def uvrd_cas4_links(G, graph_save_dir):
uvrd_profiles = [
"COG0210", "PRK11773", "TIGR01075", "cl22977", "cl26202", "cl28312",
"pfam00580", "pfam13361", "pfam13538"
]
cas4_profiles = [
"cd09637", "cd09659", "cls000170", "COG1468", "COG4343", "pfam01930",
"pfam06023"
]
all_profiles = uvrd_profiles + cas4_profiles
profile2gene = t.map_cdd_defense2gene_name()
print "Extracting subgraphs"
cas4_g = gt.subgraph(G, cas4_profiles, radius=1)
uvrd_g = gt.subgraph(G, uvrd_profiles, radius=1)
print "Joining subgraphs"
joint_g = nx.compose(cas4_g, uvrd_g)
print "Writing graph to files"
nodes_to_write = set()
with open(os.path.join(graph_save_dir, "edgelist.txt"), "w") as edge_outf:
for edge in joint_g.edges(data=True):
(p1, p2, data) = edge
# if p1 in all_profiles and p2 in all_profiles:
edge_outf.write("%s\t%s\t%f\n" % (p1, p2, data['weight']))
nodes_to_write.update([p1, p2])
with open(os.path.join(graph_save_dir, "nodes.txt"), "w") as node_outf:
for p in nodes_to_write:
if p in cas4_profiles:
_type = 1
elif p in uvrd_profiles:
_type = 2
else:
_type = 0
node_outf.write("%s\t%s\t%d\n" % (p, profile2gene[p], _type))
return joint_g
def uvrd_cas4_links(G, graph_save_dir):
uvrd_profiles = ["COG0210", "PRK11773", "TIGR01075", "cl22977", "cl26202", "cl28312", "pfam00580", "pfam13361", "pfam13538"]
cas4_profiles = ["cd09637", "cd09659", "cls000170", "COG1468", "COG4343", "pfam01930", "pfam06023"]
all_profiles = uvrd_profiles + cas4_profiles
profile2gene = t.map_cdd_defense2gene_name()
print "Extracting subgraphs"
cas4_g = gt.subgraph(G, cas4_profiles, radius=1)
uvrd_g = gt.subgraph(G, uvrd_profiles, radius=1)
print "Joining subgraphs"
joint_g = nx.compose(cas4_g, uvrd_g)
print "Writing graph to files"
nodes_to_write= set()
with open(os.path.join(graph_save_dir, "edgelist.txt"),"w") as edge_outf:
for edge in joint_g.edges(data=True):
(p1, p2, data) = edge
# if p1 in all_profiles and p2 in all_profiles:
edge_outf.write("%s\t%s\t%f\n" % (p1, p2, data['weight']))
nodes_to_write.update([p1, p2])
with open(os.path.join(graph_save_dir, "nodes.txt"),"w") as node_outf:
for p in nodes_to_write:
if p in cas4_profiles:
_type = 1
elif p in uvrd_profiles:
_type = 2
else:
_type = 0
node_outf.write("%s\t%s\t%d\n" % (p, profile2gene[p], _type))
return joint_g
def add_annotation_to_pty():
pty_dir = os.path.join(prok1603_path, "pty")
print "Loading dictionaries"
gi2profiles = map_gi2profiles()
gi2cluster = map_gi2cluster()
profile2gene = t.map_cdd_defense2gene_name()
print "Maps loaded. Starting annotations"
cnt = 1
for d in os.listdir(pty_dir):
# if cnt < 1937:
# cnt += 1
# continue
print cnt, d
for f in glob.glob("%s/*.pty" % (os.path.join(pty_dir, d))):
genes = t.parse_plain_pty_file(f)
if not genes:
continue
for gene in genes:
if gene.gid in gi2profiles:
gene.profiles = gi2profiles[gene.gid]
gene.gene_names = " ".join([profile2gene[p] for p in gene.profiles])
elif gene.gid in gi2cluster:
gene.profiles = gi2cluster[gene.gid]
new_f = f + "2"
t.write_genes_to_pty(genes, new_f)
cnt += 1
def extract_cas_communities(tree_file, node_file, G=None):
module2members = infomap.parse_tree_file(tree_file,
node_names=True,
node_desc_file=node_file)
modules = sorted([(k, v) for (k, v) in module2members.items()],
key=lambda x: len(x[1]),
reverse=True)
cl2name = cas_synonyms()
profile2gene = t.map_cdd_defense2gene_name()
def_profs_file = "/panfs/pan1/patternquest/Projects/NewSystems/data/profiles/defenseProfiles.tab"
cas_profiles = [
l.split()[0] for l in open(def_profs_file) if l.split()[1] == "CRISPR"
]
# uvrd_profiles = ["COG0210", "PRK11773", "TIGR01075", "cl22977", "cl26202", "cl28312", "pfam00580", "pfam13361",
# "pfam13538"]
cas_profiles = cas_profiles + list(cl2name.keys())
save_dir = os.path.join(work_dir, "cases/crispr_graph")
