def wrapperProfileGraph(parentFile, contactFile):
'''
Draws a graph of the number of clashes at each recombination point
'''
pdbName = contactFile.split('_')[0][-4:]
parent_list = schema.readMultipleSequenceAlignmentFile(
file(parentFile, 'r'))
parents = [p for (k, p) in parent_list]
pdb_contacts = schema.readContactFile(file(contactFile, 'r'))
clash_data = [[] for x in parents[0]]
for i in range(1, len(parents)):
print i
#This reshuffles the alignment to make the first and second sequences the ones analysed. It was needed as SCHEMA is limited to 9 sequences.
newList = [parents[0], parents[i]]
for x in range(1, len(parents)):
if not i == x:
newList.append(parents[x])
#Graphs for hotspots
for residue in range(0, len(parents[0])):
crossovers = [residue]
contacts = schema.getSCHEMAContactsWithCrossovers(
pdb_contacts, newList, crossovers)
fragments = schema.getFragments(crossovers, parents[0])
clash_data[residue].append(
schema.getChimeraDisruption('21', contacts, fragments,
newList))
means = [np.mean(values) for values in clash_data]
StDev = [np.std(values) for values in clash_data]
makeBarGraph(means, StDev, pdbName)
def frags(self, parent):
return schema.getFragments(self.crossovers(), parent)
def curve(results, parents, bin_width, max_samples=1e10):
"""Compute a curve of average energy and average mutation, with the latter binned
by bin_width.
"""
if len(results) < 1:
# Nothing to do!
return
(e, xovers, lmin, lmax) = results[0]
num_crossovers = len(xovers)
#print "# No. of RASPP results:", len(results)
# Because the RASPP curve does not involve the length min/max values, we can collapse
# the set of RASPP results to only those with unique crossovers. This can
# greatly improve performance, since crossover patterns are often duplicated
# across min/max values.
unique_results = set([(avg_energy, tuple(crossovers))
for (avg_energy, crossovers, l_min, l_max) in results
])
#print "# No. of unique RASPP results:", len(unique_results)
# Now compute the average mutation levels for these unique libraries.
avg_E_ms = []
for (avg_energy, crossovers) in unique_results:
crossovers = list(crossovers)
fragments = schema.getFragments(crossovers, parents[0])
avg_m = schema.averageMutationSampled(fragments, parents, max_samples)
avg_E_ms.append((avg_energy, avg_m, crossovers))
ms = [m for (E, m, crossovers) in avg_E_ms]
(min_m, max_m) = (min(ms), max(ms))
num_bins = int((max_m - min_m) / bin_width) + 1
# Assemble the RASPP curve. If num_samples exceeds the library size,
# then this curve is approximate.
approx_curve = []
for i in range(num_bins):
approx_curve.append(None)
for (E, m, crossovers) in avg_E_ms:
bin_number = int((m - min_m) / bin_width)
# If there's an existing value in this bin, check it
if approx_curve[bin_number]:
(E_old, m_old, crossovers_old) = approx_curve[bin_number]
# If lower E in this bin, substitute it
if E < E_old:
approx_curve[bin_number] = (E, m, crossovers)
else: # Otherwise just add it
approx_curve[bin_number] = (E, m, crossovers)
# It may be that the approximate curve is exact. If so, just return it.
library_size = len(parents)**(num_crossovers + 1)
approximate = (library_size > max_samples)
if not approximate:
return [r for r in approx_curve if r]
# If the curve IS approximate, we'll do a final pass so that
# the bin values are correct. Some libraries may still be
# incorrectly binned because we've
# Compute the exact mutation numbers for the lowest-E libraries
final_curve = []
for r in approx_curve:
if r:
(E, approx_m, crossovers) = r
fragments = schema.getFragments(crossovers, parents[0])
true_avg_m = schema.averageMutation(fragments, parents)
final_curve.append((E, true_avg_m, crossovers))
#print "%1.2f\t%1.2f" % (true_avg_m, approx_m)
return final_curve
def main(args):
arg_dict = parse_arguments(args)
if not confirm_arguments(arg_dict):
if args[0].split(os.path.sep)[-1] == "schemarandom.py":
print_usage(args)
return
# Flags and values
print_E = False
print_m = False
# Inputs:
# The alignment/fragment file name.
msa_file = arg_dict[ARG_MULTIPLE_SEQUENCE_ALIGNMENT_FILE]
