def generate_cluster(sequences, dpm_subset, idx, sampler_config):
counts = [ [ 0.0 for j in range(sampler_config.tfbs_length) ] for i in range(4) ]
counts_gap = [ 0.0 ] * sampler_config.tfbs_length
baseline_prior = get_baseline_prior(dpm_subset.dpm_subset_tag(), sampler_config)
alpha = baseline_prior[0:4]
alpha_gap = baseline_prior[4]
components = len(dpm_subset)
for position in dpm_subset:
s = position[0]
p = position[1]
# loop over the motif
for j in range(sampler_config.tfbs_length):
# loop over all nucleotides plus counts for gaps
for i in range(4):
counts[i][j] += sequences[s][p+j][i]
counts_gap[j] += sequences[s][p+j][4]
return cluster_t(counts, counts_gap, alpha, alpha_gap, components, idx, dpm_subset.dpm_subset_tag())
def load_cluster(cluster_parser, sampler_config, cluster_name):
result = re.match('cluster_([0-9]+)', cluster_name)
counts, counts_gap = read_counts(cluster_parser, cluster_name)
components = int(cluster_parser.get('Cluster', '%s_components' % cluster_name))
identifier = int(result.group(1))
cluster_type = cluster_parser.get('Cluster', '%s_type' % cluster_name)
alpha = None
alpha_gap = None
sites = None
if cluster_parser.has_option('Cluster', '%s_sites' % cluster_name):
sites_str = cluster_parser.get('Cluster', '%s_sites' % cluster_name)
sites = dpm_subset_t(cluster_type)
for elem in parse_partition_elements(sites_str):
sites.insert(elem)
for (tag, prior) in zip(sampler_config.baseline_tags, sampler_config.baseline_priors):
if tag == cluster_type:
alpha = map(list, zip(*prior))[0:4]
alpha_gap = map(list, zip(*prior))[4]
if alpha == None or alpha_gap == None:
raise IOError("Baseline prior not found in sampler config.")
return cluster_t(counts, counts_gap, alpha, alpha_gap, components, identifier, cluster_type, sites = sites)
def generate_cluster(sequences, dpm_subset, idx, sampler_config, index_error = True):
length = dpm_subset.model_id().length
counts = [ [ 0.0 for j in range(length) ] for i in range(4) ]
counts_gap = [ 0.0 ] * length
baseline_prior = get_baseline_prior(dpm_subset.model_id(), sampler_config)
alpha = [ [ baseline_prior[i][0] for j in range(length) ] for i in range(4) ]
alpha_gap = baseline_prior[4] * length
components = len(dpm_subset)
for r in dpm_subset:
s = r.index()[0]
p = r.index()[1]
# loop over the motif
for j in range(r.length()):
# loop over all nucleotides plus counts for gaps
if r.reverse():
try: sequences[s][p+j][DNA.complement(i)]
except IndexError:
if index_error:
raise
else:
print >> sys.stderr, "Warning: index %d:%d out of range!" % (s, p+j)
continue
for i in range(4):
counts[i][r.length()-j-1] += sequences[s][p+j][DNA.complement(i)]
else:
try: sequences[s][p+j][i]
except IndexError:
if index_error:
raise
else:
print >> sys.stderr, "Warning: index %d:%d out of range!" % (s, p+j)
continue
for i in range(4):
counts[i][j] += sequences[s][p+j][i]
counts_gap[j] += sequences[s][p+j][4]
return cluster_t(counts, counts_gap, alpha, alpha_gap, components, idx, dpm_subset.model_id(), sites=dpm_subset)