protein_blast()

Find the best hit of each gene in the query cluster's proteins.

Every protein is aligned globally with every protein of the query cluster (Needleman-Wunsch, as R's protein_blast() with pwalign::pairwiseAlignment(); see aligner for the scores). A gene keeps the query protein with the highest identity as its hit, when the identity is at least identity; equal best hits give a row each. The clusters are then ordered by a synteny score, the query first.

Needs Biopython: pip install "geneviewer-py[align]".

Usage

from geneviewer import protein_blast

protein_blast(
    data,
    query,
    id='protein_id',
    start='start',
    end='end',
    cluster='cluster',
    genes=None,
    identity=30,
)

Arguments

Argument	Type	Description
data	pandas.DataFrame or path	Genes with their protein sequence (column translation or sequence), or a GenBank file or folder of them (one cluster per file).
query	str	The query cluster.
id	str	Column of unique protein ids.
start, end	str	Columns of the start and end positions.
cluster	str	Column of cluster names.
genes	list of str, optional	Only these genes of the query cluster (by id). Default: all.
identity	float	The lowest identity (percent) of a hit. Default 30.

Value

pandas.DataFrame: data (rows without an id, sequence or cluster left out) with rowID, start, end, and identity, similarity, BlastP (the id of the hit, "No Hit" without one; a hit only one gene has is "No Hit" too) and score (the cluster's synteny score). Ordered by cluster, the query first, then by score.

Errors

Error	When
ImportError	Without Biopython.
ValueError	When a column or the query is not in the data.

Examples

>>> genes = pd.DataFrame({
...     "protein_id": ["A1", "A2", "B1", "B2"],
...     "cluster": ["A", "A", "B", "B"],
...     "start": [1, 200, 1, 200], "end": [150, 350, 150, 350],
...     "translation": ["MKTAYIAKQRQISFVKSHFSRQ", "MSDNELLKRLGEAVLAHPTTG",
...                     "MKTAYIAKQRQISFVKSHFSRQ", "MLPGGRTSWVQEHCDKAYNIF"]})
>>> protein_blast(genes, query="A")[["protein_id", "identity", "BlastP"]]
  protein_id  identity  BlastP
0         A1     100.0      A1
1         A2     100.0  No Hit
2         B1     100.0      A1
3         B2       NaN  No Hit