def insert_new_proteins_n_alignments(session, target_sequence, record):
database = record.database
# insert the database entity (if it doesn't already exist)
q_get_database = 'match $db isa database, has name "' + database + '"; get $db;'
q_insert_database = 'insert $db isa database, has name "' + database + '";'
database_id = insert_if_non_existent(session, q_get_database,
q_insert_database, "$db")
for alignment in record.alignments:
# insert the protein entity (if it doesn't already exist)
protein_name = alignment.hit_def.split(" >")[0].split(";")[0]
q_get_protein = 'match $pr isa protein, has name "' + protein_name + '"; get $pr;'
q_insert_protein = 'insert $pr isa protein, has name "' + protein_name + '";'
protein_id = insert_if_non_existent(session, q_get_protein,
q_insert_protein, "$pr")
for hsp in alignment.hsps:
sequence = hsp.sbjct
# insert the sequence attribute for the protein entity (if doesn't exists already)
q_get_protein_sequence = 'match $pr id ' + protein_id + ', has sequence $seq; $seq "' + sequence + '"; get $seq;'
q_insert_protein_sequence = 'match $pr id ' + protein_id + '; insert $seq isa sequence; $seq "' + sequence + '"; $pr has sequence $seq;'
insert_if_non_existent(session, q_get_protein_sequence,
q_insert_protein_sequence, "$seq")
# insert the sourcing-of-information relationship (if it doesn't already exist)
q_get_sourcing = (
'match $seq isa sequence; $seq "' + sequence + '"; $db id ' +
database_id + '; ' +
'$sourcing (information-source: $db, sourced-information: $seq) isa sourcing-of-information; '
+ 'get $sourcing;')
q_insert_sourcing = (
'match $seq isa sequence; $seq "' + sequence + '"; $db id ' +
database_id + ';' +
'insert $sourcing (information-source: $db, sourced-information: $seq) isa sourcing-of-information;'
)
insert_if_non_existent(session, q_get_sourcing, q_insert_sourcing,
"$sourcing")
# insert the alignment relationship (if it doesn't already exist)
sequence_positivity = round(hsp.positives / alignment.length, 3)
sequence_identicality = round(hsp.identities / alignment.length, 3)
sequence_gaps = round(hsp.gaps / alignment.length, 5)
sequence_midline = hsp.match
alignment_identifier = alignment.hit_id.split("|", 4)[3]
q_get_alignment = (
'match $target-seq isa sequence; $target-seq "' +
target_sequence + '"; ' +
'$matched-seq isa sequence; $matched-seq "' + sequence +
'"; ' +
'$alignment (target-sequence: $target-seq, matched-sequence: $matched-seq) isa sequence-sequence-alignment; '
+ 'get $alignment;')
q_insert_alignment = (
'match $target-seq isa sequence; $target-seq "' +
target_sequence + '"; ' +
'$matched-seq isa sequence; $matched-seq "' + sequence +
'"; ' +
'insert $alignment (target-sequence: $target-seq, matched-sequence: $matched-seq) isa sequence-sequence-alignment; '
+ '$alignment has sequence-positivity ' +
str(sequence_positivity) + ', has sequence-identicality ' +
str(sequence_identicality) + ', has sequence-gaps ' +
str(sequence_gaps) + ', has sequence-midline "' +
sequence_midline + '"' + ', has identifier "' +
alignment_identifier + '";')
insert_if_non_existent(session, q_get_alignment,
q_insert_alignment, "$alignment")
# insert the species entity (if it doesn't already exist)
if (len(alignment.hit_def.split("[")) > 1):
species = alignment.hit_def.split("[")[1].split("]")[0]
q_get_species = 'match $species isa species, has name "' + species + '"; get $species;'
q_insert_species = 'insert $species isa species, has name "' + species + '"; '
species_id = insert_if_non_existent(session, q_get_species,
q_insert_species, "$species")
# insert protein-ownership relationship (if it doesn't already exist)
q_get_protein_ownership = (
'match $sp id "' + species_id + '"; ' + '$pr id ' +
protein_id + '; ' +
'$pr-ownership (species-owner: $sp, owned-protein: $pr) isa protein-ownership;'
+ 'get $pr-ownership;')
q_insert_protein_ownership = (
'match $sp id "' + species_id + '"; ' + '$pr id ' +
protein_id + '; ' +
'insert $pr-ownership (species-owner: $sp, owned-protein: $pr) isa protein-ownership;'
)
insert_if_non_existent(session, q_get_protein_ownership,
