def get_gene_overlap( chr, pos, ori, bp ):
SERVER='http://grch37.rest.ensembl.org'
SPECIES="human"
ENDPOINT="/overlap/region/"+SPECIES+"/"+str(chr)+":"+str(pos)+"-"+str(pos)
HEADERS={"Content-Type" : "application/json"}
PARAMS={"feature": "gene"}
genes_data=EnsemblRestClient.perform_rest_action(SERVER, ENDPOINT, HEADERS, PARAMS)
fusions = dict()
for gene in genes_data:
if gene["biotype"]=="protein_coding":
fusion=[]
if not ori and gene['strand'] == 1:
if 'donor' not in fusions:
fusions['donor'] = dict()
fusions['donor'][gene['id']+"\t"+bp] = gene['start']
elif not ori and gene['strand'] == -1:
if 'acceptor' not in fusions:
fusions['acceptor'] = dict()
fusions['acceptor'][gene['id']+"\t"+bp] = gene['start']
elif ori and gene['strand'] == 1:
if 'acceptor' not in fusions:
fusions['acceptor'] = dict()
fusions['acceptor'][gene['id']+"\t"+bp] = gene['end']
elif ori and gene['strand'] == -1:
if 'donor' not in fusions:
fusions['donor'] = dict()
fusions['donor'][gene['id']+"\t"+bp] = gene['end']
return( fusions )
def mask_seq(chr, start, end, strand, seq):
ext = "/overlap/region/human/" + str(chr) + ":" + str(start) + "-" + str(
end) + ":" + str(strand) + "?feature=variation"
headers = {"Content-Type": "application/json"}
data = EnsemblRestClient.perform_rest_action(server, ext, headers)
masked_seq = seq
for snp in data:
snp_pos = snp['start'] - start
if str(strand) == '-1':
snp_pos = len(masked_seq) - snp_pos - 1
masked_seq = masked_seq[:snp_pos] + 'n' + masked_seq[snp_pos + 1:]
return (masked_seq)
def get_seq(chr, start, end, strand):
ext = "/info/assembly/homo_sapiens/" + str(chr)
headers = {"Content-Type": "application/json"}
data = EnsemblRestClient.perform_rest_action(server, ext, headers)
chrlength = data['length']
ext = "/sequence/region/human/" + str(chr) + ":" + str(start) + "-" + str(
end) + ":" + str(strand) + ""
if start < 1:
sys.stderr.write(
"\tFailed to get seq, because POS is too close to START of the chr\n"
)
seq = False
elif end > chrlength:
sys.stderr.write(
"\tFailed to get seq, because ENDPOS is too close to END of the chr\n"
)
seq = False
else:
data = EnsemblRestClient.perform_rest_action(server, ext, headers)
seq = data['seq']
if mask:
seq = mask_seq(chr, start, end, strand, seq)
return (seq)
bamfile = pysam.AlignmentFile(args.bam, "rb" )
fasta = open(args.output_dir+"/"+svid+".fasta", 'a+')
for read in bamfile.fetch(chr, start, end):
#### Breakpoints that only have supplementary reads and not a primary read spanning the breakpoint will be excluded
#### No effect when testing on the truthset in recall, but see if it ever happens in real sets
if read.query_name in exclude or read.seq == None or read.is_supplementary:
continue
fasta.write( ">"+read.query_name+"\n")
fasta.write(read.seq+"\n")
#exclude.append(read.query_name)
fasta.close()
bamfile.close()
print("Start:", datetime.datetime.now())
EnsemblRestClient=EnsemblRestClient()
vcf_reader = pyvcf.Reader(open(args.vcf, 'r'))
for record in vcf_reader:
if not isinstance(record.ALT[0], pyvcf.model._Breakend):
continue
fusions={'donor':{}, 'acceptor':{}}