for module in modules:
nodes = set(module[1])
cas_nodes = nodes.intersection(cas_profiles)
if cas_nodes:
# if "crispr" in nodes:
# print "CRISPR"
cas_ratio = len(cas_nodes) / float(len(nodes))
if cas_ratio < 0.3:
continue
print module[0], len(nodes), cas_ratio
continue
graph_file = os.path.join(save_dir,
"%s_%d.net" % (module[0], len(nodes)))
node_file = os.path.join(
save_dir, "%s_%d.nodes.txt" % (module[0], len(nodes)))
subgraph = gt.subgraph_crosslink(G, list(nodes))
gt.write_edgelist(subgraph, graph_file)
with open(node_file, "w") as outf:
outf.write("Node\tgene_name\t")
for node in subgraph:
if node in cl2name:
_name = cl2name[node]
elif node in profile2gene:
_name = profile2gene[node]
else:
_name = node
outf.write("%s\t%s\n" % (node, _name))
break
def extract_extract_cas_communities_from_oslom(G):
oslom_file = "/panfs/pan1/patternquest/Projects/NewSystems/data/prok1603/graph/infomap/merged/consensus_clustering/multi_level/norm_min/3/oslom/tp"
module2pvalue = {}
module2members = {}
with open(oslom_file) as inf:
l = inf.readline()
while l:
assert l.startswith("#")
parts = l.split()
name = parts[1]
size = int(parts[3])
if size == 1:
l = inf.readline()
l = inf.readline()
continue
pvalue = float(parts[5])
module2pvalue[name] = pvalue
module2members[name] = inf.readline().strip().split()
l = inf.readline()
# modules = [l.strip().split() for l in open(oslom_file) if not l.startswith("#")]
node_file = os.path.join(work_dir, "infomap/merged/nodes.txt")
id2node = {l.split()[0]: l.split()[1] for l in open(node_file)}
def_profs_file = "/panfs/pan1/patternquest/Projects/NewSystems/data/profiles/defenseProfiles.tab"
cas_profiles = [
l.split()[0] for l in open(def_profs_file) if l.split()[1] == "CRISPR"
]
cl2name = cas_synonyms()
cas_profiles = cas_profiles + list(cl2name.keys())
save_dir = os.path.join(work_dir, "cases/crispr_graph_2/")
profile2gene = t.map_cdd_defense2gene_name()
for (name, members) in module2members.items():
# nodes = set(module[1])
nodes = set([id2node[id] for id in members])
cas_nodes = nodes.intersection(cas_profiles)
if cas_nodes:
cas_ratio = len(cas_nodes) / float(len(nodes))
if cas_ratio < 0.3:
continue
if "crispr" in nodes:
print "CRISPR"
# print len(nodes), cas_ratio, " ".join(nodes)
# print len(nodes), cas_ratio, Counter([profile2gene[n] for n in cas_nodes])
vals = [
name,
str(module2pvalue[name]),
str(len(nodes)),
str(cas_ratio), " ".join(nodes)
]
# print "\t".join(vals)
graph_file = os.path.join(save_dir,
"%s_%d.net" % (name, len(nodes)))
node_file_s = os.path.join(save_dir,
"%s_%d.nodes.txt" % (name, len(nodes)))
subgraph = gt.subgraph_crosslink(G, list(nodes))
gt.write_edgelist(subgraph, graph_file)
with open(node_file_s, "w") as outf:
outf.write("Node\tgene_name\tcas\n")
for node in subgraph:
if node in cl2name:
_name = cl2name[node]
elif node in profile2gene:
_name = profile2gene[node]
else:
_name = node
_type = 0
if node in cas_nodes:
_type = 1
outf.write("%s\t%s\t%d\n" % (node, _name, _type))
os.path.join(os.path.expanduser('~'), 'Projects/SystemFiles/'))
elif sys.platform == 'linux2':
sys.path.append(
os.path.join(os.path.expanduser('~'), 'Projects/lib/BioPy/'))
sys.path.append(
os.path.join(os.path.expanduser('~'), 'Projects/SystemFiles/'))
import global_variables as gv
sys.path.append(gv.project_code_path)
from lib.utils import tools as t
import glob
root_path = '/home/hudaiber/data/Prok1603/pty/'
print "Loading profile2genes"
profile2gene = t.map_cdd_defense2gene_name()
print "Loading ccp"
map_file = '/panfs/pan1.be-md.ncbi.nlm.nih.gov/prokdata/db_tmp/Prok1603/Annotation/Prok1603.ccp.csv'
gi2cdd = t.map_gid2cdd(map_file)
print "Loading CDD"
cdd2gene = t.map_cdd_defense2gene_name()
cnt = 0
cas4 = 0
cas4_uvrd = 0
uvrd = 0
for dir in os.listdir(root_path):