def get_acs_for_protein_seq(self, seq):
"""
returns a list of protein accessions for a given sequence. The
list is guaranteed to contain at least one element with the
MD5-based accession (MD5_01234abc...def56789) at the end of the
list.
"""
md5 = seq_md5(seq)
return [r['ac'] for r in self._fetchall(self._queries['acs_for_protein_md5'], [md5])
] + ['MD5_' + md5]
def get_acs_for_protein_seq(self,seq):
"""
returns a list of protein accessions for a given sequence. The
list is guaranteed to contain at least one element with the
MD5-based accession (MD5_01234abc...def56789) at the end of the
list.
"""
md5 = seq_md5(seq)
cur = self._get_cursor()
cur.execute(self.sql['acs_for_protein_md5'],[md5])
return [ r['ac'] for r in cur.fetchall() ] + [ 'MD5_'+md5 ]
def load_txinfo(session, opts, cf):
self_aln_method = "transcript"
update_period = 250
sf = None # established on first use, below
@lru_cache(maxsize=100)
def _fetch_origin_by_name(name):
try:
ori = session.query(usam.Origin).filter(
usam.Origin.name == name).one()
except NoResultFound as e:
logger.error("No origin for " + ti.origin)
raise e
return ori
n_rows = len(gzip.open(opts["FILE"]).readlines()) - 1
tir = ufti.TxInfoReader(gzip.open(opts["FILE"]))
logger.info("opened " + opts["FILE"])
session.execute("set role {admin_role};".format(
admin_role=cf.get("uta", "admin_role")))
session.execute("set search_path = " + usam.schema_name)
n_new = 0
n_unchanged = 0
n_cds_changed = 0
n_exons_changed = 0
for i_ti, ti in enumerate(tir):
if ti.exons_se_i == "":
logger.warning(ti.ac + ": no exons?!; skipping.")
continue
if ti.cds_se_i:
cds_start_i, cds_end_i = map(int, ti.cds_se_i.split(","))
else:
cds_start_i = cds_end_i = None
cds_md5 = None
# 1. Fetch or make the Transcript record
existing = session.query(usam.Transcript).filter(
usam.Transcript.ac == ti.ac,
)
assert existing.count() <= 1, "Expected max 1 existing transcripts with accession {ti.ac}".format(ti=ti)
u_tx = None
if existing.count() == 1:
u_tx = existing[0]
if (u_tx.cds_start_i, u_tx.cds_end_i) != (cds_start_i, cds_end_i):
u_tx.ac = "{u_tx.ac}/{u_tx.cds_start_i}..{u_tx.cds_end_i}".format(u_tx=u_tx)
logger.warn("Transcript {ti.ac}: CDS coordinates changed!; renamed to {u_tx.ac}".format(ti=ti, u_tx=u_tx))
session.flush()
u_tx = None
n_cds_changed += 1
# state: u_tx is set if a transcript was found and was
# unchanged, or None if 1) no such was found or 2) was found
# and had updated CDS coords.
if u_tx is None:
ori = _fetch_origin_by_name(ti.origin)
if ti.cds_se_i:
if sf is None:
sf = _get_seqfetcher(cf)
try:
cds_seq = sf.fetch(ti.ac, cds_start_i, cds_end_i)
except KeyError:
raise Exception("{ac}: not in sequence database".format(ac=ti.ac))
cds_md5 = seq_md5(cds_seq)
else:
cds_md5 = None
assert (cds_start_i is not None) ^ (cds_md5 is None), "failed: cds_start_i is None i.f.f. cds_md5_is None"
u_tx = usam.Transcript(
ac=ti.ac,
origin=ori,
hgnc=ti.hgnc,
cds_start_i=cds_start_i,
cds_end_i=cds_end_i,
cds_md5=cds_md5,
)
session.add(u_tx)
if u_tx.hgnc != ti.hgnc:
logger.warn("{ti.ac}: HGNC symbol changed from {u_tx.hgnc} to {ti.hgnc}".format(
u_tx=u_tx, ti=ti))
u_tx.hgnc = ti.hgnc
# state: transcript now exists, either existing or freshly-created
# 2. Upsert an ExonSet attached to the Transcript
n, o = _upsert_exon_set_record(session, ti.ac, ti.ac, 1, self_aln_method, ti.exons_se_i)
(no) = (n is not None, o is not None)
if no == (True, False):
n_new += 1
elif no == (True, True):
logger.warn("Transcript {ti.ac} exon structure changed".format(ti=ti))
n_exons_changed += 1
elif no == (False, True):
logger.debug("Transcript {ti.ac} exon structure unchanged".format(ti=ti))
n_unchanged += 1
if i_ti % update_period == 0 or i_ti + 1 == n_rows:
session.commit()
logger.info("{i_ti}/{n_rows} {p:.1f}%; {n_new} new, {n_unchanged} unchanged, "
"{n_cds_changed} cds changed, {n_exons_changed} exons changed; commited".format(
i_ti=i_ti, n_rows=n_rows,
n_new=n_new, n_unchanged=n_unchanged, n_cds_changed=n_cds_changed, n_exons_changed=n_exons_changed,
p=(i_ti + 1) / n_rows * 100))