q_insert_protein_ownership, "$pr-ownership")
species = protein_details[2]
# insert the protein entity
q_insert_protein = ("insert $pr isa protein " +
'has identifier "' + identifier + '" ' +
'has name "' + name + '" ' + 'has sequence "' +
sequence + '";')
protein_id = insert_anyway(session, q_insert_protein)
# insert the sourcing-of-information relationship
q_insert_sourcing_of_information = (
'match $pr id ' + protein_id + '; ' + '$db id ' + db_id +
'; ' +
"insert (information-source: $db, sourced-information: $pr) isa sourcing-of-information;"
)
insert_anyway(session, q_insert_sourcing_of_information)
# insert the species entity (if it doesn't already exist)
q_insert_species = 'insert $species isa species has name "' + species + '"; '
species_id = insert_if_non_existent(session, q_insert_species,
"$species")
# insert protein-ownership relationship (protein <> species)
q_insert_protein_ownership = (
'match $species id ' + species_id + '; ' + '$protein id ' +
protein_id + '; ' +
"insert (species-owner: $species, owned-protein: $protein) isa protein-ownership;"
)
insert_anyway(session, q_insert_protein_ownership)
print("- - - - - - - - - - - - - - - - -")
def init(data_path):
"""
1. creates a Grakn session to talk to the 'proteins' keyspace
2. inserts the database entity named 'UniProt'
3. for each protein stored in target-protein-sequences.fasta, inserts the:
- protein entity
- species entity
- species <> protein relationship
- protein <> database relationship
"""
with GraknClient(uri="localhost:48555") as client:
with client.session(keyspace="blast") as session:
# insert the database entity
q_get_database = 'match $db isa database, has name "uniprot"; get $db;'
q_insert_database = 'insert $db isa database, has name "uniprot";'
database_id = insert_if_non_existent(session, q_get_database,
q_insert_database, "$db")
with open(data_path) as data:
for first_line, sequence in SimpleFastaParser(data):
# extra relevant edata from first_line of each fasta (protein)
protein_details = re.split(',| OS=| OX=',
first_line.replace(' ', ',', 1))
identifier = protein_details[0].split("|")[1]
name = protein_details[1]
species = protein_details[2]
# insert the protein entity
q_get_protein = ('match $pr isa protein ' +
', has identifier "' + identifier + '" ' +
', has name "' + name + '" ' +
', has sequence "' + sequence + '"; ' +
'get $pr;')
q_insert_protein = ('insert $pr isa protein ' +
', has identifier "' + identifier +
'" ' + ', has name "' + name + '" ' +
', has sequence "' + sequence + '";')
protein_id = insert_if_non_existent(
session, q_get_protein, q_insert_protein, "$pr")
# insert the sourcing-of-information relationship
q_get_sourcing_of_information = (
'match $pr id ' + protein_id + '; ' + '$db id ' +
database_id + '; ' +
'$sourcing (information-source: $db, sourced-information: $pr) isa sourcing-of-information; '
+ 'get $sourcing;')
q_insert_sourcing_of_information = (
'match $pr id ' + protein_id + '; ' + '$db id ' +
database_id + '; ' +
'insert $sourcing (information-source: $db, sourced-information: $pr) isa sourcing-of-information;'
)
insert_if_non_existent(session,
q_get_sourcing_of_information,
q_insert_sourcing_of_information,
"$sourcing")
# insert the species entity
q_get_species = 'match $species isa species, has name "' + species + '"; get $species;'
q_insert_species = 'insert $species isa species, has name "' + species + '";'
species_id = insert_if_non_existent(
session, q_get_species, q_insert_species, "$species")
# insert protein-ownership relationship
q_get_protein_ownership = (
'match $species id ' + species_id + '; ' +
'$protein id ' + protein_id + '; ' +
'$pr-ownership (species-owner: $species, owned-protein: $protein) isa protein-ownership; '
+ 'get $pr-ownership;')
q_insert_protein_ownership = (
'match $species id ' + species_id + '; ' +
'$protein id ' + protein_id + '; ' +
'insert $pr-ownership (species-owner: $species, owned-protein: $protein) isa protein-ownership;'
)
insert_if_non_existent(session, q_get_protein_ownership,
q_insert_protein_ownership,
"$pr-ownership")