def addGene(self, node1, node2):
"""
Add a randomly weighted gene between two nodes.
Args:
node1: the first node
node2: the second node
"""
# connect the two valid, unconnected nodes
g = Gene()
g.source_neuron = node1
g.target_neuron = node2
g.weight = random.random() * 4 - 2
# check to see if the gene already exists in the index
for index, gene in gene_index.items():
if gene.source_neuron == node1 and gene.target_neuron == node2:
# if found, use that same innovation number
self.genes[index] = g
return
# if it's not in the index, add it and record in the index
if gene_index:
self.genes[max(gene_index) + 1] = g
gene_index[max(gene_index) + 1] = g
else:
gene_index[0] = g
self.genes[0] = g
def mutateAllWeights(self):
"""Find all genes, and alter their weights a little."""
g = Gene()
for innov in self.genes:
g.copy(self.genes[innov])
g.weight += random.random() * 0.2 - 1
self.genes[innov] = g
def init_genes(self, keys):
for x in range(len(self.maze)):
for y in range(len(self.maze)):
if self.maze[x][y] != "_":
self.maze[x][y] = Gene(self.maze[x][y], True)
else:
self.maze[x][y] = Gene(keys[randint(0, len(keys) - 1)])
def __init__(self, timetables):
self.fitness = 1
self.genes = []
for timetable in timetables:
gene = Gene(timetable)
gene.permutate()
self.genes.append(gene)
def gen_population(self, size):
population = []
for i in range(0, size - 1, 1):
g = Gene("")
g.gen_gene(size)
population.append(g)
return population
def addGene(self, geneName, transcript, chromosome, strand, startPosition, endPosition, exonStarts, exonEnds):
#create the gene object
gene = Gene(geneName, chromosome, strand, startPosition, endPosition)
gene.addTranscriptID(transcript)
gene.addExonsToGene(exonStarts, exonEnds)
#add the gene to the hash
self.__genes[geneName] = gene
def __init__(self, timetables):
self.genes = []
self.fitness = 1
for tt in timetables:
temp_gene = Gene(tt)
temp_gene.perm()
self.genes.append(temp_gene)
def mutateOneWeight(self):
"""Find a random gene, and alter its weight a little."""
if self.genes:
innov = random.sample(self.genes, 1)[0]
else:
return
v = random.random()
g = Gene()
g.copy(self.genes[innov])
g.weight += v * 0.2 - 0.1
self.genes[innov] = g
def mutation(self, mutate_gene: gene.Gene):
"""Insert random small gene into given gene."""
mutation_size = int(np.random.exponential(self.mutation_scale))
if mutation_size == 0:
return
mutation_gene = gene.Gene(
np.random.randint(0, 2, (mutation_size, mutation_size)))
mutatepoint_x = np.random.randint(0, self.genesize_x)
mutatepoint_y = np.random.randint(0, self.genesize_y)
mutate_gene.insert_subgene(mutation_gene, mutatepoint_x, mutatepoint_y)
return
def init_genome(self):
return [[
Node(0, 0, role='input'),
Node(1, 1, role='flatten'),
Node(2, 2, role='output')
],
[
Gene(3, 0, 1, mutate_to=[[KernelGene, DenseGene],
[1, 0]]).mutate_random(),
Gene(4, 1, 2, mutate_to=[[KernelGene, DenseGene],
[0, 1]]).mutate_random()
]]
def generate_genes(self, max_genes, max_conds, seed):
'''
Randomly generate a variable size chromosome.
'''
random.seed(seed)
self.default_class = random.randint(0, 9)
for i in range(random.randint(1, max_genes)):
gene = Gene()
gene.generate_conditions(max_conds, random.random())
if self.one_vs_all != -1:
gene.hand_class = self.one_vs_all
self.genes.append(gene)
def insertion(self, max_genes, max_conds, insertion_rate, seed):
'''
Mutation Function
Randomly inserts genes into the chromosome.
'''
random.seed(seed)
for i in range(max_genes):
if random.random() < (insertion_rate):
gene = Gene()
gene.generate_conditions(max_conds, random.random())
if self.one_vs_all != -1:
gene.hand_class = self.one_vs_all
self.genes.append(gene)
def load(self, s):
""" Load the Chromosome from
save passed as parameter.
