def random_human_genes(num):
genes_set = set()
for rec in SeqIO.parse(
open(
"data\UniProt full Proteomes\uniprot-proteome-full H**o sapiens (71599 proteins).fasta"
), 'fasta'):
seq = str(rec.seq)
uniqueIdentifier, entryName, proteinName, organismName, geneName = \
utils.parse_UniProtKB_header(rec.description)
genes_set.add(geneName)
return random.sample(genes_set, num)
def get_human_proteins_that_contain_saar(SAAR):
rcp_set = set()
genes_set = set()
for rec in SeqIO.parse(
open(
"data\UniProt full Proteomes\uniprot-proteome-full H**o sapiens (71599 proteins).fasta"
), 'fasta'):
seq = str(rec.seq)
uniqueIdentifier, entryName, proteinName, organismName, geneName = \
utils.parse_UniProtKB_header(rec.description)
genes_set.add(geneName)
if SAAR in seq:
rcp_set.add(geneName)
return rcp_set
kmer for kmer in self.repeating_non_overlapping_kmers
if self.seq.find(kmer) +
params.MIN_DIST_BETWEEN_REPETITIONS < self.seq.rfind(kmer)
]
all_proteins = list() # all_proteins[i] = PROTEIN()_OBJECT
protein_seq = dict() # protein_seq[GENE_NAME] = AMINO_ACID_SEQUENCE
print "Reading Uniprot file and generating k-mers list for each protein..."
created_protein_names = set(
) # prevent creation of two similar protein objects
for rec in SeqIO.parse(open(params.HUMAN_PROTEOME), 'fasta'):
seq = str(rec.seq)
uniqueIdentifier, entryName, proteinName, organismName, geneName = \
utils.parse_UniProtKB_header(rec.description)
if geneName == '':
print 'Ignoring unknown gene: %s' % rec.description
continue
if geneName in created_protein_names:
print 'Ignoring duplicate gene: %s' % rec.description
continue
# create a new Protein object
created_protein_names.add(geneName)
protein_seq[geneName] = seq
all_proteins.append(Protein(geneName, seq, params.K))
print
print "Counted k-mer (k=%d) for %d different genes (proteins)." % (
params.K, len(all_proteins))
def main(proteome_file, similar_diluted):
all_proteins = list() # all_proteins[i] = PROTEIN()_OBJECT
protein_seq = dict() # protein_seq[GENE_NAME] = AMINO_ACID_SEQUENCE
print "Reading Uniprot file and generating k-mers list for each protein..."
created_protein_names = set(
) # prevent creation of two similar protein objects
duplicate_genes_ignored = 0
for rec in SeqIO.parse(open(proteome_file), 'fasta'):
seq = str(rec.seq)
uniqueIdentifier, entryName, proteinName, organismName, geneName = \
utils.parse_UniProtKB_header(rec.description)
if geneName == '':
geneName = uniqueIdentifier
#print 'Using uncharacterized gene with identifier %s' % uniqueIdentifier
#print 'Ignoring unknown gene: %s' % rec.description
#continue
if geneName in created_protein_names:
duplicate_genes_ignored += 1
#print 'Ignoring duplicate gene: %s' % rec.description
continue
# create a new Protein object
created_protein_names.add(geneName)
protein_seq[geneName] = seq
all_proteins.append(Protein(geneName, seq, params.K))
print
print "Ignored %d duplicate genes." % duplicate_genes_ignored
print "Counted k-mer (k=%d) for %d different genes (proteins)." % (
params.K, len(all_proteins))
pb = Progressbar('Generating frequency dictionary for k-mers')
kmers_frequency = dict() # track popularity of kmer accross all proteins
i = 0
for prot in all_proteins:
i += 1
pb.update_progress(i, len(all_proteins))
for kmer in prot.kmers:
if kmer not in kmers_frequency:
kmers_frequency[kmer] = set()
# add the new protein only if it's dissimilar enough from all other
# proteins that were already added and contain this kmer
protein_names = kmers_frequency[
kmer] # list of all prots that share this kmer
redundantProt = False
if similar_diluted:
for protein_name in protein_names:
#print '%s: Checking similarity of %s and %s' % (kmer, protein_name, prot.geneName)
if not utils.proteins_are_dissimilar(
protein_name, prot.geneName,
protein_seq[protein_name], prot.seq):
redundantProt = True
break
if not redundantProt:
kmers_frequency[kmer].add(prot.geneName)
redundantProt = False
print "Sorting frequent k-mers by frequency..."
