Basis¶

Query Primary databases¶

pyeed has a set of build in query methods which allow retrieving data from the primary databases such as UniProt and NCBI. Using the fetch_from_primary_db or fetch_nucleotide_from_db methods, either protein or nucleotide sequences can be retrieved from the primary databases. Besides the sequence, also Gene Ontology, Organism, region, and site annotations are retrieved and added to the database as defined in the pyeed graph model.

In [ ]:

from pyeed import Pyeed

uri = "bolt://127.0.0.1:7687"
user = "neo4j"
password = "12345678"

# Create a Pyeed object, automatically connecting to the database
eedb = Pyeed(uri, user, password)


# UniProt IDs of proteins to be fetched from primary database
ids = [
    "P04182",
    "Q6QDP7",
    "P04182",
    "P29758",
    "A0A1G4JJF2",
    "G8ZTZ5",
    "A0A1G4MBD6",
    "A0A7H9HSJ3",
    "J7SA96",
    "G0VK69",
]

# Fetch proteins from primary database
eedb.fetch_from_primary_db(ids)

# number of nodes and edges in db
print(eedb.db.stats())

from pyeed import Pyeed uri = "bolt://127.0.0.1:7687" user = "neo4j" password = "12345678" # Create a Pyeed object, automatically connecting to the database eedb = Pyeed(uri, user, password) # UniProt IDs of proteins to be fetched from primary database ids = [ "P04182", "Q6QDP7", "P04182", "P29758", "A0A1G4JJF2", "G8ZTZ5", "A0A1G4MBD6", "A0A7H9HSJ3", "J7SA96", "G0VK69", ] # Fetch proteins from primary database eedb.fetch_from_primary_db(ids) # number of nodes and edges in db print(eedb.db.stats())

Sequence embeddings¶

The sequence embeddings are generated using Meta's ESM2 (esm2_t33_650M_UR50D) model. By default, embeddings are calculated in batches of 16 sequences and subsequently added to the database after the calculation.

In [ ]:

# Calculate embeddings for all sequences in the database
eedb.calculate_sequence_embeddings()

# Calculate embeddings for all sequences in the database eedb.calculate_sequence_embeddings()

Visualize what's in the DB¶

To use the web interface, open a browser and go to http://localhost:7474/.

Query DB¶

In [4]:

from pyeed.model import Protein

## Query using pyeed graph objects
# Get all proteins
proteins = Protein.nodes.all()
print("Number of proteins in database: ", len(proteins))

# Get protein with id P04182
protein = Protein.nodes.get(accession_id="P04182")
print(
    f"id: {protein.accession_id} | mol weight: {protein.mol_weight} | ec: {protein.ec_number} | seq length: {protein.seq_length}"
)

## Or execute cypher query
# Get all organisms that have at least one connected proteins
query = """
MATCH (o:Organism)<-[:ORIGINATES_FROM]-(p:Protein)
WITH o, COUNT(p) AS proteinCount
WHERE proteinCount >= 1
RETURN o
"""

organisms = eedb.db.execute_read(query)
print("Number of organisms with at least one protein: ", len(organisms))

# Get 5 most similar proteins to protein with accession_id P04182 based on sequence embedding
query = """
MATCH (p:Protein {accession_id: 'P04182'})
CALL db.index.vector.queryNodes('vector_index_Protein_embedding', 5, p.embedding)
YIELD node AS similarProtein, score
RETURN similarProtein.accession_id AS accession_id, 
       similarProtein.name AS protein_name, 
       score
ORDER BY score DESC
"""

similar_proteins = eedb.db.execute_read(query)
similar_proteins

from pyeed.model import Protein ## Query using pyeed graph objects # Get all proteins proteins = Protein.nodes.all() print("Number of proteins in database: ", len(proteins)) # Get protein with id P04182 protein = Protein.nodes.get(accession_id="P04182") print( f"id: {protein.accession_id} | mol weight: {protein.mol_weight} | ec: {protein.ec_number} | seq length: {protein.seq_length}" ) ## Or execute cypher query # Get all organisms that have at least one connected proteins query = """ MATCH (o:Organism)<-[:ORIGINATES_FROM]-(p:Protein) WITH o, COUNT(p) AS proteinCount WHERE proteinCount >= 1 RETURN o """ organisms = eedb.db.execute_read(query) print("Number of organisms with at least one protein: ", len(organisms)) # Get 5 most similar proteins to protein with accession_id P04182 based on sequence embedding query = """ MATCH (p:Protein {accession_id: 'P04182'}) CALL db.index.vector.queryNodes('vector_index_Protein_embedding', 5, p.embedding) YIELD node AS similarProtein, score RETURN similarProtein.accession_id AS accession_id, similarProtein.name AS protein_name, score ORDER BY score DESC """ similar_proteins = eedb.db.execute_read(query) similar_proteins

Number of proteins in database:  44
id: P04182 | mol weight: 48333.0 | ec: 2.6.1.13 | seq length: 439
Number of proteins associated with GO:0005739:  11
Number of organisms with at least two proteins:  9

Manually add a Protein to the DB¶

Besides automatically fetching data from the primary databases, it is also possible to manually add a protein to the database. This can be done by creating a Protein object and adding it to the database.

For more examples on query and adding data to the database, refere to the neomodel documentation.

In [ ]:

my_protein = Protein(
    accession_id="12345!",
    name="my_protein",
    seq="MYAYAYAYA",
    seq_length=9,
    mol_weight=2,
)

my_protein.save()

my_protein = Protein( accession_id="12345!", name="my_protein", seq="MYAYAYAYA", seq_length=9, mol_weight=2, ) my_protein.save()

In [8]:

# Close the connection to the database
eedb.db.close()

# Close the connection to the database eedb.db.close()

🔌 Connection closed.