# Read the alignment file to create a list of parents.
# The parents will appear in the list in the order in which they appear in the file.
parent_list = schema.readMultipleSequenceAlignmentFile(file(msa_file, "r"))
parents = [p for (k, p) in parent_list]
# Get the contacts
pdb_contacts = schema.readContactFile(file(arg_dict[ARG_CONTACT_FILE], "r"))
# Establish connection to output, either file or, if no output file is
# specified, to standard output.
if arg_dict.has_key(ARG_OUTPUT_FILE):
output_file = file(arg_dict[ARG_OUTPUT_FILE], "w")
else:
output_file = sys.stdout
# Get the number of libraries to evaluate.
if arg_dict.has_key(ARG_NUM_LIBRARIES):
num_libraries = int(arg_dict[ARG_NUM_LIBRARIES])
else:
num_libraries = int(1e3)
# Get the minimum fragment size.
if arg_dict.has_key(ARG_MIN_FRAGMENT_SIZE):
min_length = int(arg_dict[ARG_MIN_FRAGMENT_SIZE])
else:
min_length = 4
# Get the number of fragments -- one more than the number of crossovers.
num_fragments = int(arg_dict[ARG_NUM_CROSSOVERS]) + 1
num_parents = len(parents)
library_size = num_parents ** num_fragments
if arg_dict.has_key(ARG_MAX_CHIMERAS_PER_LIBRARY):
max_chimeras = min(library_size, int(arg_dict[ARG_MAX_CHIMERAS_PER_LIBRARY]))
else:
max_chimeras = library_size
if arg_dict.has_key(ARG_RANDOM_SEED):
random.seed(int(arg_dict[ARG_RANDOM_SEED]))
# Make libraries consistent with RASPP
(new_parents, identical_sites) = raspp.collapse_parents(parents)
if len(new_parents[0]) < num_fragments * min_length:
error_msg = (
"Minimum diversity length of %d is too large.\n%d "
+ "fragments with diversity %d cannot be found in a "
+ "sequence of length %d (with identities removed). Aborting..."
)
print error_msg % (min_length, num_fragments, min_length, len(parents[0]))
return
start_time = time.clock()
output_file.write("# <E>\t<m>\tcrossover points\n")
random_crossovers = []
for libnum in range(num_libraries):
crossovers = schema.generateRandomCrossovers(len(new_parents[0]), num_fragments - 1, min_length)
crossovers = raspp.translate_collapsed_indices(crossovers, identical_sites)
random_crossovers.append(crossovers)
for crossovers in random_crossovers:
fragments = schema.getFragments(crossovers, parents[0])
filtered_contacts = schema.getSCHEMAContactsWithCrossovers(pdb_contacts, parents, crossovers)
all_chimeras = []
if max_chimeras < library_size:
# Assemble a random sample of chimeras, with replacement
for n_chim in range(max_chimeras):
chim_index = random.randint(0, library_size - 1)
n2c = schema.base(chim_index, num_parents)
chimera_blocks = "".join(["1"] * (num_fragments - len(n2c)) + ["%d" % (int(x) + 1,) for x in n2c])
all_chimeras.append(chimera_blocks)
else: # We'll be covering all chimeras in the library; might as well get a good sample.
# The number of parents and fragments specifies all possible chimeras, regardless of
# crossover point positions, so pre-generate all chimeras.
max_chimeras = library_size
for i in range(library_size):
# The next two lines turn i into a chimera block pattern
# (e.g., 0 -> '11111111', 1 -> '11111112', 2 -> '11111113'...)
n2c = schema.base(i, num_parents)
chimera_blocks = "".join(["1"] * (num_fragments - len(n2c)) + ["%d" % (int(x) + 1,) for x in n2c])
all_chimeras.append(chimera_blocks)
# Randomly assort the chimeras
random.shuffle(all_chimeras)
# Calculate average E and m for the library or subsample
E_values = []
m_values = []
for chim_index in range(max_chimeras):
chimera_blocks = all_chimeras[chim_index]
E = schema.getChimeraDisruption(chimera_blocks, filtered_contacts, fragments, parents)
m = schema.getChimeraShortestDistance(chimera_blocks, fragments, parents)
E_values.append(E)
m_values.append(m)
average_E = schema.mean(E_values)
average_m = schema.mean(m_values)
xover_pat = "%d " * len(crossovers)
xover_str = xover_pat % tuple(crossovers)
output_file.write(("%1.4f\t%1.4f\t%s\n") % (average_E, average_m, xover_str))
output_file.flush()
total_time = time.clock() - start_time
output_file.write(
"# Finished in %1.2f seconds (%d libraries, %d chimeras)\n"
% (total_time, num_libraries, num_libraries * max_chimeras)
)
if arg_dict.has_key(ARG_OUTPUT_FILE):
output_file.close()
def main(args):
arg_dict = parse_arguments(args)
if not confirm_arguments(arg_dict):
if args[0].split(os.path.sep)[-1] == "schemaenergy.py":
print_usage(args)
return
# Flags and values
print_E = False
print_m = False
output_file = sys.stdout
# Inputs:
# The alignment/fragment file name.
msa_file = arg_dict[ARG_MULTIPLE_SEQUENCE_ALIGNMENT_FILE]
if arg_dict.has_key(ARG_PRINT_E):
print_E = True
if arg_dict.has_key(ARG_PRINT_M):