"""
self._size = s["size"]
self._gene_count = s["gene_count"]
self._genes = []
for g in s["genes"]:
new_gene = Gene(self._size)
new_gene.load(g)
self._genes.append(new_gene)
return s["generation"]
def cross_over(geneOne, geneTwo, pivot):
g1 = ""
g2 = ""
for i in range(0, pivot):
g1 += str(geneOne.info[i])
g2 += str(geneTwo.info[i])
for i in range(pivot, len(geneOne.info)):
g1 += str(geneTwo.info[i])
g2 += str(geneOne.info[i])
g1 = Gene(g1)
g2 = Gene(g2)
return [g1,g2]
def __init__(self, InputUnits=16, OutputUnits=4, graph=None, N_hidden=1):
self.inovationNumber = 0
self.calculated_fitness = 0
self.genome = nx.DiGraph()
N = 16 # input
self.inputGenes = []
self.outputGenes = []
self.graph = graph
self.HLayers = N_hidden
#adding inputnodes
for i in range(0, InputUnits):
gn = Gene(1.0,
input_gene=True,
response=1.0,
graph=self.graph,
indexLayer=0)
self.inputGenes.append(gn)
self.genome.add_node(gn)
#adding outputnodes
for i in range(0, OutputUnits):
gn = Gene(0.0,
output_gene=True,
graph=self.graph,
indexLayer=N_hidden + 1)
self.outputGenes.append(gn)
self.genome.add_node(gn)
bias = Gene(0.0, bias_gene=True, graph=self.graph, indexLayer=0)
self.genome.add_node(bias)
# random conection
#self.add_conn_rnd()
#self.add_conn_rnd()
self.add_connection2()
self.add_connection2()
self.add_node_hidden()
self.specie = ""
self.champion = False
self.name = self.new_name()
def elitism(population):
population.sort(reverse=True)
tophalf = []
for _ in range(generationsize / 2):
tophalf.append(population[_][1])
newpopulation = []
for _ in range(generationsize / 4):
parent1 = random.choice(tophalf)
parent2 = random.choice(tophalf)
children = crossover(parent1.chromosone, parent2.chromosone)
newpopulation.append(Gene(mutation(children[0])))
newpopulation.append(Gene(mutation(children[1])))
return tophalf + newpopulation
def __init__(self, sample_ID, sample_ICD, gene_name, Output_geno,
amplicon_name, Active_score, Drug_action, Low_coverage, Range,
Gene_KB):
Gene.__init__(self, sample_ID, sample_ICD, gene_name, Drug_action,
amplicon_name, Range)
if self.ICD_relevant:
self.genotype = self.get_geno(Output_geno)
self.phenotype, self.score_allele = self.get_pheno(Active_score)
self.potential_allele = self.get_pot_allele(Gene_KB)
self.potential_phenotype = self.get_pot_pheno(Active_score)
else:
self.genotype = None
self.phenotype, self.score_allele = None, None
self.potential_allele = None
self.potential_phenotype = True
def before_first_request():
options = app.config['OPTIONS']
objects = [
Gene.fromfile(f) for f in content_file_iterator(options.content)
]
objects = [o for o in objects if o is not None]
objects = dict([(g.id, g) for g in objects])
app.config['GENES'] = objects
print("In __init__ before_first_request, objects={}".format(objects))
# Configure logging
if not app.debug:
from logging.handlers import SMTPHandler, RotatingFileHandler
from logging import Formatter
mail_handler = SMTPHandler('127.0.0.1', '*****@*****.**',
app.config['ADMINS'], 'ogrdb failed')
