return concatenated
def main():
no_threads = 1
special = None
if len(sys.argv) > 1 and len(sys.argv) < 3:
print "usage: %s <set name> <number of threads> " % sys.argv[0]
exit(1)
elif len(sys.argv) == 3:
special = sys.argv[1]
special = special.lower()
if special == 'none': special = None
no_threads = int(sys.argv[2])
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
# find db ids adn common names for each species db
[all_species, ensembl_db_name] = get_species(cursor)
species = 'homo_sapiens'
switch_to_db(cursor, ensembl_db_name[species])
if special:
print "using", special, "set"
gene_list = get_theme_ids(cursor, ensembl_db_name, cfg, special)
else:
print "using all protein coding genes"
switch_to_db(cursor, ensembl_db_name['homo_sapiens'])
gene_list = get_gene_ids(cursor, biotype='protein_coding', is_known=1)
incomplete = 0
genes_checked = 0
#for gene_id in gene_list:
#for gene_id in [743609]:
for sampling_count in range(1000):
gene_id = choice(gene_list)
genes_checked += 1
with_map = 0
tot = 0
switch_to_db(cursor, ensembl_db_name['homo_sapiens'])
print gene2stable(cursor, gene_id), get_description(cursor, gene_id)
# find all exons we are tracking in the database
human_exons = gene2exon_list(cursor, gene_id)
human_exons.sort(key=lambda exon: exon.start_in_gene)
has_a_map = False
for human_exon in human_exons:
if (not human_exon.is_canonical or not human_exon.is_coding):
continue
if verbose:
print
print "\t human", human_exon.exon_id, human_exon.is_known
print "\t ", get_exon_pepseq(cursor, human_exon,
ensembl_db_name['homo_sapiens'])
print "\t checking maps ..."
maps = get_maps(cursor, ensembl_db_name, human_exon.exon_id,
human_exon.is_known)
tot += 1
if maps:
has_a_map = True
with_map += 1
#print "ok"
else:
print "no maps for exon", human_exon.exon_id
continue
if verbose:
for map in maps:
species = map.species_2
exon = map2exon(cursor, ensembl_db_name, map)
unaligned_sequence = get_exon_pepseq(
cursor, exon, ensembl_db_name[species])
if (map.similarity):
print "\t", species, map.source, map.exon_id_2, map.exon_known_2
print "\tmaps to ", map.exon_id_1, map.exon_known_1
print "\tsim", map.similarity,
print "\tsource", map.source
print "\t", unaligned_sequence
if not map.bitmap:
print "\t bitmap not assigned"
else:
bs = Bits(bytes=map.bitmap)
reconst_pepseq = ''
if (not bs.count(1) == len(unaligned_sequence)):
print "\talnd seq mismatch"
else:
usi = iter(unaligned_sequence)
for c in bs.bin:
if c == '0': reconst_pepseq += '-'
else: reconst_pepseq += next(usi)
print "\tbinary : ", bs.bin
print "\talnd seq: ", reconst_pepseq
print
if not tot == with_map:
print "#### gene id: %d total exons: %d with map: %d ( = %d%%) " % \
(gene_id, tot, with_map, int(float(with_map)/tot*100) )
incomplete += 1
print "genes checked: %d, incomplete: %d" % (genes_checked, incomplete)
cursor.close()
db.close()
print tot, with_map
def dump_exons (species_list, db_info):
[local_db, ensembl_db_name] = db_info
db = connect_to_mysql()
cfg = ConfigurationReader()
cursor = db.cursor()
out_path = "{0}/exons".format(cfg.get_path('afs_dumps'))
if not os.path.exists(out_path):
print out_path, "not found"
exit (1) # exit on failed output dir check
for species in species_list:
#if (not species=='homo_sapiens'):
# continue
outfile = "{0}/{1}_exon_dump.txt".format(out_path, species)
of = erropen (outfile,"w")
if not of: continue
switch_to_db (cursor, ensembl_db_name[species])
if (species=='homo_sapiens'):
gene_ids = get_gene_ids (cursor, biotype='protein_coding', is_known=1, ref_only=True)
else:
gene_ids = get_gene_ids (cursor, biotype='protein_coding')
source = get_analysis_dict(cursor)
ct = 0
for gene_id in gene_ids:
ct += 1
if (not ct%1000):
print species, ct, len(gene_ids)
# get _all_ exons
exons = gene2exon_list(cursor, gene_id)
if (not exons):
print 'no exons for ', gene_id
continue
for exon in exons:
if exon.covering_exon > 0: continue
# exons seqs are its aa translation, left_flank, right_flank, and dna_seq
exon_seqs = get_exon_seqs(cursor, exon.exon_id, exon.is_known)
if (not exon_seqs):
continue
# human readable string describing the source of annotation for this exon
if exon.is_known==2:
analysis = 'sw_sharp'
elif exon.is_known==3:
analysis = 'usearch'
else:
analysis = source[exon.analysis_id]
# the first field return by get_exon_seqs is the exon_seq_id, so get rid of it
gene_stable_id = gene2stable(cursor,gene_id)
if ( exon.is_known == 1):
exon_stable_id = exon2stable(cursor,exon.exon_id)
elif ( exon.is_known == 2):
exon_stable_id = 'sw_sharp_'+str(exon.exon_id)
elif ( exon.is_known == 3):
exon_stable_id = 'usearch_'+str(exon.exon_id)
else:
exon_stable_id = "anon"
print >> of, exon_tabstring (exon, gene_stable_id, exon_stable_id, species, analysis, exon_seqs[1:])
of.close()
print species, "done"
cursor.close()
db .close()
