import pprint
import pandas as pd
from obnb.alltypes import List, Optional
from obnb.data.annotation.base import BaseAnnotationData
[docs]class GeneOntologyAnnotation(BaseAnnotationData):
"""Gene Ontology annotations.
Gene ontology annotations are retrieved from geneontology.org.
There are sevone categories of gene annotation evidences from the Gene
Ontology. By default, we only use *Experimental evidences*, *Author
Statements*, and the *Curator inferred* evidence types to ensure the
quality of the annotations. See the gene ontology `evidence codes
<http://geneontology.org/docs/guide-go-evidence-codes/>`_ documentation
page for more information.
- Experimental evidences (``EXPERIMENTAL``):
- ``EXP`` Experiment
- ``IDA`` Direct assay
- ``IPI`` Physical interaction
- ``IMP`` Mutant phenotype
- ``IGI`` Genetic interaction
- ``IEP`` Expression pattern
- Phylogenetically-inffered (``PHYLOGENIC``):
- ``IBA`` Biological aspect of ancestor
- ``IBD`` Biological aspect of descendant
- ``IKR`` Key residues
- ``IRD`` Rapid divergence
- Computational analysis (``COMPUTATIONAL``):
- ``ISS`` Sequence or structural similarity
- ``ISO`` Sequence orthology
- ``ISA`` Sequence alignment
- ``ISM`` Sequence model
- ``IGC`` Genomic context
- ``RCA`` Reviewed computational analysis
- Author statements (``AUTHOR``):
- ``TAS`` Tracable author statement
- ``NAS`` Nontracable author statement
- Curator statements (``CURATOR``):
- ``IC`` Inferred by curator
- ``ND`` No biological data available
- Electronic annotation evidences (``ELECTRONIC``):
- ``IEA`` Electronic annotation
**[Last updated: 2023-03-10]**
Args:
root: Root directory of the data.
data_sources: List of evidene types to be considered. If not set,
then use the default channels (experimental evidences, author and
curator statements).
"""
annotation_file_name = "goa_human.gaf"
annotation_url = "http://geneontology.org/gene-associations/goa_human.gaf.gz"
def __init__(
self,
root: str,
*,
data_sources: Optional[List[str]] = None,
**kwargs,
):
"""Initialize GeneOntology annotation data object."""
self._data_sources = data_sources
super().__init__(root, **kwargs)
@property
def data_sources(self) -> List[str]:
if self._data_sources is None:
return [
"EXP", # Experiment
"IDA", # Direct Assay
"IPI", # Physical Interaction
"IMP", # Mutant Phenotype
"IGI", # Genetic Interaction
"IEP", # Expression Pattern
"TAS", # Traceable Author Statement
"NAS", # Non-traceable Author Statement
"IC", # Inferred by Curator
]
else:
return self._data_sources
[docs] def load_processed_data(self):
path = self.raw_file_path(0)
self.plogger.info(f"Loading raw annotation from {path}")
# Load gene annotation data (gaf-version: 2.2)
# http://geneontology.org/docs/go-annotation-file-gaf-format-2.2/
annot_df = pd.read_csv(
path,
sep="\t",
comment="!",
header=0,
names=[
"db",
"db_id",
"db_symbol",
"qual",
"go_id",
"db_ref",
"ec",
"wof",
"aspect",
"eb_name",
"db_syn",
"db_type",
"taxon",
"date",
"assigned_by",
"annot_ext",
"gene_prod_id",
],
low_memory=False,
)
# Select specified channels
evidence_str = pprint.pformat(self.data_sources)
self.plogger.info(f"Subsetting annotations to evidences:\n{evidence_str}")
ind = annot_df["ec"].isin(self.data_sources)
self.plogger.info(f"{ind.sum():,} (out of {ind.shape[0]:,}) entries selected")
annot_df = annot_df[ind]
# Convert gene ids
gene_id_converter = self.get_gene_id_converter()
gene_id_converter.map_df(annot_df, "db_id", "gene_id")
annot_df["term_id"] = annot_df["go_id"]
# Save attributes
self.data = annot_df[["gene_id", "term_id"]].copy()