mail_handler.setLevel(logging.ERROR)
log_path = app.config['LOG_PATH']
file_handler = RotatingFileHandler(log_path,
maxBytes=1 << 20,
backupCount=5)
file_handler.setFormatter(
Formatter('%(asctime)s %(levelname)s: %(message)s '
'[in %(pathname)s:%(lineno)d]'))
app.logger.setLevel(logging.INFO)
file_handler.setLevel(logging.INFO)
app.logger.addHandler(file_handler)
app.logger.addHandler(mail_handler)
def makeNewPopulation(self,
list_of_genes=None,
population_size=None,
num_of_chromo=None,
length_of_chromo=None):
if list_of_genes == None:
for i in range(population_size):
gene = Gene([], num_of_chromo, length_of_chromo)
gene.makeRandomGene()
player = Player(gene)
self.addPlayer(player)
else:
for i in range(population_size):
player = Player(list_of_genes[i])
self.addPlayer(player)
def selection_1(self, fit_data, strength_threshhold):
# normalize accuracy for pool selection for i in fit_data:
i['accuracy'] /= accuracy_a.max()
# pool selection
mating_pool = []
while len(mating_pool) < mating_pool_size:
choice = random.choice(fit_data)
r = np.random.rand()
if choice['accuracy'] > r:
mating_pool.append(choice)
# perish connections with weak strength from gene
for i in mating_pool:
temp = i['strength_matrix'] >= strength_threshhold
i['gene'].connections &= temp
i['gene'].connections_number = i['gene'].connections.sum()
# mating pool for reproduction
mating_pool = [i['gene'] for i in mating_pool]
# Sample two parents from mating pool, reproduce children and impose muatation on the newborn.
gene_pool = []
for _ in range(P):
new_gene = Gene.crossover(random.sample(mating_pool, 2))
gene_pool.append(new_gene)
# mutation
return gene_pool
def read_genome_from_file(self, filename):
f = open(filename, "r")
f.readline()
line = f.readline()
var = line.split(" ")
fitness = var[0]
orig_fitness = var[1]
f.readline()
nodes = []
while line != "gene":
line = f.readline()
if line == "gene":
break
var = line.split(" ")
from node import Node
newnode = Node().new_node(var[0], var[1])
nodes.append(newnode)
genes = []
while line != "":
line = f.readline()
if line == "":
break
var = line.split(" ")
from gene import Gene
newgene = Gene().add_gene_with_no_trait()
genes.append(newgene)
def get_son(self, parents):
cbti = random.randint(0, (len(parents) - 1))
cbt = parents[cbti].col_by_tree
ssi = random.randint(0, (len(parents) - 1))
ss = parents[ssi].subsample
mcwi = random.randint(0, (len(parents) - 1))
mcw = parents[mcwi].min_child_weight
mdi = random.randint(0, (len(parents) - 1))
md = parents[mdi].max_depth
nei = random.randint(0, (len(parents) - 1))
ne = parents[nei].n_estimators
lri = random.randint(0, (len(parents) - 1))
lr = parents[lri].learning_rate
wayi = random.randint(0, (len(parents) - 1))
way = parents[wayi].way
nnei = random.randint(0, (len(parents) - 1))
n_neighbors = parents[nnei].way
son = Gene(cbt, ss, mcw, md, ne, lr, way, n_neighbors)
return son
def to_chromosome(solution):
"""Create an simple/lighter internal representation of a solution to be used by the GA operations.
