def build_poa_affinity_tree(p: graph.Poagraph,
blosum: Optional[parameters.Blosum],
output_dir: Path,
hbmin: parameters.Hbmin,
verbose: bool) -> tree.AffinityTree:
"""Builds Affinity Tree coherent with poa software.
This method builds a simple version of Affinity Tree
as it uses a single call to poa software. Poa provides
division of sequences in Poagraph into consistent groups
with a consensus path assigned to each group. These groups
are converted in this method to Affinity Tree nodes and
connected with a dummy root node so the result is coherent
with pangtree definition of Affinity Tree.
Args:
p: Poagraph containing sequences to be divided into
groups (Affinity Tree nodes).
optional blosum: BLOSUM matrix. If not provided, default Blosum80.mat is used.
output_dir: Path to a directory that can be used by poa software.
hbmin: Parameter required by poa software. The minimum value of
sequence compatibility to generated consensus.
verbose: Switch to control logging intensity.
Raises:
AffinityTreeGenerationException: if consensuses cannot be found.
"""
def _convert_consensus_paths_to_affinity_tree_nodes():
at_nodes = []
assigned_sequences = []
for c_id, c_info in consensus_paths.items():
assigned_sequences += c_info.assigned_sequences_ids
all_seq = p.get_sequences_ids()
compatibilities = p.get_compatibilities(all_seq, c_info.path)
if len(c_info.assigned_sequences_ids):
assigned_seq_comp = [c
for seq_id, c in compatibilities.items()
if seq_id in c_info.assigned_sequences_ids]
mincomp = min(assigned_seq_comp)
else:
mincomp = 0
new_node = tree.AffinityNode(id_=tree.AffinityNodeID(c_id + 1),
parent=tree.AffinityNodeID(0),
sequences=c_info.assigned_sequences_ids,
mincomp=mincomp,
compatibilities=compatibilities,
consensus=c_info.path,
children=[])
at_nodes.append(new_node)
node_for_unassigned_sequences = tree.AffinityNode(parent=tree.AffinityNodeID(0),
sequences=[seq_id
for seq_id in p.get_sequences_ids()
if seq_id not in assigned_sequences],
id_=tree.AffinityNodeID(len(at_nodes) + 1),
mincomp=graph.Compatibility(0),
children=[])
at_nodes.append(node_for_unassigned_sequences)
return at_nodes
global_logger.info("POA defined affinity tree generation started.")
if blosum is None:
blosum = get_default_blosum()
_raise_error_if_invalid_poagraph(p)
try:
consensus_paths = poa.get_consensuses(p,
p.get_sequences_ids(),
output_dir,
"poa_tree",
blosum.filepath,
hbmin)
except poa.NoConsensusError:
raise AffinityTreeBuildException("No consensus in the Affinity Tree.")
consensus_nodes = _convert_consensus_paths_to_affinity_tree_nodes()
root_node = tree.AffinityNode(id_=tree.AffinityNodeID(0),
children=[c_node.id_
for c_node in consensus_nodes])
affinity_tree = tree.AffinityTree([root_node] + consensus_nodes)
global_logger.info("POA defined affinity tree generation finished.")
