def setUpClass(cls):
fasta_path = resource_filename(__name__, 'data/test.fa')
fasta_without_prefix_path = resource_filename(
__name__, 'data/test_without_prefix.fa')
cls.ann = Annotator(fasta_path, 'grch37')
cls.ann_without_prefix = Annotator(fasta_without_prefix_path, 'grch37')
def run_serial(args):
"""
串行运行
"""
try:
vcf = pysam.VariantFile(args.I)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
header = vcf.header
header.add_line(
'##INFO=<ID=SpliceAI,Number=.,Type=String,Description="SpliceAIv1.3 variant '
'annotation. These include delta scores (DS) and delta positions (DP) for '
'acceptor gain (AG), acceptor loss (AL), donor gain (DG), and donor loss (DL). '
'Format: ALLELE|SYMBOL|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL">'
)
try:
output = pysam.VariantFile(args.O, mode='w', header=header)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
ann = Annotator(args.R, args.A)
for record in vcf:
scores = get_delta_scores(record, ann, args.D, args.M)
if len(scores) > 0:
record.info['SpliceAI'] = scores
output.write(record)
vcf.close()
output.close()
def main():
args = get_options()
if None in [args.I, args.O, args.D, args.M]:
logging.error(
'Usage: spliceai [-h] [-I [input]] [-O [output]] -R reference -A annotation '
'[-D [distance]] [-M [mask]]')
exit()
try:
vcf = pysam.VariantFile(args.I)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
header = vcf.header
header.add_line(
'##INFO=<ID=SpliceAI,Number=.,Type=String,Description="SpliceAIv1.3 variant '
'annotation. These include delta scores (DS) and delta positions (DP) for '
'acceptor gain (AG), acceptor loss (AL), donor gain (DG), and donor loss (DL). '
'This version also includes the distance (DIST) to the nearest splice site.'
'Format: ALLELE|SYMBOL|DIST|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL">'
)
try:
output = pysam.VariantFile(args.O, mode='w', header=header)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
# loading prescored files
prescored_files = []
try:
for filename in args.P:
vcf_file = pysam.VariantFile(filename)
prescored_files.append(vcf_file)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
ann = Annotator(args.R, args.A)
for record in vcf:
scores = get_delta_scores(record, ann, args.D, args.M, prescored_files)
if len(scores) > 0:
record.info['SpliceAI'] = scores
output.write(record)
vcf.close()
output.close()
def main():
args = get_options()
if None in [args.I, args.O, args.D, args.M]:
logging.error(
'Usage: spliceai [-h] [-I [input]] [-O [output]] -R reference -A annotation '
'[-D [distance]] [-M [mask]]')
exit()
try:
vcf = pysam.VariantFile(args.I)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
header = vcf.header
###Adding header lines required to satisfy vcf format output.write()
header.add_line('###fileDate=20191004')
header.add_line('##reference=GRCh37/hg19')
header.add_line(
'##INFO=<ID=SpliceAI,Number=.,Type=String,Description="SpliceAIv1.3.1 variant '
'annotation. These include delta scores (DS) and delta positions (DP) for '
'acceptor gain (AG), acceptor loss (AL), donor gain (DG), and donor loss (DL). '
'Format: ALLELE|SYMBOL|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL">'
)
header.add_line(
'##INFO=<ID=NS,Number=20000,Type=Integer,Description="Dummy NS">')
try:
output = pysam.VariantFile(args.O, mode='w', header=header)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
ann = Annotator(args.R, args.A)
for record in vcf:
scores = get_delta_scores(record, ann, args.D, args.M)
if len(scores) > 0:
record.info['SpliceAI'] = scores
output.write(record)
vcf.close()
output.close()
def process_record(records, results, ref_fasta, annotations, dist_var, mask):
# 创建一个注释助手类
ann = Annotator(ref_fasta, annotations)
# 监听队列
while True:
# 尝试从队列获得一个待打分的变异
try:
record = records.get_nowait()
except queue.Empty:
continue
# 判断队列是否结束
if record != 'END':
# 对变异进行打分并把结果放入队列
scores = get_delta_scores(record, ann, dist_var, mask)