"""
genes = {}
host_failure_rates = {}
for h in solution["hosts"]:
genes[h["host_name"]] = []
if "failure_rate" in h:
host_failure_rates[h["host_name"]] = h["failure_rate"]
else:
host_failure_rates[h["host_name"]] = 0.0
for v in solution["vnfs"]:
if "place_at" in v and v["place_at"] is not None:
for p in v[
"place_at"]: # v[.] is a list, although it normally has just 1 element
if "failure_rate" in v:
vfrate = v["failure_rate"]
else:
vfrate = 0.0
genes[p].append({
"vnf_name": v["vnf_name"],
"failure_rate": vfrate
})
G = []
for hname, vnfs in genes.iteritems():
G.append(Gene(hname, vnfs,
host_failure_rate=host_failure_rates[hname]))
return Chromosome(G)
def build_site_id_dictionaries():
if os.path.isfile('site_id_dictionaries.pkl'):
with open('site_id_dictionaries.pkl', 'rb') as f:
print('loading site_id_dictionaries.pkl... ', end='', flush=True)
global site_id_dictionary_left
global site_id_dictionary_right
global gene_id_dictionary_left
global gene_id_dictionary_right
global gene_id_site_id_arcs
global site_id_gene_id_arcs
site_id_dictionary_left, site_id_dictionary_right, gene_id_dictionary_left, gene_id_dictionary_right, gene_id_site_id_arcs, site_id_gene_id_arcs = pickle.load(
f)
print('DONE')
else:
with open('processed/sites_with_scored_interactions.tsv',
'r') as infile:
content = csv.reader(infile,
delimiter='\t',
quoting=csv.QUOTE_NONE)
header = content.__next__()
i = 0
j = 0
for row in content:
gene_id = row[0]
gene_symbol = row[1]
transcript_id = row[2]
utr_start = row[3]
utr_end = row[4]
seed_match_type = row[5]
site = Site(gene_id, gene_symbol, transcript_id, utr_start,
utr_end, seed_match_type)
site_id_dictionary_left[site] = i
site_id_dictionary_right[i] = site
gene = Gene(gene_id, gene_symbol, transcript_id)
i += 1
if not gene in gene_id_dictionary_left:
gene_id_dictionary_left[gene] = j
gene_id_dictionary_right[j] = gene
j += 1
site_id = i - 1
gene_id = gene_id_dictionary_left[gene]
if not gene_id in gene_id_site_id_arcs:
gene_id_site_id_arcs[gene_id] = list()
gene_id_site_id_arcs[gene_id].append(site_id)
if site_id in site_id_gene_id_arcs:
print(
f'error: site_id = {site_id} already in site_id_gene_id_arcs'
)
os._exit(1)
site_id_gene_id_arcs[site_id] = gene_id
with open('site_id_dictionaries.pkl', 'wb') as f:
print('saving site_id_dictionaries.pkl... ', end='', flush=True)
pickle.dump([
site_id_dictionary_left, site_id_dictionary_right,
gene_id_dictionary_left, gene_id_dictionary_right,
gene_id_site_id_arcs, site_id_gene_id_arcs
], f)
print('DONE')
def openDB(self, fileName=''):
if fileName == '':
fileName = self.geneDBname
with open(fileName, 'r+') as f:
self.geneList = []
for x in f:
self.geneList.append(Gene(x))
def kill(self):
self.Genes.clear()
for in_node, out_node in it.product(self.inputs, self.outputs):
self.Genes[in_node].append(Gene((in_node, out_node)))
for gene in sum(self.Genes.values(), []):
gene.mutate()
self._add_node_()
self._add_link_()
def testFromBounds(self):
bound1 = SeqFeature(FeatureLocation(300, 400, strand=-1), qualifiers={'label': ['bound1']})
bound2 = SeqFeature(FeatureLocation(100, 200, strand=1), qualifiers={'label': ['bound2']})
self.record.features += [bound1, bound2]
gene = Gene.from_bounds(self.record, 'test_gene', 'bound1', 'bound2')
gene_seq = gene.extract(gene.record).seq
expected_gene_seq = self.record[100:400].reverse_complement().seq
self.assertEqual(str(gene_seq), str(expected_gene_seq))
def add_genes(self, unpacked_json):
self.genes_json = unpacked_json
for gene in self.genes_json['result']['Genes']:
# Use preexisting gene objects to save time on searching for name matches and so forth
# if str(gene['GeneSymbol']) in genesdict:
# self.genes.append(genesdict[gene['GeneSymbol']])
# else:
g = Gene(self, json=gene, hgncHandler=self.hgncHandler)
self.genes.append(g)
def create_population(population_size, input_board, width, height):
population = []
for x in range(population_size):
gene = Gene(input_board, width, height)
population.append(gene)
population.sort()
#pp.pprint(population)
return population
def reproduce(self, other):
if self == other:
return
childGenes = [
Gene(myGene.choose(), otherGene.choose())
for myGene, otherGene in zip(self.genes, other.genes)
]
return Person(self, self.genome, childGenes)
def add_gene(self):
""" Add a new Gene to the chromosome.