print_m = True
# Read the alignment file to create a list of parents.
# The parents will appear in the list in the order in which they appear in the file.
parent_list = schema.readMultipleSequenceAlignmentFile(file(msa_file, 'r'))
parents = [p for (k,p) in parent_list]
crossovers = schema.readCrossoverFile(file(arg_dict[ARG_CROSSOVER_FILE], 'r'))
fragments = schema.getFragments(crossovers, parents[0])
# Get the contacts
pdb_contacts = schema.readContactFile(file(arg_dict[ARG_CONTACT_FILE], 'r'))
contacts = schema.getSCHEMAContactsWithCrossovers(pdb_contacts, parents, crossovers)
if arg_dict.has_key(ARG_OUTPUT_FILE):
output_file = file(arg_dict[ARG_OUTPUT_FILE], 'w')
# Now, what does the user want?
output_string = '%s'
output_file.write('# chimera')
if print_E:
output_string += '\t%d'
output_file.write('\tE')
if print_m:
output_string += '\t%d'
output_file.write('\tm')
output_string += '\n'
output_file.write('\n')
if arg_dict.has_key(ARG_CHIMERAS): # Print values for chimeras
chimeras = arg_dict[ARG_CHIMERAS]
# Could be a) a chimera, b) a list of chimeras, or c) a file of chimeras.
if type(chimeras) is list:
# It's a list of chimeras
for chimera_blocks in chimeras:
outputEnergies(chimera_blocks, contacts, fragments, parents, output_file, output_string, print_E, print_m)
elif os.path.isfile(chimeras):
# It's a file of chimeras
for line in file(chimeras,'r').readlines():
chimera_blocks = line.strip()
outputEnergies(chimera_blocks, contacts, fragments, parents, output_file, output_string, print_E, print_m)
else:
# It's a single chimera sequence
chimera_blocks = chimeras
outputEnergies(chimera_blocks, contacts, fragments, parents, output_file, output_string, print_E, print_m)
else:
# Enumerates all possible chimeras and their disruption and mutation values.
p = len(parents)
n = len(fragments)
Es = []
ms = []
for i in xrange(len(parents)**len(fragments)):
# The next two lines turn i into a chimera block pattern
# (e.g., 0 -> '11111111', 1 -> '11111112', 2 -> '11111113'...)
n2c = schema.base(i,p)
chimera_blocks = ''.join(['1']*(n-len(n2c))+['%d'%(int(x)+1,) for x in n2c])
(E, m) = outputEnergies(chimera_blocks, contacts, fragments, parents, output_file, output_string, print_E, print_m)
if (print_E):
Es.append(E)
if (print_m):
ms.append(m)
if (print_E):
mean_str = "# Average disruption <E> = %1.4f\n" % schema.mean(Es)
output_file.write(mean_str)
if (print_m):
mean_str = "# Average mutation <m> = %1.4f\n" % schema.mean(ms)
output_file.write(mean_str)
if arg_dict.has_key(ARG_OUTPUT_FILE):
output_file.close()
def frags(self,parent): return schema.getFragments(self.crossovers(), parent)
def curve(results, parents, bin_width, max_samples=1e10): """Compute a curve of average energy and average mutation, with the latter binned by bin_width. """ if len(results) < 1: # Nothing to do! return (e, xovers, lmin, lmax) = results[0] num_crossovers = len(xovers) #print "# No. of RASPP results:", len(results) # Because the RASPP curve does not involve the length min/max values, we can collapse # the set of RASPP results to only those with unique crossovers. This can # greatly improve performance, since crossover patterns are often duplicated # across min/max values. unique_results = set([(avg_energy, tuple(crossovers)) for (avg_energy, crossovers, l_min, l_max) in results]) #print "# No. of unique RASPP results:", len(unique_results) # Now compute the average mutation levels for these unique libraries. avg_E_ms = [] for (avg_energy, crossovers) in unique_results: crossovers = list(crossovers) fragments = schema.getFragments(crossovers, parents[0]) avg_m = schema.averageMutationSampled(fragments, parents, max_samples) avg_E_ms.append((avg_energy, avg_m, crossovers)) ms = [m for (E, m, crossovers) in avg_E_ms] (min_m, max_m) = (min(ms), max(ms)) num_bins = int((max_m-min_m)/bin_width) + 1 # Assemble the RASPP curve. If num_samples exceeds the library size, # then this curve is approximate. approx_curve = [] for i in range(num_bins): approx_curve.append(None) for (E, m, crossovers) in avg_E_ms: bin_number = int((m - min_m)/bin_width) # If there's an existing value in this bin, check it if approx_curve[bin_number]: (E_old, m_old, crossovers_old) = approx_curve[bin_number] # If lower E in this bin, substitute it if E < E_old: approx_curve[bin_number] = (E, m, crossovers) else: # Otherwise just add it approx_curve[bin_number] = (E, m, crossovers) # It may be that the approximate curve is exact. If so, just return it. library_size = len(parents)**(num_crossovers+1) approximate = (library_size > max_samples) if not approximate: return [r for r in approx_curve if r] # If the curve IS approximate, we'll do a final pass so that # the bin values are correct. Some libraries may still be # incorrectly binned because we've # Compute the exact mutation numbers for the lowest-E libraries final_curve = [] for r in approx_curve: if r: (E, approx_m, crossovers) = r fragments = schema.getFragments(crossovers, parents[0]) true_avg_m = schema.averageMutation(fragments, parents) final_curve.append((E, true_avg_m, crossovers)) #print "%1.2f\t%1.2f" % (true_avg_m, approx_m) return final_curve
def main(args):
arg_dict = parse_arguments(args)
if not confirm_arguments(arg_dict):
if args[0].split(os.path.sep)[-1] == "schemaenergy.py":
print_usage(args)
return
# Flags and values
print_E = False
print_m = False
output_file = sys.stdout
# Inputs:
# The alignment/fragment file name.
msa_file = arg_dict[ARG_MULTIPLE_SEQUENCE_ALIGNMENT_FILE]
if arg_dict.has_key(ARG_PRINT_E):
print_E = True
if arg_dict.has_key(ARG_PRINT_M):
print_m = True
# Read the alignment file to create a list of parents.
# The parents will appear in the list in the order in which they appear in the file.
parent_list = schema.readMultipleSequenceAlignmentFile(file(msa_file, 'r'))
parents = [p for (k, p) in parent_list]
crossovers = schema.readCrossoverFile(
file(arg_dict[ARG_CROSSOVER_FILE], 'r'))
fragments = schema.getFragments(crossovers, parents[0])
# Get the contacts
pdb_contacts = schema.readContactFile(file(arg_dict[ARG_CONTACT_FILE],
'r'))
contacts = schema.getSCHEMAContactsWithCrossovers(pdb_contacts, parents,
crossovers)
if arg_dict.has_key(ARG_OUTPUT_FILE):
output_file = file(arg_dict[ARG_OUTPUT_FILE], 'w')
# Now, what does the user want?
output_string = '%s'
output_file.write('# chimera')
if print_E:
output_string += '\t%d'
output_file.write('\tE')
if print_m:
output_string += '\t%d'
output_file.write('\tm')
output_string += '\n'
output_file.write('\n')
if arg_dict.has_key(ARG_CHIMERAS): # Print values for chimeras
chimeras = arg_dict[ARG_CHIMERAS]
# Could be a) a chimera, b) a list of chimeras, or c) a file of chimeras.
if type(chimeras) is list:
# It's a list of chimeras
for chimera_blocks in chimeras:
outputEnergies(chimera_blocks, contacts, fragments, parents,
output_file, output_string, print_E, print_m)
elif os.path.isfile(chimeras):
# It's a file of chimeras
for line in file(chimeras, 'r').readlines():
chimera_blocks = line.strip()
outputEnergies(chimera_blocks, contacts, fragments, parents,
output_file, output_string, print_E, print_m)
else:
# It's a single chimera sequence
chimera_blocks = chimeras
outputEnergies(chimera_blocks, contacts, fragments, parents,
output_file, output_string, print_E, print_m)
else:
# Enumerates all possible chimeras and their disruption and mutation values.
p = len(parents)
n = len(fragments)
Es = []
ms = []
for i in xrange(len(parents)**len(fragments)):
# The next two lines turn i into a chimera block pattern
# (e.g., 0 -> '11111111', 1 -> '11111112', 2 -> '11111113'...)
n2c = schema.base(i, p)
chimera_blocks = ''.join(['1'] * (n - len(n2c)) +
['%d' % (int(x) + 1, ) for x in n2c])
(E, m) = outputEnergies(chimera_blocks, contacts, fragments,
parents, output_file, output_string,
print_E, print_m)
if (print_E):
Es.append(E)
if (print_m):
ms.append(m)
if (print_E):
mean_str = "# Average disruption <E> = %1.4f\n" % schema.mean(Es)
output_file.write(mean_str)
if (print_m):
mean_str = "# Average mutation <m> = %1.4f\n" % schema.mean(ms)
output_file.write(mean_str)
if arg_dict.has_key(ARG_OUTPUT_FILE):
output_file.close()