def main():
# parse_cdna('972_973delinsAT')
# # #print(get_cdna()[3810])
# # #parse_cdna('4510_4535del')
# # #print(get_cdna()[4509:4535])
# exit()
if len(argv) < 2:
print(f"usage: {argv[0]} <input tsv>")
exit()
in_tsv = argv[1]
cases = parse_in(in_tsv)
db, cursor = abca4_connect()
for case in cases:
[
pubmed_id, ref, patient_id, c1, p1, c2, p2, value, onset,
progression_string
] = case
print(case)
# for allele1, allele 2
allele_ids = []
for [cdna_variants, protein_variants] in [[c1, p1], [c2, p2]]:
# store or retrieve variant ids
variant_ids = store_variants(cursor, cdna_variants,
protein_variants)
# store or retrieve allele id
allele_id = store_allele(cursor, variant_ids)
if not allele_id: panic(["no allele id for "] + variant_ids)
allele_ids.append(allele_id)
# store or retrieve publication id
publication_id = store_publication(cursor, pubmed_id, ref)
fixed_fields = {
'publication_id': publication_id,
'patient_xref_id': patient_id
}
update_fields = {
'allele_id_1': allele_ids[0],
'allele_id_2': allele_ids[1],
'onset_age': onset,
'acuity_type': 'decimal',
'eye': 'better',
'progression': progression_string
}
#store case: allele_id_1, allele_id_2, publication_id, patient_id, onset, value, better, progression
store_or_update(cursor,
"cases",
fixed_fields=fixed_fields,
update_fields=update_fields)
# print(fixed_fields)
# print(update_fields)
# print()
#exit()
cursor.close()
db.close()
def store_variants(cursor, c, p, verbose=False):
cdna_vars = [v.replace(" ", "") for v in c.split(";")]
protein_vars = [v.replace(" ", "") for v in p.split(";")]
if len(cdna_vars) == 0:
print(f"cdna vars empty {c} {p}")
exit()
if len(protein_vars) == 0:
print(f"proteins vars empty {c} {p}")
exit()
if len(cdna_vars) != len(protein_vars):
print(f"var lengths not equal {c} {p}")
exit()
var_ids = []
for i in range(len(cdna_vars)):
cdna_var = cdna_vars[i].replace(" ", "")
protein_var = protein_vars[i].replace(" ", "")
if cdna_var == "np":
ok = store_variant_w_cdna_uknown(cursor, protein_var, var_ids,
verbose)
if not ok: panic([cdna_vars, protein_vars])
else:
ok = store_variant_w_known_cdna(cursor, cdna_var, protein_var,
var_ids, verbose)
if not ok: panic([cdna_vars, protein_vars])
if len(var_ids) != len(cdna_vars):
print(f"missing variant id for {cdna_vars} {protein_vars} (?)")
exit()
return var_ids
def main():
db, cursor = abca4_connect()
# find cases with at least three progression points
qry = "select id, allele_id_1, allele_id_2, onset_age, progression, publication_id from cases where "
qry += "(progression like '%:%:%:%' or (progression like '%:%:%' and onset_age is not null and onset_age>0)) "
qry += "and (notes is null or notes not like '%caveat%')"
# find the variants corresponding to those cases
for [case_id, allele_id_1, allele_id_2, onset_age, progression, publication_id] in hard_landing_search(cursor,qry):
params = {}
variants = {}
for ai in [allele_id_1, allele_id_2]:
variants[ai] = hard_landing_search(cursor, f"select variant_ids from alleles where id={ai}")[0][0].strip("-").split("-")
for v in variants[ai]:
ret = error_intolerant_search(cursor, f"select * from parametrization_literature where variant_id={v}")
if ret:
if len(ret)>1: panic([f"multiple parametrization_literature for varid {v}"])
params[v] = ret[0]
else:
# is this null by any chance?
ret = error_intolerant_search(cursor, f"select * from parametrization where variant_id={v}")
if ret:
if len(ret)>1: panic([f"multiple parametrization for varid {v}"])
[prm_id, var_id, e, t, notes] = ret[0]
if e>0.001: continue # this is expressing
params[v] = [prm_id, var_id, e, t, 0]
if len(params)!=(len(variants[allele_id_1])+len(variants[allele_id_2])): continue
# keep only if all variants have experimental support
# i am still not ready to deal with multiple variants per allele
if len(variants[allele_id_1])>1 or len(variants[allele_id_2])>1: continue
print()
print(case_id, "onset age:", onset_age)
print(progression, publication_id)
for ai in [allele_id_1, allele_id_2]:
print("\t", ai, variants[ai])
for v in variants[ai]:
print(f"\t\t variant {v} {params[v]}")
varid1 = variants[allele_id_1][0]
varid2 = variants[allele_id_2][0]
params1 = params[varid1][2:4]
params2 = params[varid2][2:4]
print(varid1, params1)
print(varid2, params2)
age, va = unpack_progression(progression)
if onset_age and onset_age>0:
age = [max(onset_age-1, 0)] + age
va = [1.0] + va
# plot_sim_results_vs_data(age, va, varid1, varid2, params1, params2, rpe_baseline)
plot_sim_results_vs_data(age, va, varid1, varid2, params1, params2, 0.1)
cursor.close()
db.close()
def store_publication(cursor, pubmed_id, reference):
if (not pubmed_id or pubmed_id.lower() != "none") and not reference:
panic(["null entry for publication"])
if pubmed_id and pubmed_id.lower() != "none":
return store_publication_by_pubmed_id(cursor, pubmed_id, reference)
return store_publication_by_reference(cursor, reference)
def store_publication_by_reference(cursor, reference):
publication_id = None
ret = error_intolerant_search(
cursor, f"select id from publications where reference='{reference}'")
if not ret:
qry = f"insert into publications (reference) values ('{reference}')"
if search_db(cursor, qry, verbose=True): exit()
publication_id = hard_landing_search(
cursor, "select max(id) from publications")[0][0]
elif len(ret) > 1:
panic(["multiple returns for", reference])
else:
publication_id = ret[0][0]
return publication_id
def store_allele(cursor, variant_ids):
allele_id = None
variant_signature = "-" + "-".join([str(i)
for i in sorted(variant_ids)]) + "-"
ret = error_intolerant_search(
cursor,
f"select id from alleles where variant_ids='{variant_signature}'")
if not ret:
qry = f"insert into alleles (variant_ids) values ('{variant_signature}')"
if search_db(cursor, qry, verbose=True): exit()
allele_id = hard_landing_search(cursor,
"select max(id) from alleles")[0][0]
elif len(ret) > 1:
panic(["multiple returns for", variant_signature])
else:
allele_id = ret[0][0]
return allele_id
def store_publication(cursor, url, pmc, pubmed_id, ref):
publication_id = None
ret = error_intolerant_search(
cursor, f"select id from publications where other_xref='{url}'")
if not ret:
qry = f"insert into publications (reference, other_xref) values ('{ref}', '{url}')"
if search_db(cursor, qry, verbose=True): exit()
publication_id = hard_landing_search(
cursor, "select max(id) from publications")[0][0]
if pubmed_id:
qry = f"update publications set pubmed={pubmed_id} where id={publication_id}"
if search_db(cursor, qry, verbose=True): exit()
if pmc:
qry = f"update publications set pubmedcentral='{pmc}' where id={publication_id}"
if search_db(cursor, qry, verbose=True): exit()
elif len(ret) > 1:
panic(["multiple returns for", pubmed_id])
else:
publication_id = ret[0][0]
return publication_id