"""
if Chromosome.ADD_CHANCE < random.random():
self._gene_count += 1
self._genes.append(Gene(self._size))
else:
pass
def __init__(self, inputs, outputs):
# create list of input and output Nodes
self.inputs = [InputNode(i) for i in range(inputs)]
self.outputs = [OutputNode(i) for i in range(inputs, inputs + outputs)]
Gene.node_counter = inputs + outputs
# create every possible link from input to output nodes
self.Genes = defaultdict(list)
for in_node, out_node in it.product(self.inputs, self.outputs):
self.Genes[in_node].append(Gene((in_node, out_node)))
def __init__(self, timetables, cRate = 1e0, mRate = 1e-1):
self.genes = []
self.crossoverRate = cRate
self.mutationRate = mRate
self.fitness = 0
for tt in timetables:
self.genes.append(Gene(tt))
self.fitness = Chromosome.getFitness()
def crossover(self, creature1: Creature, creature2: Creature) -> Creature:
# 형질들은 다 똑같이 갖고있어야함
halfChromosome1 = creature1.getHalfChromosomeAsList(
) # random half chromosome
halfChromosome2 = creature2.getHalfChromosomeAsList()
resultChromosome = Chromosome()
for i in range(len(halfChromosome1)):
gene1 = halfChromosome1[i]
gene2 = halfChromosome2[i]
resultChromosome.appendGene(
Gene(gene1.geneName, gene1.stat, gene1.isSuperior),
Gene(gene2.geneName, gene2.stat, gene2.isSuperior))
x, y = 0, 0
while True:
x = random.randrange(0, c.MAP_WIDTH)
y = random.randrange(0, c.MAP_HEIGHT)
if self.map_[x][y] != c.OCCUPIED:
break
return Creature(x, y, resultChromosome)
def create(self, k):
generation = []
for i in range(0, k):
v1 = random.random()
v2 = random.random()
v3 = random.random()
g = Gene(v1, v2, v3)
generation.append(g)
return generation
def mate(gene1,gene2):
# print'in mate'
w1=gene1.weights
w2=gene2.weights
# print'w1:',len(w1),w1
# print'w2:',len(w2),w2
w3=[]
for i in range(0,len(w1)):
if randrange(0,2):
w3.append(w1[i])
else:
w3.append(w2[i])
t1=gene1.thresholds
t2=gene2.thresholds
t3=[]
for i in range(0,len(t1)):
if randrange(0,2):
t3.append(t1[i])
else:
t3.append(t2[i])
g=Gene(structure=gene1.structure,weights=w3,thresholds=t3)
g.set()
return g
def breed(job_data, elites, peasants):
new_genes = []
parents = elites + peasants
print("Parents")
print(parents)
print("************")
# The elite get to pass their genes directly
for e in elites:
file_name = job_data.genes[e[0]][0].file_name
# file_name = 'test'
gene = Gene(file_name)
new_genes.append(gene)
# Now let everyone randomly mate with one another,
# using the fitness as the chance of mating
for i in range(job_data.population - job_data.elites):
file_name = job_data.genes[0][0].file_name
gene = Gene(file_name)
r1 = select_parent(parents)
different = False
while not different:
r2 = select_parent(parents)
if r2 != r1:
different = True
parent1 = parents[r1]
parent2 = parents[r2]
p_floats = mate(job_data, parent1[0], parent2[0])
gene.p_floats = p_floats
new_genes.append(gene)
# Update the naming
for i, n in enumerate(new_genes):
n.file_name = job_data.file_name + "AM1_0P" + str(i)
build_input(n.file_name + ".com", n.header, n.coordinates,
n.params, n.p_floats)
for i in range(job_data.population):
file_name = job_data.file_name + "AM1_0P" + str(i)
job_data.genes[i][0] = Gene(file_name)
def mutateAddNeuron(self):
"""
Add a neuron at the location of a random gene.
This is done by disabling the original gene,
then adding a new node and connections between
the new node and the original gene nodes.
