def intersect_fusions_by_breakpoints():
lines = [line for line in open(cff, "r")]
fusion = pygeneann.CffFusion(lines[0])
header = fusion.zone1_attrs + fusion.zone2_attrs + fusion.zone3_attrs + fusion.zone4_attrs
df_cff = pd.read_csv(cff,
sep='\t',
keep_default_na=False,
index_col=False,
names=header)
#create BedTools object with appropriate column names
print >> sys.stderr, "create BedTools object with appropriate column names"
df_bed = df_cff[[
'chr1', 'pos1', 'pos1', 'chr2', 'pos2', 'pos2', 'fusion_id'
]]
df_bed.columns = [
'chr1', 'pos1', 'pos1_2', 'chr2', 'pos2', 'pos2_2', 'fusion_id'
]
df_bed.loc[:, ['pos1_2', 'pos2_2']] += 1
df_bed = bedtools.BedTool.from_dataframe(df_bed)
#Intersect fusions: NOTE: only keeps fusions that intersect
#print >> sys.stderr, "Intersect fusions: NOTE: only keeps fusions that intersect"
#df_intersect=df_bed.pair_to_pair(df_bed, slop=100, rdn=True)
print >> sys.stderr, "Intersect fusions: NOTE: rdn=False, keeps self-intersections"
df_intersect = df_bed.pair_to_pair(df_bed, slop=100, rdn=False)
df = df_intersect.to_dataframe(header=None).iloc[:, 0:14]
df.columns = [
'chr1', 'pos1', 'pos1_2', 'chr2', 'pos2', 'pos2_2', 'fusion_id',
'chr1_1', 'pos1_1', 'pos1_2_1', 'chr2_1', 'pos2_1', 'pos2_2_1',
'fusion_id_lst'
]
df = df[['fusion_id', 'fusion_id_lst']]
#write paired F_IDs to tsv
return df
def intersect_fusions_by_genes(cff_file):
fusion_dict = {}
fusion_list_for_bp_cmp = []
common_key_dict = {}
# cluster fusions by gene pairs, save in fusion_dict
for line in open(cff_file, "r"):
if line.startswith("#"):
continue
fusion = pygeneann.CffFusion(line)
if fusion.t_gene1 == "NA" or fusion.t_gene2 == "NA":
continue
else:
key = ",".join(
sorted([
fusion.t_gene1 + "|" + fusion.chr1,
fusion.t_gene2 + "|" + fusion.chr2
]))
fusion_dict.setdefault(key, []).append(fusion.fusion_id)
return fusion_dict
def cluster_fusions_by_genes(cff_file):
fusion_dict = {}
fusion_list_for_bp_cmp = []
common_key_dict = {}
# cluster fusions by gene pairs, save in fusion_dict
for line in open(cff_file, "r"):
if line.startswith("#"):
continue
fusion = pygeneann.CffFusion(line)
if fusion.t_gene1 == "NA" or fusion.t_gene2 == "NA":
continue
else:
key = ",".join(
sorted([
fusion.t_gene1 + "|" + fusion.chr1,
fusion.t_gene2 + "|" + fusion.chr2
]))
fusion_dict.setdefault(key, []).append(fusion.fusion_id)
# output clustered fusions
#for key in fusion_dict:
# fusion_list = fusion_dict[key]
#self.output_clustered_fusions(fusion_list, "Gene_Cluster")
return fusion_dict
parser.add_argument('ref_fa', action='store', help='Reference genome file')
args = parser.parse_args()
gene_ann = pygeneann.GeneAnnotation(args.ensbed)
ref = pysam.FastaFile(args.ref_fa)
def remove_underscores(gene):
return gene.replace("_", ".")
seq_dict = {}
for line in open(args.cff_file, "r"):
fusion = pygeneann.CffFusion(line)
# in a downstream script, "_" is used as a field separator. Need to remove "_" from gene names
gene1 = remove_underscores(fusion.reann_gene1)
gene2 = remove_underscores(fusion.reann_gene2)
lib = fusion.library
fusion_id = fusion.fusion_id
#print fusion.tostring()
#fusion.check_codon(gene_ann, ref_fa)
head_seqs = pygeneann.build_transcript_and_fusion_seq(
gene_ann, fusion, ref, fusion.pos1, "head")
#print(set(head_seqs))
#continue
tail_seqs = pygeneann.build_transcript_and_fusion_seq(
list(set(gene2_list))), max_split_cnt, max_span_cnt,
",".join(list(set(sample_type_list))), ",".join(
list(set(disease_list))), ",".join(list(set(tool_list))),
",".join(list(set(category_list))), gene1_on_bndry,
gene1_close_to_bndry, gene2_on_bndry, gene2_close_to_bndry,
dna_supp_cluster_num, ",".join(list(set(sample_list))), ",".join(
list(set(chr1_list))), ",".join(list(set(breakpoint_1_list))),
",".join(list(set(chr2_list))), ",".join(
list(set(breakpoint_2_list))), captured_reads_tumor_mean,
captured_reads_normal_mean, ",".join(list(set(fusion_IDs)))
]))
# Load cff file
lines = [line for line in open(cff, "r")]
fusion = pygeneann.CffFusion(lines[0])
header = fusion.zone1_attrs + fusion.zone2_attrs + fusion.zone3_attrs + fusion.zone4_attrs
df_cff = pd.read_csv(cff,
sep='\t',
keep_default_na=False,
index_col=False,
names=header)
# load FIDs file
FID_clusters = [line for line in open(FIDs, "r")]
#FID_clusters = FID_clusters[1]
for cluster in FID_clusters:
if cluster.startswith('FIDs'): continue
FID_lst = cluster.rstrip().split(",")
#print(FID_lst)
#print(df_cff[df_cff['fusion_id'].isin(FID_lst)][['t_gene1', 't_gene2', 'pos1', 'pos2']])