return affinity_tree
def main():
parser = cli.get_parser()
args = parser.parse_args()
start = datetime.datetime.now()
if not args.quiet and args.verbose:
logprocess.add_file_handler_to_logger(args.output_dir,
"details",
"details.log",
propagate=False)
logprocess.add_file_handler_to_logger(args.output_dir,
"",
"details.log",
propagate=False)
if args.quiet:
logprocess.disable_all_loggers()
poagraph, dagmaf, fasta_provider = None, None, None
if isinstance(args.multialignment, Maf) and args.raw_maf:
poagraph = Poagraph.build_from_maf(args.multialignment, args.metadata)
elif isinstance(args.multialignment, Maf) and not args.raw_maf:
fasta_provider = cli.resolve_fasta_provider(args)
poagraph, dagmaf = Poagraph.build_from_dagmaf(args.multialignment,
fasta_provider,
args.metadata)
elif isinstance(args.multialignment, Po):
poagraph = Poagraph.build_from_po(args.multialignment, args.metadata)
consensus_tree = None
if args.consensus is not None:
blosum = args.blosum if args.blosum else cli.get_default_blosum()
if fasta_provider is not None and isinstance(fasta_provider,
ConstSymbolProvider):
blosum.check_if_symbol_is_present(
fasta_provider.missing_symbol.as_str())
consensus_output_dir = pathtools.get_child_dir(args.output_dir,
"consensus")
if args.consensus == 'poa':
consensus_tree = simple_tree_generator.get_simple_consensus_tree(
poagraph, blosum, consensus_output_dir, args.hbmin,
args.verbose)
elif args.consensus == 'tree':
max_strategy = cli.resolve_max_strategy(args)
node_strategy = cli.resolve_node_strategy(args)
consensus_tree = tree_generator.get_consensus_tree(
poagraph, blosum, consensus_output_dir, args.stop, args.p,
max_strategy, node_strategy, args.verbose)
try:
seq_id_to_name = {
seq_id: seq.seqmetadata["name"]
for seq_id, seq in poagraph.sequences.items()
}
except:
seq_id_to_name = None
newick_consensus_tree = consensus_tree.as_newick(seq_id_to_name)
pathtools.save_to_file(
newick_consensus_tree,
pathtools.get_child_path(args.output_dir, "consensus_tree.newick"))
if args.output_po:
pangenome_po = poagraph_to_PangenomePO(poagraph)
pathtools.save_to_file(
pangenome_po,
pathtools.get_child_path(args.output_dir, "poagraph.po"))
if args.output_fasta:
sequences_fasta = poagraph_to_fasta(poagraph)
pathtools.save_to_file(
sequences_fasta,
pathtools.get_child_path(args.output_dir, "sequences.fasta"))
if consensus_tree:
consensuses_fasta = consensuses_tree_to_fasta(
poagraph, consensus_tree)
pathtools.save_to_file(
consensuses_fasta,
pathtools.get_child_path(args.output_dir, "consensuses.fasta"))
end = datetime.datetime.now()
pangenomejson = to_PangenomeJSON(task_parameters=cli.get_task_parameters(
args, running_time=f"{end-start}s"),
poagraph=poagraph,
dagmaf=dagmaf,
consensuses_tree=consensus_tree)
pangenome_json_str = to_json(pangenomejson)
pathtools.save_to_file(
pangenome_json_str,
pathtools.get_child_path(args.output_dir, "pangenome.json"))
def build_affinity_tree(poagraph: graph.Poagraph,
blosum: Optional[parameters.Blosum],
output_dir: Path,
stop: parameters.Stop,
p: parameters.P,
verbose: bool) -> tree.AffinityTree:
"""Builds Affinity Tree.
Affinity Tree is defined in paper 'Getting insight into the
pan-genome structure with Pangtree'. This method builds
an Affinity Tree by iterative calls to poa software.
Full algorithm and idea are described in the above-mentioned paper.
Args:
poagraph: Poagraph containing _sequences to be divided into groups
(Affinity Tree nodes).
optional blosum: BLOSUM matrix. If not provided, default Blosum80.mat is used.
output_dir: Path to a directory that can be used by poa software.
stop: Value of mincomp above which an affinity tree node is no more
split.
p: Value changing the linear meaning of compatibility when searching
for cutoff.
verbose: Switch to control logging intensity.
Raises:
AffinityTreeGenerationException: if consensuses cannot be found.
Returns:
Affinity Tree generated with Pangtree algorithm.