results.put((record.id, scores))
else:
# 队列结束，重新把结束标志放入队列，以终止其他进程
records.put('END')
break
def main():
args = get_options()
vcf = pysam.VariantFile(args.I)
header = vcf.header
header.add_line('##INFO=<ID=SpliceAI,Number=.,Type=String,Description="SpliceAIv1.2 variant '
'annotation. These include delta scores (DS) and delta positions (DP) for '
'acceptor gain (AG), acceptor loss (AL), donor gain (DG), and donor loss (DL). '
'Format: ALLELE|SYMBOL|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL">')
output = pysam.VariantFile(args.O, mode='w', header=header)
ann = Annotator(args.R, args.A)
for record in vcf:
scores = get_delta_scores(record, ann)
if len(scores) > 0:
record.info['SpliceAI'] = scores
output.write(record)
def main():
args = get_options()
if None in [args.I, args.O, args.D, args.M]:
logging.error(
'Usage: spliceai [-h] [-I [input]] [-O [output]] -R reference -A annotation '
'[-D [distance]] [-M [mask]]')
exit()
try:
vcf = pysam.VariantFile(args.I)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
header = vcf.header
header.add_line(
'##INFO=<ID=SpliceAI,Number=.,Type=String,Description="SpliceAIv1.3 variant '
'annotation. These include delta scores (DS) and delta positions (DP) for '
'acceptor gain (AG), acceptor loss (AL), donor gain (DG), and donor loss (DL). '
'Format: ALLELE|SYMBOL|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL">'
)
try:
output = pysam.VariantFile(args.O, mode='w', header=header)
except (IOError, ValueError) as e:
logging.error('{}'.format(e))
exit()
ann = Annotator(args.R, args.A)
run = partial(process_record, ann, args.D, args.M)
with mp.Pool(args.t) as p:
for record in p.map(run, vcf):
output.write(record)
vcf.close()
output.close()
def get_variant_assessment():
logger.debug("Invoking get_variant_assessment")
logger.debug("Chromosome: " + request.form['chrom'])
logger.debug("Position: " + request.form['pos'])
logger.debug("Ref: " + request.form['ref'])
logger.debug("Alt: " + request.form['alt'])
logger.debug("Assembly: " + request.form['assembly'])
logger.debug("Distance: " + request.form['distance'])
logger.debug("Mask: " + request.form['mask'])
Record.chrom = request.form['chrom']
Record.pos = int(request.form['pos'])
Record.ref = request.form['ref']
Record.alts = [request.form['alt']]
assembly = request.form['assembly']
distance = int(request.form['distance'])
mask = int(request.form['mask'])
ann = Annotator(app.config['HG19'], assembly)
pred = get_delta_scores(Record, ann, distance, mask)
return json.dumps(pred)
GRCH38_ANNOTATIONS = "./annotations/gencode.v38.annotation.txt.gz"
ANNOTATION_INTERVAL_TREES = {
"37": defaultdict(IntervalTree),
"38": defaultdict(IntervalTree),
}
for genome_version, annotation_path in ("37", GRCH37_ANNOTATIONS), ("38", GRCH38_ANNOTATIONS):
print(f"Loading {annotation_path}", flush=True)
df = pd.read_table(annotation_path, dtype={"TX_START": int, "TX_END": int})
for _, row in df.iterrows():
chrom = row["CHROM"].replace("chr", "")
ANNOTATION_INTERVAL_TREES[genome_version][chrom].add(Interval(row["TX_START"], row["TX_END"] + 0.1, row["#NAME"]))
SPLICEAI_ANNOTATOR = {
"37": Annotator(HG19_FASTA_PATH, GRCH37_ANNOTATIONS),
"38": Annotator(HG38_FASTA_PATH, GRCH38_ANNOTATIONS),
}
SPLICEAI_MAX_DISTANCE_LIMIT = 10000
SPLICEAI_DEFAULT_DISTANCE = 50 # maximum distance between the variant and gained/lost splice site, defaults to 50
SPLICEAI_DEFAULT_MASK = 0 # mask scores representing annotated acceptor/donor gain and unannotated acceptor/donor loss, defaults to 0
USE_PRECOMPUTED_SCORES = 1 # whether to use precomputed scores by default
SPLICEAI_SCORE_FIELDS = "ALLELE|SYMBOL|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL".split("|")
SPLICEAI_EXAMPLE = f"/spliceai/?hg=38&distance=50&mask=0&precomputed=1&variant=chr8-140300615-C-G"
VARIANT_RE = re.compile(
"(chr)?(?P<chrom>[0-9XYMTt]{1,2})"
"[-\s:]+"