most_frequenct_kmers = sorted(kmers_frequency,
key=lambda k: len(kmers_frequency[k]),
reverse=True)
print "Writing results to file..."
import datetime, time, csv
timestamp = datetime.datetime.fromtimestamp(
time.time()).strftime('%Y-%m-%d-%H%M%S')
filename_without_extension = os.path.splitext(
os.path.basename(proteome_file))[0]
dilution_status = 'with dilution' if similar_diluted else 'without dilution'
outfile = "{0} - frequent k{1}-mers - {2} - {3}.csv".format(
filename_without_extension, params.K, dilution_status, timestamp)
outfile = os.path.join(os.path.dirname(proteome_file), outfile)
with open(outfile, "wb") as csv_file:
writer = csv.writer(csv_file, delimiter=',')
writer.writerow([
'k-mer', 'number of proteins', 'percentage', 'all',
'(Out of %d proteins in total)' % len(all_proteins)
])
for kmer in most_frequenct_kmers:
total_proteins = len(kmers_frequency[kmer])
if total_proteins < 5:
break
percentage = round(float(total_proteins) / len(all_proteins), 6)
geneList = list(kmers_frequency[kmer])
#geneList = '\r\n'.join(geneList)
row = [kmer, total_proteins, percentage, geneList]
writer.writerow(row)
def main(proteome_file, output_dir):
all_proteins = list() # all_proteins[i] = PROTEIN()_OBJECT
protein_seq = dict() # protein_seq[GENE_NAME] = AMINO_ACID_SEQUENCE
print "Reading Uniprot file and generating k-mers list for each protein..."
created_protein_names = set(
) # prevent creation of two similar protein objects
for rec in SeqIO.parse(open(proteome_file), 'fasta'):
seq = str(rec.seq)
uniqueIdentifier, entryName, proteinName, organismName, geneName = \
utils.parse_UniProtKB_header(rec.description)
if geneName == '':
geneName = uniqueIdentifier
#print 'Using uncharacterized gene with identifier %s' % uniqueIdentifier
#print 'Ignoring unknown gene: %s' % rec.description
#continue
if geneName in created_protein_names:
print 'Ignoring duplicate gene: %s' % rec.description
continue
# create a new Protein object
created_protein_names.add(geneName)
protein_seq[geneName] = seq
all_proteins.append(Protein(geneName, seq, params.K))
print
print "Counted k-mer (k=%d) for %d different genes (proteins)." % (
params.K, len(all_proteins))
pb = Progressbar('Generating frequency dictionary for k-mers')
skipped_prots = 0
kmers_frequency = dict() # track popularity of kmer accross all proteins
i = 0
for prot in all_proteins:
i += 1
pb.update_progress(i, len(all_proteins))
"""
if prot.geneName.startswith('ZNF') or prot.geneName.startswith('ZF'):
skipped_prots += 1
continue
if prot.geneName.startswith('OR'):
skipped_prots += 1
continue
if prot.geneName.startswith('HOX'):
skipped_prots += 1
continue
if prot.geneName.startswith('IGKV'):
skipped_prots += 1
continue
"""
for kmer in prot.kmers:
if kmer not in kmers_frequency:
kmers_frequency[kmer] = set()
kmers_frequency[kmer].add(prot.geneName)
print "Sorting frequent k-mers by frequency..."
most_frequenct_kmers = sorted(kmers_frequency,
key=lambda k: len(kmers_frequency[k]),
reverse=True)
import datetime, time, csv
timestamp = datetime.datetime.fromtimestamp(
time.time()).strftime('%Y-%m-%d-%H%M%S')
outfile = '{}/frequent k{}-mers - {}.csv'.format(output_dir, params.K,
timestamp)
with open(outfile, "wb") as csv_file:
writer = csv.writer(csv_file, delimiter=',')
writer.writerow([
'k-mer', 'number of proteins', 'percentage of total', 'all',
'(Out of %d proteins in total)' %
(len(all_proteins) - skipped_prots)
])
for kmer in most_frequenct_kmers:
total_proteins = len(kmers_frequency[kmer])
if total_proteins < 10:
break
percentage = "{0:.4f}".format(
float(total_proteins) / (len(all_proteins) - skipped_prots))
geneList = list(kmers_frequency[kmer])
#geneList = '\r\n'.join(geneList)
row = [kmer, total_proteins, percentage, geneList]
writer.writerow(row)