"""
# get one random gene
if self.genes:
innov = random.sample(self.genes, 1)[0]
else:
return
gene = self.genes[innov]
node1 = gene.source_neuron
node2 = gene.target_neuron
# make sure our nodes aren't two consecutive integers
if node1 + 1 > node2 - 1:
return
g = Gene()
g.copy(gene)
# create a new node
node = random.randint(node1 + 1, node2 - 1)
# make sure our node isn't the same as any existing nodes
if node in self.nodes:
return
# should be a hidden node
assert node >= 1 and node <= MAX_LAYER
# morph the gene into two genes and add the node
g1 = Gene()
g1.source_neuron = node1
g1.target_neuron = node
g1.weight = 1
g2 = Gene()
g2.source_neuron = node
g2.target_neuron = node2
g2.weight = gene.weight
self.nodes.add(node)
# disable the original gene, replace it with the new ones
g.enabled = False
self.genes[innov] = g
self.genes[max(gene_index) + 1] = g1
self.genes[max(gene_index) + 2] = g2
# add the new genes to the index
gene_index[max(gene_index) + 1] = g1
gene_index[max(gene_index) + 1] = g2
def read_reference(self,file_ref):
print "Read reference ..."
#get file size
file_size=os.stat(file_ref).st_size
#Parse GTF file
in_reference = open(file_ref)
genes={}
no_line = 0
current_position=0
for gtf_line in in_reference:
no_line = no_line + 1
current_position+=len(gtf_line)
#if (no_line< 1e5 and no_line % 1000 == 0) or (no_line<1e6 and no_line % 1e4 ==0) or (no_line>1e6 and no_line % 1e5 ==0) :
if no_line % 1e5 ==0 :
self.log_already_completed("{0} lines read from reference".format(no_line),file_size,current_position)
if re.match("^#.*$",gtf_line):
continue
gtf_line = gtf_line.rstrip('\r\n')
elmts = gtf_line.split('\t')
gene_chr=elmts[0]
gene_chr=gene_chr.lower().replace("chr","")
start=int(elmts[3])
end=int(elmts[4])
if Join_with_gtf.debug and gene_chr != '1' :
break
feature=elmts[2]
annot=elmts[8]
me=re.match('^gene_id "([^"]+)".*$',annot)
if me :
gene_id=me.group(1)
else :
#Feature not related to a gene_id
gene_id=""
#sys.exit("Unable to find gene_id value on line #{0} of file '{1}'. Exiting".format(no_line,file_ref))
if feature == "gene":
gene_start=start
gene_end=end
strand=elmts[6]
if strand == "-" : strand=-1
elif strand == '+' : strand=1
else: sys.exit("Unexpected strand value on line #{0} of file '{1}' : '{2}'. Exiting".format(no_line,file_ref,strand))
if gene_id not in genes :
gene=Gene(gene_id,gene_chr,gene_start,gene_end,strand)
genes[gene_id]=gene
else :
gene=genes[gene_id]
gene.set_location(gene_chr,gene_start,gene_end)
gene.set_strand(strand)
#gene start and end are defined in this line, therefore we can compute :
#tss, promoter and tss
self.features_found["promoter"]=1
self.features_found["tss"]=1
self.features_found["tts"]=1
self.features_found["gene"]=1
gene.gene_model_has_been_defined()
#elif feature not in("CDS","UTR","transcript") :
else :
if gene_id not in genes :
gene=Gene(gene_id,gene_chr)
genes[gene_id]=gene
else :
gene=genes[gene_id]
if feature == "start_codon" :
self.features_found["utr5"]=1
elif feature == "stop_codon" :
self.features_found["utr3"]=1
elif feature == "exon" :
self.features_found["exon"]=1
self.features_found["intron"]=1
else :
self.features_found[feature.lower()]=1
gene.add_feature(feature,start,end)
in_reference.close()
print "\n\t{0} lines read from reference in total.".format(no_line)
#Check that all features listed in configuration file has been found at least once
for feature in self.features_found :
if self.features_found[feature.lower()] == 0 :
sys.stderr.write(("Warning : feature named '{0}' found in 'feature_priorities' parameter. "+
"This feature has never been found in reference file '{1}'.\n").format(
feature, file_ref
))
#Complete feature_properties with the one found in gtf files but not requested by user
#Otherwise when we will try to order feature overlapping with a given region
#sorted(overlaps, key=lambda ovlp: self.feature_priorities[ ovlp.value[0] ])
#It will raise an exception.
