def setline(self, irandom=None): # don't access <self.line> directly
if self.line is not None:
return self.line
line = {}
for region in utils.regions:
line[region + '_gene'] = self.genes[region]
for boundary in utils.boundaries:
line[boundary + '_insertion'] = self.insertions[boundary]
for boundary in utils.effective_boundaries:
line[boundary + '_insertion'] = '' # NOTE 'fv' and 'jf' insertions are hereby hardcoded to zero (I'm just writing this here to make it easily searchable -- I don't remember why it's set up that way)
for erosion in utils.real_erosions:
line[erosion + '_del'] = self.erosions[erosion]
for erosion in utils.effective_erosions:
line[erosion + '_del'] = self.effective_erosions[erosion]
line['input_seqs'] = self.final_seqs
line['indelfos'] = self.indelfos
line['seqs'] = [self.indelfos[iseq]['reversed_seq'] if indelutils.has_indels(self.indelfos[iseq]) else line['input_seqs'][iseq] for iseq in range(len(line['input_seqs']))]
self.set_ids(line, irandom=irandom)
treeutils.translate_labels(self.tree, zip(self.leaf_names, line['unique_ids'])) # ordering in <self.leaf_names> is set in recombinator.add_mutants()
line['affinities'] = [None for _ in line['unique_ids']]
line['tree'] = self.tree.as_string(schema='newick')
utils.add_implicit_info(self.glfo, line)
self.line = line
def read_input_tree_file(self, outfname):
if self.args.debug:
print ' reading trees from %s' % self.args.input_simulation_treefname
utils.simplerun('cp %s %s' %
(self.args.input_simulation_treefname, outfname),
debug=False)
ages, treestrs = [], []
with open(outfname) as treefile:
for line in treefile:
tstr = line.strip()
if tstr == '': # skip empty lines
continue
dtree = treeutils.get_dendro_tree(
treestr=tstr, suppress_internal_node_taxa=True)
if dtree.seed_node.edge_length is None: # make sure root edge length is set (otherwise bppseqgen barfs)
dtree.seed_node.edge_length = 0.
old_new_label_pairs = [
(l.taxon.label, 't%d' % (i + 1))
for i, l in enumerate(dtree.leaf_node_iter())
]
treeutils.translate_labels(
dtree, old_new_label_pairs
) # rename the leaves to t1, t2, etc. (it would be nice to not have to do this, but a bunch of stuff in recombinator uses this to check that e.g. bppseqgen didn't screw up the ordering)
age = self.choose_full_sequence_branch_length()
if self.args.debug > 1: # it's easier to keep this debug line separate up here than make a tmp variable to keep track of the old height
print ' input tree %d (rescaled depth %.3f --> %.3f):' % (
len(ages), treeutils.get_mean_leaf_height(tree=dtree),
age)
treeutils.rescale_tree(
age, dtree=dtree
) # I think this gets rescaled again for each event, so we could probably in principle avoid this rescaling, but if the input depth is greater than one stuff starts breaking, so may as well do it now
ages.append(age)
treestrs.append(dtree.as_string(schema='newick').strip())
if self.args.debug > 1:
print utils.pad_lines(treeutils.get_ascii_tree(dtree))
if any(a > 1. for a in ages):
raise Exception(
'tree depths must be less than 1., but trees read from %s don\'t satisfy this: %s'
% (self.args.input_simulation_treefname, ages))
if len(ages) != self.args.n_trees:
print ' resetting --n-trees from %d to %d to match trees read from %s' % (
self.args.n_trees, len(ages),
self.args.input_simulation_treefname)
self.args.n_trees = len(ages)
return ages, treestrs
def get_mature_line(sfos,
naive_line,
glfo,
nodefo,
dtree,
target_sfos,
locus=None):
assert len(
naive_line['unique_ids']
) == 1 # enforces that we ran naive-only, 1-leaf partis simulation above
assert not indelutils.has_indels(
