def compute_hgvs(x):
v = VCFVariant(x[CHR_COL], x[POS_COL], x[REF_COL], x[ALT_COL])
v = v.to_hgvs_obj(hgvs_proc.contig_maps[HgvsWrapper.GRCh38_Assem])
if normalize:
vn = hgvs_proc.normalizing(v)
v = vn if vn else v
return hgvs_proc.genomic_to_cdna(v)
def convert_to_hg37(vars, brca_resources_dir):
def pseudo_vcf_entry(v):
entries = [v.chr, v.pos, '.', v.ref, v.alt, '', '', '']
return '\t'.join([str(s) for s in entries])
lst = [pseudo_vcf_entry(v) for v in vars]
vcf_tmp = tempfile.mktemp('.vcf')
with open(vcf_tmp, 'w') as f:
f.write('\n'.join(lst))
vcf_tmp_out = tempfile.mktemp('.vcf')
args = [
"CrossMap.py", "vcf", brca_resources_dir + "/hg38ToHg19.over.chain.gz",
vcf_tmp, brca_resources_dir + "/hg19.fa", vcf_tmp_out
]
logging.info("Running CrossMap.py to convert to hg19")
sp = subprocess.Popen(args)
out, err = sp.communicate()
if out:
logging.info("standard output of subprocess: {}".format(out))
if err:
logging.info("standard output of subprocess: {}".format(err))
vcf_out_lines = open(vcf_tmp_out, 'r').readlines()
return [
VCFVariant(v[0], int(v[1]), v[3], v[4])
for v in [l.strip().split('\t') for l in vcf_out_lines]
]
def test_hg19_to_hg38(hgvs_wrapper):
vars38 = [
'chr13:g.32316477:AAG>A', 'chr17:g.43079204:A>C',
'chr17:g.43053755:G>A', 'chr17:g.43125273:T>C', None
] # outside transcript boundaries
vars37 = [
'chr13:g.32890614:AAG>A', 'chr17:g.41231221:A>C',
'chr17:g.41205772:G>A', 'chr17:g.41277290:T>C',
'chr17:g.41279883:GAC>G'
]
for (v37, v38) in zip(vars37, vars38):
v37_obj = VCFVariant.from_str(v37).to_hgvs_obj(
hgvs_wrapper.contig_maps[hgvs_utils.HgvsWrapper.GRCh37_Assem])
if v38:
v38_obj = VCFVariant.from_hgvs_obj(
hgvs_wrapper.hg19_to_hg38(v37_obj))
assert VCFVariant.from_str(v38) == v38_obj
else:
with pytest.raises(ValueError):
hgvs_wrapper.hg19_to_hg38(v37_obj)
def from_genomic(parsedLine, fieldIdxDict, hgvs_wrapper, seq_fetcher37,
errorsFile):
acc = 'NC_0000' + str(parsedLine[fieldIdxDict['chromosome']].replace(
'chr', '')) + '.10'
var_str = acc + ":" + parsedLine[fieldIdxDict['gDNA']]
try:
var_hgvs = hgvs_wrapper.hgvs_parser.parse(var_str)
var_hgvs_norm = hgvs.normalizer.Normalizer(
hgvs_wrapper.hgvs_dp, shuffle_direction=5).normalize(var_hgvs)
return VCFVariant.from_hgvs_obj(var_hgvs_norm, seq_fetcher37)
except HGVSError as e:
print('Could not parse genomic field ' + str(var_str) +
'. Error was ' + str(e),
file=errorsFile)
return None
def hgvs_variant(hgvs_wrapper):
return (VCFVariant.from_str('chr13:g.32316477:AAG>A').to_hgvs_obj(
hgvs_wrapper.contig_maps[hgvs_utils.HgvsWrapper.GRCh38_Assem]))
def main(input, output, pkl, log_path, config_file, resources):
logging.basicConfig(filename=log_path,
filemode="w",
level=logging.INFO,
format=' %(asctime)s %(filename)-15s %(message)s')
cfg_df = config.load_config(config_file)
syn_ac_dict = {
x[config.SYMBOL_COL]: x[config.SYNONYM_AC_COL].split(';')
for _, x in cfg_df.iterrows()
}
cdna_default_ac_dict = {
x[config.SYMBOL_COL]: x[config.HGVS_CDNA_DEFAULT_AC]
for _, x in cfg_df.iterrows()
}
hgvs_proc = HgvsWrapper()
df = pd.read_csv(input, sep='\t')
df[VAR_OBJ_FIELD] = df.apply(
lambda x: VCFVariant(x[CHR_COL], x[POS_COL], x[REF_COL], x[ALT_COL]),
axis=1)
#### CDNA conversions
df[TMP_CDNA_NORM_FIELD] = _get_cdna(df,
pkl,
hgvs_proc,
cdna_default_ac_dict,
normalize=True)
df[PYHGVS_CDNA_COL] = df[TMP_CDNA_NORM_FIELD].apply(str)
available_cdna = df[PYHGVS_CDNA_COL].str.startswith("NM_")
df.loc[available_cdna, REFERENCE_SEQUENCE_COL] = df.loc[
available_cdna, PYHGVS_CDNA_COL].str.split(':').apply(lambda l: l[0])
df.loc[available_cdna, HGVS_CDNA_COL] = df.loc[
available_cdna, PYHGVS_CDNA_COL].str.split(':').apply(lambda l: l[1])
# still setting a reference sequence for downstream steps, even though no cDNA could be determined
df.loc[~available_cdna, REFERENCE_SEQUENCE_COL] = df.loc[
~available_cdna,
GENE_SYMBOL_COL].apply(lambda g: cdna_default_ac_dict[g])
df.loc[~available_cdna, HGVS_CDNA_COL] = '-'
#### Genomic Coordinates
df[PYHGVS_GENOMIC_COORDINATE_38_COL] = df[VAR_OBJ_FIELD].apply(
lambda v: str(v))
var_objs_hg37 = convert_to_hg37(df[VAR_OBJ_FIELD], resources)
df[PYHGVS_GENOMIC_COORDINATE_37_COL] = pd.Series(
[str(v) for v in var_objs_hg37])
df[PYHGVS_HG37_START_COL] = pd.Series([v.pos for v in var_objs_hg37])
df[PYHGVS_HG37_END_COL] = df[PYHGVS_HG37_START_COL] + (df[HG38_END_COL] -
df[HG38_START_COL])
#### Protein
df[PYHGVS_PROTEIN_COL] = df[TMP_CDNA_NORM_FIELD].apply(
lambda x: str(hgvs_proc.to_protein(x)))
#### Synonyms
df[NEW_SYNONYMS_FIELD] = df.apply(
lambda s: get_synonyms(s, hgvs_proc, syn_ac_dict), axis=1)
df[SYNONYMS_COL] = df[SYNONYMS_COL].fillna('').str.strip()
# merge existing synonyms with generated ones and sort them
df[SYNONYMS_COL] = df.apply(_merge_synonyms, axis=1)
#### Writing out
# cleaning up temporary fields
df = df.drop(
columns=[VAR_OBJ_FIELD, NEW_SYNONYMS_FIELD, TMP_CDNA_NORM_FIELD])
df.to_csv(output, sep='\t', index=False)