def test_get_delta_score_acceptor(self):
record = Record('10', 94077, 'A', ['C'])
scores = get_delta_scores(record, self.ann, 500, 0)
self.assertEqual(scores, ['C|TUBB8|0.15|0.27|0.00|0.05|89|-23|-267|193'])
scores = get_delta_scores(record, self.ann_without_prefix, 500, 0)
self.assertEqual(scores, ['C|TUBB8|0.15|0.27|0.00|0.05|89|-23|-267|193'])
record = Record('chr10', 94077, 'A', ['C'])
scores = get_delta_scores(record, self.ann, 500, 0)
self.assertEqual(scores, ['C|TUBB8|0.15|0.27|0.00|0.05|89|-23|-267|193'])
scores = get_delta_scores(record, self.ann_without_prefix, 500, 0)
self.assertEqual(scores, ['C|TUBB8|0.15|0.27|0.00|0.05|89|-23|-267|193'])
def test_get_delta_score_donor(self):
record = Record('10', 94555, 'C', ['T'])
scores = get_delta_scores(record, self.ann)
self.assertEqual(scores, ['T|TUBB8|0.01|0.18|0.15|0.62|-2|110|-190|0'])
scores = get_delta_scores(record, self.ann_without_prefix)
self.assertEqual(scores, ['T|TUBB8|0.01|0.18|0.15|0.62|-2|110|-190|0'])
record = Record('chr10', 94555, 'C', ['T'])
scores = get_delta_scores(record, self.ann)
self.assertEqual(scores, ['T|TUBB8|0.01|0.18|0.15|0.62|-2|110|-190|0'])
scores = get_delta_scores(record, self.ann_without_prefix)
self.assertEqual(scores, ['T|TUBB8|0.01|0.18|0.15|0.62|-2|110|-190|0'])
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 test_get_delta_score_donor(self):
''' test get_delta_scores for a predicted donor
'''
class Record():
chrom, pos, ref, alts = '10', 94555, 'C', ['T']
record = Record()
scores = get_delta_scores(record, self.ann)
self.assertEqual(scores, ['T|TUBB8|0.01|0.18|0.15|0.62|-2|110|-190|0'])
def test_get_delta_score_acceptor(self):
''' test get_delta_scores for a predicted acceptor
'''
class Record():
chrom, pos, ref, alts = '10', 94077, 'A', ['C']
record = Record()
scores = get_delta_scores(record, self.ann)
self.assertEqual(scores,
['C|TUBB8|0.15|0.27|0.00|0.05|89|-23|-267|193'])
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()
vcf = pysam.VariantFile(args.I)
header = vcf.header
header.add_line(
'##INFO=<ID=SpliceAI,Number=.,Type=String,Description="SpliceAI 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
###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 process_record(ann, distance, mask, record):
scores = get_delta_scores(record, ann, distance, mask)
if len(scores) > 0:
record.info['SpliceAI'] = scores
return record
def process_variant(variant, genome_version, spliceai_distance, spliceai_mask, use_precomputed_scores):
try:
chrom, pos, ref, alt = parse_variant(variant)
except ValueError as e:
return {
"variant": variant,
"error": f"ERROR: {e}",
}
if len(ref) > 1 and len(alt) > 1:
return {
"variant": variant,
"error": f"ERROR: SpliceAI does not currently support complex InDels like {chrom}-{pos}-{ref}-{alt}",
}
# generate error message if variant falls outside annotated exons or introns
OTHER_GENOME_VERSION = {"37": "38", "38": "37"}
chrom_without_chr = chrom.replace("chr", "")
if not ANNOTATION_INTERVAL_TREES[genome_version][chrom_without_chr].at(pos):
other_genome_version = OTHER_GENOME_VERSION[genome_version]
other_genome_overlapping_intervals = ANNOTATION_INTERVAL_TREES[other_genome_version][chrom_without_chr].at(pos)
if other_genome_overlapping_intervals:
other_genome_genes = " and ".join(sorted(set([str(i.data).split("---")[0] for i in other_genome_overlapping_intervals])))
return {
"variant": variant,
"error": f"ERROR: In GRCh{genome_version}, {chrom}-{pos}-{ref}-{alt} falls outside all gencode exons and introns."
f"SpliceAI only works for variants within known exons or introns. However, in GRCh{other_genome_version}, "
f"{chrom}:{pos} falls within {other_genome_genes}, so perhaps GRCh{genome_version} is not the correct genome version?"
}
else:
return {
"variant": variant,
"error": f"ERROR: {chrom}-{pos}-{ref}-{alt} falls outside all Gencode exons and introns on "
f"GRCh{genome_version}. SpliceAI only works for variants that are within known exons or introns.",
}
"""
NOTE: The reason SpliceAI currently works only for variants "
f"within annotated exons or introns is that, although the SpliceAI neural net takes any "
f"arbitrary nucleotide sequence as input, SpliceAI needs 1) the transcript strand "
f"to determine whether to reverse-complement the reference genome sequence before passing it "
f"to the neural net, and 2) transcript start and end positions to determine where to truncate "
f"the reference genome sequence.
"""
source = None
scores = []
if (len(ref) <= 5 or len(alt) <= 2) and str(spliceai_distance) == str(SPLICEAI_DEFAULT_DISTANCE) and str(use_precomputed_scores) == "1":
# examples: ("masked", "snv", "hg19") ("raw", "indel", "hg38")
key = (
"masked" if str(spliceai_mask) == "1" else ("raw" if str(spliceai_mask) == "0" else None),
"snv" if len(ref) == 1 and len(alt) == 1 else "indel",
"hg19" if genome_version == "37" else ("hg38" if genome_version == "38" else None),
)
try:
results = SPLICEAI_CACHE_FILES[key].fetch(chrom, pos-1, pos+1)
for line in results:
# ['1', '739023', '.', 'C', 'CT', '.', '.', 'SpliceAI=CT|AL669831.1|0.00|0.00|0.00|0.00|-1|-37|-48|-37']
fields = line.split("\t")
if fields[0] == chrom and int(fields[1]) == pos and fields[3] == ref and fields[4] == alt:
scores.append(fields[7])
if scores:
source = "lookup"
#print(f"Fetched: ", scores, flush=True)
except Exception as e:
print(f"ERROR: couldn't retrieve scores using tabix: {type(e)}: {e}", flush=True)
if not scores:
if exceeds_rate_limit(request.remote_addr, request_type="SpliceAI: computed"):
return {
"variant": variant,
"error": f"ERROR: Rate limit reached. To prevent a user from overwhelming the server and making it "
f"unavailable to other users, this tool allows no more than "
f"{RATE_LIMIT_REQUESTS_PER_USER_PER_MINUTE['SpliceAI: computed']} computed requests per minute per user.",
}
record = VariantRecord(chrom, pos, ref, alt)
try:
scores = get_delta_scores(
record,
SPLICEAI_ANNOTATOR[genome_version],
spliceai_distance,
spliceai_mask)
source = "computed"
#print(f"Computed: ", scores, flush=True)
except Exception as e:
return {
"variant": variant,
"error": f"ERROR: {type(e)}: {e}",
}
if not scores:
return {
"variant": variant,
"error": f"ERROR: The SpliceAI model did not return any scores for {variant}. This is typically due to the "
f"variant falling outside of all Gencode exons and introns.",
}
scores = [s[s.index("|")+1:] for s in scores] # drop allele field
return {
"variant": variant,
"genome_version": genome_version,
"chrom": chrom,
"pos": pos,
"ref": ref,
"alt": alt,
"scores": scores,
"source": source,
}