for feature in self.features_found :
if feature.lower() not in self.feature_priorities :
self.feature_priorities[feature.lower()]=None
#define downstream/upstream boundaries
promoter_downstream= self.theme.get_parameter("promoter_downstream")
promoter_upstream= self.theme.get_parameter("promoter_upstream")
tss_downstream= self.theme.get_parameter("tss_downstream")
tss_upstream= self.theme.get_parameter("tss_upstream")
tts_downstream= self.theme.get_parameter("tts_downstream")
tts_upstream= self.theme.get_parameter("tts_upstream")
#print "promoter_upstream={0}".format(promoter_upstream)
#print "promoter_downstream={0}".format(promoter_downstream)
#print "tss_upstream={0}".format(tss_upstream)
#print "tss_downstream={0}".format(tss_downstream)
#print "tts_upstream={0}".format(tts_upstream)
#print "tts_downstream={0}".format(tts_downstream)
#Initialize dictionnaries
features={}
gene_boundaries={}
#Build gene model
print "Build gene model ..."
no_gene=0
for gene_id in genes :
gene=genes[gene_id]
(gene_chr,gene_start,gene_end)=gene.get_coordinates()
no_gene+=1
if no_gene % 1000 == 0 :
self.log_already_completed("{0} genes treated".format(no_gene),len(genes),no_gene)
if gene_chr not in features :
features[gene_chr]=IntervalTree()
gene_boundaries[gene_chr]=IntervalTree()
if gene.gene_model_is_defined() :
if gene_chr not in gene_boundaries :
gene_boundaries[gene_chr]=IntervalTree()
gene_boundaries[gene_chr].insert_interval( Interval(gene_start,gene_end, value=["gene",gene_id] ) )
#Promoter
if gene.strand == 1 :
(start,end)=gene.get_promoter(promoter_upstream,promoter_downstream)
else :
(start,end)=gene.get_promoter(promoter_downstream,promoter_upstream)
features[gene_chr].insert_interval( Interval(start,end, value=["promoter",gene_id] ) )
#5' UTR
(start,end)=gene.get_utr5()
if start is not None:
features[gene_chr].insert_interval( Interval(start,end, value=["utr5",gene_id] ) )
#TTS
if gene.strand == 1 :
(start,end)=gene.get_tss(tss_upstream,tss_downstream)
else :
(start,end)=gene.get_tss(tss_downstream,tss_upstream)
features[gene_chr].insert_interval( Interval(start,end, value=["tss",gene_id] ) )
#Intron / Exon
(intron_coords,exon_coords)=gene.get_introns_exons()
#Debug
#if gene.gene_id == "ENSBTAG00000000010" :
# print "gene_id '{0} / intron={1} / exon={2}".format(gene.gene_id,intron_coords,exon_coords)
for exon_coord in exon_coords :
(start,end)=exon_coord
features[gene_chr].insert_interval( Interval(start,end, value=["exon",gene_id] ) )
for intron_coord in intron_coords :
(start,end)=intron_coord
features[gene_chr].insert_interval( Interval(start,end, value=["intron",gene_id] ) )
#TTS
if gene.strand == 1 :
(start,end)=gene.get_tts(tts_upstream,tts_downstream)
else :
(start,end)=gene.get_tts(tts_downstream,tts_upstream)
features[gene_chr].insert_interval( Interval(start,end, value=["tts",gene_id] ) )
#3' UTR
(start,end)=gene.get_utr3()
if start is not None:
features[gene_chr].insert_interval( Interval(start,end, value=["utr3",gene_id] ) )
#Other features
for feature in gene.get_other_features() :
(start,end,feature)=feature
features[gene_chr].insert_interval( Interval(start,end, value=[feature,gene_id] ) )
print "\n\t{0} genes treated in total.".format(no_gene)
return (features,gene_boundaries)