"""
global_logger.info("Affinity Tree generation started.")
if blosum is None:
blosum = get_default_blosum()
if verbose:
logprocess.add_file_handler_to_logger(output_dir,
"tresholdsCSV",
"tresholds.csv",
"%(message)s", False)
_raise_error_if_invalid_poagraph(poagraph)
root_node = _get_root_node(poagraph, blosum.filepath, output_dir, p)
affinity_tree = tree.AffinityTree([root_node])
nodes_to_process = deque([affinity_tree.get_node(tree.AffinityNodeID(0))])
while nodes_to_process:
node = nodes_to_process.pop()
children_nodes = _get_children_nodes_looping(node,
poagraph,
output_dir,
blosum.filepath,
p,
affinity_tree.get_max_node_id())
if len(children_nodes) == 1:
continue
for child in children_nodes:
all_sequences = [*poagraph.sequences.keys()]
child.compatibilities = poagraph.get_compatibilities(sequences_ids=all_sequences,
consensus_path=child.consensus,
p=p)
node.children.append(child.id_)
affinity_tree.nodes.append(child)
if not _node_is_ready(child, stop):
nodes_to_process.append(child)
global_logger.info("Affinity Tree generation finished.\n")
return affinity_tree
def main():
parser = cli.get_parser()
args = parser.parse_args()
start = datetime.datetime.now()
if not args.quiet and args.verbose:
logprocess.add_file_handler_to_logger(args.output_dir,
"details",
"details.log",
propagate=False)
logprocess.add_file_handler_to_logger(args.output_dir,
"",
"details.log",
propagate=False)
if args.quiet:
logprocess.disable_all_loggers()
poagraph, dagmaf, fasta_provider = None, None, None
if isinstance(args.multialignment, msa.Maf) and args.raw_maf:
poagraph = builder.build_from_maf(args.multialignment, args.metadata)
elif isinstance(args.multialignment, msa.Maf) and not args.raw_maf:
fasta_provider = cli.resolve_fasta_provider(args)
poagraph, dagmaf = builder.build_from_dagmaf(args.multialignment,
fasta_provider,
args.metadata)
elif isinstance(args.multialignment, msa.Po):
poagraph = builder.build_from_po(args.multialignment, args.metadata)
affinity_tree = None
if args.affinity is not None:
blosum = args.blosum if args.blosum else cli.get_default_blosum()
if fasta_provider is not None and isinstance(
fasta_provider, missings.ConstBaseProvider):
blosum.check_if_symbol_is_present(
fasta_provider.missing_base.as_str())
consensus_output_dir = pathtools.get_child_dir(args.output_dir,
"affinitytree")
if args.affinity == 'poa':
affinity_tree = at_builders.build_poa_affinity_tree(
poagraph, blosum, consensus_output_dir, args.hbmin,
args.verbose)
elif args.affinity == 'tree':
affinity_tree = at_builders.build_affinity_tree(
poagraph, blosum, consensus_output_dir, args.stop, args.p,
args.verbose)
if args.metadata is not None:
seq_id_to_metadata = {
seq_id: seq.seqmetadata
for seq_id, seq in poagraph.sequences.items()
}
else:
seq_id_to_metadata = None
affinity_tree_newick = affinity_tree.as_newick(seq_id_to_metadata,
separate_leaves=True)
pathtools.save_to_file(
affinity_tree_newick,
pathtools.get_child_path(consensus_output_dir,
"affinity_tree.newick"))
if args.output_po:
pangenome_po = po.poagraph_to_PangenomePO(poagraph)
pathtools.save_to_file(
pangenome_po,
pathtools.get_child_path(args.output_dir, "poagraph.po"))
if args.output_fasta:
sequences_fasta = fasta.poagraph_to_fasta(poagraph)
pathtools.save_to_file(
sequences_fasta,
pathtools.get_child_path(args.output_dir, "_sequences.fasta"))
if affinity_tree:
consensuses_fasta = fasta.affinity_tree_to_fasta(
poagraph, affinity_tree)
pathtools.save_to_file(
consensuses_fasta,
pathtools.get_child_path(args.output_dir,
"affinitytree.fasta"))
end = datetime.datetime.now()
pangenomejson = json.to_PangenomeJSON(
task_parameters=cli.get_task_parameters(args,
running_time=f"{end-start}s"),
poagraph=poagraph,
dagmaf=dagmaf,
affinity_tree=affinity_tree)
pangenome_json_str = json.to_json(pangenomejson)
pathtools.save_to_file(
pangenome_json_str,
pathtools.get_child_path(args.output_dir, "pangenome.json"))