bnd1_fusions = get_gene_overlap(record.CHROM, record.POS, record.ALT[0].orientation, '1')
### Skip next request if the first BND already falls outside of a gene
if not fusions:
continue
bnd2_fusions=get_gene_overlap(record.ALT[0].chr, record.ALT[0].pos, record.ALT[0].remoteOrientation, '2')
fusions.update(bnd1_fusions)
if 'donor' in bnd2_fusions:
fusions['donor'].update(bnd2_fusions['donor'])
type=str,
help='Output bed file',
required=True)
parser.add_argument(
'-r',
'--region',
type=str,
help='List of genes or regions to select from the bam-file',
required=True)
args = parser.parse_args()
bed = args.bed
selection = args.region
selection = selection.split(",")
EnsemblRestClient = EnsemblRestClient()
### Create a bed-file that contains all the regions that need to be selected from the full bam-file
with open(bed, "w") as bed:
for item in selection:
### If a region is given in a chr:pos1-pos2 format, this format is directly used
if re.match(".+:\d+-\d+", item):
chrom = item.split(":")[0]
pos1 = item.split(":")[1].split("-")[0]
pos2 = item.split(":")[1].split("-")[1]
else:
server = 'http://grch37.rest.ensembl.org'
### If a desired region is given by a ensembl identifier (e.g. ENSG00000141510) or the gene name, the positions are taken from ensembl
def EnsemblAnnotation(CHROM, POS):
SERVER = 'http://grch37.rest.ensembl.org'
ENDPOINT = "/overlap/region/human/" + str(CHROM) + ":" + str(
POS) + "-" + str(POS)
HEADERS = {"Content-Type": "application/json"}
PARAMS = {"feature": "transcript"}
genes_data = EnsemblRestClient.perform_rest_action(SERVER, ENDPOINT,
HEADERS, PARAMS)
transcript_ccds = {}
UNIQUE_GENES = []
for hit in genes_data:
if hit["biotype"] == "protein_coding":
if "ccdsid" in hit:
transcript_ccds[hit["id"]] = hit["ccdsid"]
else:
transcript_ccds[hit["id"]] = None
if hit["Parent"] not in UNIQUE_GENES:
UNIQUE_GENES.append(hit["Parent"])
HITS = []
for GENE_ID in UNIQUE_GENES:
#GENE_ID=hit
ENDPOINT = "/lookup/id/" + str(GENE_ID)
PARAMS = {"expand": "1"}
gene_info = EnsemblRestClient.perform_rest_action(
SERVER, ENDPOINT, HEADERS, PARAMS)
if gene_info["biotype"] == "protein_coding":
INFO = {}
INFO["Gene_id"] = gene_info["id"]
INFO["Gene_name"] = gene_info["display_name"]
INFO["Strand"] = gene_info["strand"]
INFO["Gene_start"] = gene_info["start"]
INFO["Gene_end"] = gene_info["end"]
INFO["Chromosome"] = CHROM
INFO["Flags"] = []
if "description" in gene_info:
if "readthrough" in gene_info["description"]:
INFO["Flags"].append("fusion-with-readthrough")
##### FLAG for CTC-... and RP..... proteins (Often not well characterized or readthrough genes)
for transcript in gene_info["Transcript"]:
if transcript["is_canonical"] == 1:
if transcript["id"] in transcript_ccds:
if transcript_ccds[transcript["id"]] is None:
INFO["Flags"].append("No-CCDS")
LENGTH_CDS = 0
CDS = False
INFO["Transcript_id"] = transcript["id"]
INFO["Biotype"] = transcript["biotype"]
INFO["Transcript_start"] = transcript["start"]
INFO["Transcript_end"] = transcript["end"]
INFO["Total_exons"] = len(transcript["Exon"])
INFO["Original_CDS_length"] = (