naive_line['indelfos'][0]) # would have to handle this below
if args.debug:
utils.print_reco_event(naive_line)
reco_info = collections.OrderedDict()
for sfo in sfos:
mline = utils.get_non_implicit_copy(naive_line)
del mline['tree']
mline['unique_ids'] = [sfo['name']]
mline['seqs'] = [sfo['seq']]
mline['input_seqs'] = [
sfo['seq']
] # it's really important to set both the seqs (since they're both already in there from the naive line)
mline['duplicates'] = [[]]
reco_info[sfo['name']] = mline
try:
utils.add_implicit_info(glfo, mline)
except: # TODO not sure if I really want to leave this in long term, but it shouldn't hurt anything (it's crashing on unequal naive/mature sequence lengths, and I need this to track down which event it is) UPDATE: yeah it was just because something crashed in the middle of writing a .fa file
print 'implicit info adding failed for ievent %d in %s' % (
ievent, outdir)
lines = traceback.format_exception(*sys.exc_info())
print utils.pad_lines(
''.join(lines)
) # NOTE this will still crash on the next line if implicit info adding failed
final_line = utils.synthesize_multi_seq_line_from_reco_info(
[sfo['name'] for sfo in sfos], reco_info)
ftree = copy.deepcopy(dtree)
if locus is not None:
def ltr(u):
return u + '-' + locus
new_nodefo = {}
for u_old in nodefo:
new_nodefo[ltr(u_old)] = nodefo[u_old]
nodefo = new_nodefo
treeutils.translate_labels(ftree,
[(u, ltr(u))
for u in final_line['unique_ids']])
final_line['unique_ids'] = [
ltr(u) for u in final_line['unique_ids']
]
assert len(sfos) == len(final_line['unique_ids'])
for iseq, sfo in enumerate(sfos):
naive_id = naive_line['unique_ids'][0]
assert naive_id.count('-') == 1
bstr = naive_id.replace('-' + locus, '')
pids = final_line['paired-uids'][iseq]
assert len(pids) == 1 and pids[0].find(
bstr
) == 0 and pids[0].count('-') == 1 and pids[0].split(
'-'
)[1] in utils.loci # if uid is xxx-igh, paired id shoud be e.g. xxx-igk
final_line['paired-uids'][iseq] = [
p.replace(bstr, sfo['name']) for p in pids
]
if args.debug:
utils.print_reco_event(final_line)
# extract kd values from pickle file (use a separate script since it requires ete/anaconda to read)
if len(
set(nodefo) - set(final_line['unique_ids'])
) > 0: # uids in the kd file but not the <line> (i.e. not in the newick/fasta files) are probably just bcr-phylo discarding internal nodes
print ' in kd file, but missing from final_line (probably just internal nodes that bcr-phylo wrote to the tree without names): %s' % (
set(nodefo) - set(final_line['unique_ids']))
if len(set(final_line['unique_ids']) - set(nodefo)) > 0:
print ' in final_line, but missing from kdvals: %s' % ' '.join(
set(final_line['unique_ids']) - set(nodefo))
final_line['affinities'] = [
1. / nodefo[u]['kd'] for u in final_line['unique_ids']
]
final_line['relative_affinities'] = [
1. / nodefo[u]['relative_kd'] for u in final_line['unique_ids']
]
final_line['lambdas'] = [
nodefo[u]['lambda'] for u in final_line['unique_ids']
]
final_line['nearest_target_indices'] = [
nodefo[u]['target_index'] for u in final_line['unique_ids']
]
ftree.scale_edges(1. / numpy.mean([len(s)
for s in final_line['seqs']]))
if args.debug:
print utils.pad_lines(treeutils.get_ascii_tree(dendro_tree=ftree),
padwidth=12)
final_line['tree'] = ftree.as_string(schema='newick')
tmp_event = RecombinationEvent(
glfo
) # I don't want to move the function out of event.py right now
tmp_event.set_reco_id(
final_line, irandom=ievent
) # not sure that setting <irandom> here actually does anything
final_line['target_seqs'] = [tfo['seq'] for tfo in target_sfos]
return final_line