def __init__(self, target, splice_graph, ID, cutoff, upstream=None, downstream=None):
'''
As the purpose of ExonSeek is to flanking constitutive exons, it is
necessary to know what needs to be "flanked", the target, and have a
splice graph representation of gene structure (splice_graph). The ID
variable is meant to prevent overwriting of files.
'''
self.id = ID # id is to prevent overwriting files in self.save_path_info
self.cutoff = cutoff
self.target = target # (start, end)
self.upstream, self.downstream = upstream, downstream # if set, the user specified upstream or downstream exons
self.graph = splice_graph.get_graph() # convenience variable (could just use splice_graph)
if self.target not in self.graph.nodes():
raise utils.PrimerSeqError('Error: The target was not found in the splice graph')
self.strand = splice_graph.strand # convenience variable
self.splice_graph = splice_graph
biconnected_comp = filter(lambda x: target in x, algs.get_biconnected(self.graph))
self.total_components = None # these will be defined after calling methods
self.psi_upstream, self.psi_target, self.psi_downstream = None, None, None # these will be defined after calling methods
self.all_paths = None
self.num_of_biconnected = len(biconnected_comp)
if len(self.graph.predecessors(self.target)) == 0 or len(self.graph.successors(self.target)) == 0:
self.component = None # no flanking exon case
elif self.num_of_biconnected == 0:
self.no_biconnected_case()
elif self.num_of_biconnected == 1:
self.component = sorted(biconnected_comp[0], key=lambda x: (x[0], x[1])) # make sure component is sorted by position
self.one_biconnected_case()
elif self.num_of_biconnected == 2:
self.component = map(lambda x: sorted(x, key=lambda y: (y[0], y[1])), biconnected_comp)
self.two_biconnected_case()
else:
raise ValueError('Error: There to be either 0, 1, or 2 biconnected components. Received %s' % self.num_of_biconnected)
def get_flanking_biconnected_exons(name, target, sGraph, genome):
'''
Defines flanking exons as exons that cannot be skipped in
the graph structure. Theese exons are 100% included and do not
need estimation of inclusion level.
'''
graph = sGraph.get_graph() # nx.DiGraph
# search through each biconnected component
for component in algs.get_biconnected(graph):
component = sorted(
component, key=lambda x:
(x[0], x[1])) # ensure first component is first exon, etc
if target in component[1:-1]:
# define upstream/downstream flanking exon
if sGraph.strand == '+':
upstream = component[0]
downstream = component[-1]
else:
upstream = component[-1]
downstream = component[0]
# get possible lengths
all_paths = algs.AllPaths(sGraph,
component,
target,
chr=sGraph.chr,
strand=sGraph.strand)
# all_paths.set_all_path_lengths() # should no longer need this since it is done in primer.py
all_paths.set_all_path_coordinates()
# get sequence of upstream/target/downstream combo
genome_chr = genome[sGraph.chr] # chr object from pygr
upstream_seq, target_seq, downstream_seq = genome_chr[
upstream[0]:upstream[1]], genome_chr[
target[0]:target[1]], genome_chr[
downstream[0]:downstream[1]]
if sGraph.strand == '-':
upstream_seq, target_seq, downstream_seq = \
-upstream_seq, -target_seq, -downstream_seq
return [
sGraph.strand, name[1:], 'NA',
sGraph.chr + ':' + '-'.join(map(str, upstream)), '1.0',
sGraph.chr + ':' + '-'.join(map(str, downstream)), '1.0',
all_paths,
str(upstream_seq).upper(),
str(target_seq).upper(),
str(downstream_seq).upper()
]
return ['Error: ' + name + ' was not found in a biconnected component']
def get_flanking_biconnected_exons(name, target, sGraph, genome):
'''
Defines flanking exons as exons that cannot be skipped in
the graph structure. Theese exons are 100% included and do not
need estimation of inclusion level.
'''
graph = sGraph.get_graph() # nx.DiGraph
# search through each biconnected component
for component in algs.get_biconnected(graph):
component = sorted(component, key=lambda x: (x[0], x[1])) # ensure first component is first exon, etc
if target in component[1:-1]:
# define upstream/downstream flanking exon
if sGraph.strand == '+':
upstream = component[0]
downstream = component[-1]
else:
upstream = component[-1]
downstream = component[0]
# get possible lengths
all_paths = algs.AllPaths(sGraph, component, target,
chr=sGraph.chr, strand=sGraph.strand)
# all_paths.set_all_path_lengths() # should no longer need this since it is done in primer.py
all_paths.set_all_path_coordinates()
# get sequence of upstream/target/downstream combo
genome_chr = genome[sGraph.chr] # chr object from pygr
upstream_seq, target_seq, downstream_seq = genome_chr[upstream[0]:upstream[1]], genome_chr[target[0]:target[1]], genome_chr[downstream[0]:downstream[1]]
if sGraph.strand == '-':
upstream_seq, target_seq, downstream_seq = \
-upstream_seq, -target_seq, -downstream_seq
return [sGraph.strand, name[1:], 'NA',
sGraph.chr + ':' + '-'.join(map(str, upstream)), '1.0',
sGraph.chr + ':' + '-'.join(map(str, downstream)), '1.0',
all_paths, str(upstream_seq).upper(),
str(target_seq).upper(), str(downstream_seq).upper()]
return ['Error: ' + name + ' was not found in a biconnected component']