Setup

python -m pip install ontologize

Usage

Vignette: Genes

Suppose we have a list of genes, perhaps ones that are upregulated in a certain environment, and we wish to understand the functional changes in the cell.

We can first build an Ontology object from a list of the genes' BioCyc IDs:

from ontologize.ontology import build_ontology

# cadA EG10131
# lacA EG10524 
# xylA EG11074 
ont = build_ontology(objects=["EG10131", "EG10524", "EG11074"], schema_type="Gene")

Ontology objects store an annotated ontology graph, as a networkX DiGraph:

import networkx as nx
assert isinstance(ont.graph, nx.DiGraph)

Rich printing options are supported, including truncation of the graph at a given depth, inclusion/exclusion of leaf nodes, whether to color by depth.

print(ont.to_string(max_depth=None, include_leaves=False, colors=True))

In this example, we see that lacA and xylA are both involved in carbon utilization, while cadA is related to pH adaptation.

Command-Line Interface

Once exposed, ontologize exposes a runnable script, and can also be called as a module:

ontologize <file> <schema_type> [flags]
python -m ontologize <file> <schema_type> [flags]

The required arguments are given as follows:

• file: Path to a .csv, .tsv, or .xlsx file with BioCyc object IDs to ontologize. By default, assumes a (header-less, if .csv or .tsv) first column containing the IDs to be ontologized. If a .xlsx file is given, then by default, IDs are assumed to be in the first sheet in the first column, treating the first entry as a header.

• schema_type : Type of the objects (or properties) to be ontologized in the Biocyc Schema. For example, this might be Gene, Pathway, Compound, etc.

Note that schema_type uses the singular form of the class name!

Example:

# TODO

Flags

Ontology-building options:

• -s <sheet_name>, --sheet <sheet_name>: For a .xlsx file, the name of the sheet containing BioCyc IDs. Ignored if file is not a .xlsx file.
• -o <objects>, --objects <objects>: For a multi-column file, the name of the column containing BioCyc IDs for the objects to ontologize. Requires a header row containing column names.
• -p <objects>, --property <objects>: For a multi-column file, the name of the column containing BioCyc IDs for the property to ontologize. Requires a header row containing column names. When using this option, the objects must also be specified using the -o option.

• WARNING: -p, --property NOT YET IMPLEMENTED

• --database <orgid>: BioCyc organism ID, used to specify the organism-specific database within to search. ECOLI by default.

Printing options:

• --depth <depth>: Maximum depth of the ontology to print. No limit by default.
• --leaves: Whether to show leaf nodes, i.e., the ontologized objects themselves. Not shown by default.
• --coloroff: Turns off colorful printing.

TODO: graph options (not implemented), pkl options, --interactive (allows maintaining session)

References

[BioCyc19] Karp, P.D., et al., The BioCyc collection of microbial genomes and metabolic pathways Briefings in Bioinformatics (2019).