def split_alt_nodes(graph: GraphContainer, max_len=300, padding_len=150):
"""
Split long alternate nodes
:param graph: graph to work on
:param max_len: max length of reference node with no sequences
:param padding_len: length of sequence to keep
"""
assert max_len >= 2 * padding_len
for node in list(graph.altNodes()):
if len(node["sequence"]) <= max_len:
continue
logging.info(f"Splitting long ALT node: {node['name']}")
n1 = graph.add_altNode(node["chrom"], node["start"], node["end"],
node["sequence"][:padding_len],
node["sequences"])
n2 = graph.add_altNode(node["chrom"], node["start"], node["end"],
node["sequence"][-padding_len:],
node["sequences"])
for e in list(graph.inEdges(node)):
graph.add_edge(graph.nodes[e["from"]], n1, e["sequences"])
for e in list(graph.outEdges(node)):
graph.add_edge(n2, graph.nodes[e["to"]], e["sequences"])
graph.del_node(node)
def add_graph(graph1: GraphContainer, graph2: GraphContainer):
"""
Add all nodes, edges and paths from graph2 to graph1 (inplace)
"""
for node in graph2.refNodes():
graph1.add_refNode(
node["chrom"], node["start"], node["end"], node["sequences"])
for node in graph2.altNodes():
graph1.add_altNode(
node["chrom"], node["start"], node["end"], node["sequence"], node["sequences"])
for edge in graph2.edges.values():
graph1.add_edge(graph1.nodes[edge["from"]], graph1.nodes[edge["to"]], edge["sequences"])
graph1.paths += graph2.paths
def get_graph(self, allele_graph=False):
""" Create the paragraph representation of nodes and edges for this graph
:param alleleGraph: create edges between any compatible allele pair (rather
than just following reference and given haplotypes)
:return GraphContainer object
"""
logging.info("Creating output graph")
graph = GraphContainer()
# create ref nodes
pnode = None
for ref in self.get_ref_alleles():
node = graph.add_refNode(self.chrom, ref.begin, ref.end - 1,
ref.data.haplotypes)
if pnode:
assert pnode["end"] + 1 == node["start"]
graph.add_edge(pnode, node)
pnode = node
# Create alt nodes
for alt in self.alts.values():
graph.add_altNode(self.chrom, alt.start, alt.end, alt.sequence,
alt.haplotypes)
# Create edges connecting nodes along a haplotype (or allele in alleleGraph mode)
for haplo in self.get_haplotypes():
nodes = graph.nodes_by_haplo(haplo)
logging.info(
f"Linking nodes in sequence {haplo}:\t{', '.join(n['name'] for n in nodes)}"
)
pnode = None
for node in nodes:
if pnode:
if pnode["end"] == node["start"] - 1:
graph.add_edge(pnode, node, [haplo])
pnode_is_ref_dummy = pnode[
"end"] == pnode["start"] - 1 and not pnode["sequence"]
pnode_ends_before_node = pnode["end"] < node[
"start"] and pnode["start"] < node["start"]
if not pnode_is_ref_dummy and not pnode_ends_before_node:
raise Exception(
f"Inconsistent nodes for haplotype {haplo}: {pnode['name']}, {node['name']}"
)
pnode = node
# In alleleGraph mode link each alt node to all neighboring nodes
# In haplotype mode link nodes without in/out edges to reference
for node in graph.altNodes():
if allele_graph or not any(graph.inEdges(node)):
graph.add_edge(
graph.refNode_ending_at[node["chrom"], node["start"] - 1],
node)
if not any(graph.outEdges(node)):
graph.add_edge(
node, graph.refNode_starting_at[node["chrom"],
node["end"] + 1])
if allele_graph:
isInsertion = node["end"] < node["start"]
for n in graph.nodes_starting_at[node["end"] + 1]:
# Don't loop by connecting multiple insertions at the same position
if not (isInsertion and n["end"] < n["start"]):
graph.add_edge(node, n)
# For nodes that do not have determined in/out edges for a given haplotype
# label all in/out edges as compatible with that haplotype
# excluding edges that connect to another allele at the same vcfVariant (e.g. insertions)
for haplo in self.get_haplotypes():
for node in graph.nodes_by_haplo(haplo):
if not any(graph.inEdges(node, haplo)):
for e in graph.inEdges(node):
graph.add_edge(graph.nodes[e["from"]], node, [haplo])
assert any(graph.inEdges(node, haplo))
if not any(graph.outEdges(node, haplo)):
for e in graph.outEdges(node):
graph.add_edge(node, graph.nodes[e["to"]], [haplo])
return graph