transcript["Translation"]["length"] * 3) + 3
if INFO["Strand"] == 1:
INFO["CDS_start"] = transcript["Translation"]["start"]
INFO["CDS_end"] = transcript["Translation"]["end"]
else:
INFO["CDS_start"] = transcript["Translation"]["end"]
INFO["CDS_end"] = transcript["Translation"]["start"]
INFO["Exons"] = []
### EXONS
for rank, exon in enumerate(transcript["Exon"]):
EXON_INFO = {}
EXON_INFO["Rank"] = rank + 1
EXON_INFO["Type"] = "exon"
CHRON_START = exon["start"]
CHRON_END = exon["end"]
if INFO["Strand"] == 1:
EXON_INFO["Start"] = exon["start"]
EXON_INFO["End"] = exon["end"]
else:
EXON_INFO["Start"] = exon["end"]
EXON_INFO["End"] = exon["start"]
EXON_INFO["Contains_start_CDS"] = False
EXON_INFO["Contains_end_CDS"] = False
if CDS:
if INFO["CDS_end"] >= CHRON_START and INFO[
"CDS_end"] <= CHRON_END:
EXON_INFO["Contains_end_CDS"] = True
EXON_INFO["CDS"] = True
EXON_INFO["Start_phase"] = PHASE
#EXON_INFO["End_phase"]=-1
EXON_INFO["End_phase"] = 0
EXON_INFO["CDS_length"] = abs(
INFO["CDS_end"] - EXON_INFO["Start"]) + 1
CDS = False
else:
EXON_INFO["CDS"] = True
EXON_INFO["Start_phase"] = PHASE
EXON_INFO["End_phase"] = (
abs(EXON_INFO["End"] - EXON_INFO["Start"])
+ 1 + PHASE) % 3
EXON_INFO["CDS_length"] = abs(
EXON_INFO["End"] - EXON_INFO["Start"]) + 1
elif INFO["CDS_start"] >= CHRON_START and INFO[
"CDS_start"] <= CHRON_END:
EXON_INFO["Contains_start_CDS"] = True
EXON_INFO["CDS"] = True
#EXON_INFO["Start_phase"]=-1
EXON_INFO["Start_phase"] = 0
if INFO["CDS_end"] >= CHRON_START and INFO[
"CDS_end"] <= CHRON_END:
EXON_INFO["Contains_end_CDS"] = True
EXON_INFO["End_phase"] = 0
EXON_INFO["CDS_length"] = abs(
INFO["CDS_end"] - INFO["CDS_start"]) + 1
else:
EXON_INFO["End_phase"] = (
abs(EXON_INFO["End"] - INFO["CDS_start"]) +
1) % 3
EXON_INFO["CDS_length"] = abs(
EXON_INFO["End"] - INFO["CDS_start"]) + 1
CDS = True
else:
EXON_INFO["CDS"] = False
EXON_INFO[
"Start_phase"] = "-1" #ADD SOMETHING ELSE SO AN ERROR IS GIVEN IF USED
EXON_INFO["End_phase"] = "-1"
EXON_INFO["CDS_length"] = 0
PHASE = EXON_INFO["End_phase"]
LENGTH_CDS += EXON_INFO["CDS_length"]
INFO["Exons"].append(EXON_INFO)
### INTRONS
if rank < len(transcript["Exon"]) - 1:
INTRON_INFO = {}
INTRON_INFO["Type"] = "intron"
INTRON_INFO["Rank"] = rank + 1
INTRON_INFO["Phase"] = EXON_INFO["End_phase"]
if INFO["Strand"] == 1:
INTRON_INFO["Start"] = transcript["Exon"][
rank]["end"] + 1
INTRON_INFO["End"] = transcript["Exon"][
rank + 1]["start"] - 1
else:
INTRON_INFO["Start"] = transcript["Exon"][
rank]["start"] - 1
INTRON_INFO["End"] = transcript["Exon"][
rank + 1]["end"] + 1
# INTRON_INFO["Start"]=transcript["Exon"][rank]["end"]
# INTRON_INFO["End"]=transcript["Exon"][rank+1]["start"]
INTRON_INFO["CDS"] = CDS
INFO["Exons"].append(INTRON_INFO)
#print(INFO["Gene_id"],LENGTH_CDS, transcript["Translation"]["length"]*3)
if LENGTH_CDS - 3 != transcript["Translation"][
"length"] * 3:
INFO["Flags"].append(
"Possible-incomplete-CDS"
) #as currently I am unable to request the given ENSEMBL flags. Bias towards incomplete but bases%3=0
HITS.append(INFO)
return HITS