catomatic 0.1.0

A tool for automatically building catalogues of antibiotic resistance-associated variants

catomatic

Python code that algorithmically builds antimicrobial resistance catalogues of mutations.

Introduction

This method relies on the logic that mutations that do not cause resistance can co-occur with those that do, and if a mutation in isolation (solo) does not cause resistance, then it will also not contribute to the phenotype when not in isolation.

Mutations that occur in isolation across specified genes are traversed in sequence, and if their proportion of drug-susceptibility (vs resistance) passes the specified statistical test, they are characterized as benign and removed from the dataset. This step repeats while there are susceptible mutations in isolation. Once the dataset has been 'cleaned' of benign mutations, resistant mutations are classified via their proportions by the specified test, failing which they are added to the catalogue as 'Unclassified'.

Construction can either rely on homogenous susceptibility for the particular mutation (and no explicit phenotyping is carried out, other than to unlock susceptible variants), use a Binomial test where the proportion of resistance is tested against a specified background rate, or a Fisher's test where the proportion of resistance is tested against a calculated background rate.

Although the method is entirely algorithmic, there are 2 entry points for intervention. Firstly, one is able to 'seed' the method with neutral mutations (such as those gathered in a literature search - often helpful if a gene contains phylogenetic mutations with high prevalence that add noise), and secondly one can add or overwrite classifications and entries to the catalogue, although not recommended unless aggregating.

Because the method uses and understands GARC1 grammar, one can supply 'rules' to the catalogue post-hoc - such as {rpoB@*_fs:R} for frameshifts in rpoB, which can either simply be added (and would have lower prediction priority to finer grain mutations, such as rpoB@44_ins) or can replace any mutations that fall under that rule, effectively aggregating relevant variants.

The generated catalogue can be returned either as a dictionary, or as a Pandas dataframe which can be exported in a Piezo compatible format for rapid parsing and resistance predictions.

Contingency tables, proportions, p_values, and Wilson's Confidence Intervals are logged under the 'EVIDENCE' column of the catalogue.

A workflow with example parameters:

Installation

Using Conda

It is recommended to manage the Python environment and dependencies through Conda. You can install Catomatic within a Conda environment by following these steps:

Create and Activate Environment

First, ensure that you have Conda installed. Then, create and activate a new environment, and install catomatic:

conda env create -f env.yml
conda activate catomatic
pip install .

Running catomatic

At the most basic level, the method takes 2 input dataframes: a samples dataframe which contains 1 row per sample with 'R' vs 'S' binary phenotypes, and a mutations dataframe which contains 1 row per mutation. They have to be joinable on their UNIQUEID columns.

If exporting to Piezo format, the MUTATION column must contain GARC1 grammer (ie gene@mutation). One must also supply a path to the wildcards.json file, which should contain Piezo wildcards in a json object/dictionary (example file in /data/bdq_wildcards.json).

If seeding or updating the catalogue, the mutation grammar must match that of the MUTATION column.

CLI

After installation, the simplest way to run the catomatic catalogue builder is via the command line interface. --to_piezo or --to_json flags will need to specified to save the catalogue (with additional arguments if using --to_piezo)

BuildCatalogue --samples path/to/samples.csv --mutations path/to/mutations.csv --to_json --outfile path/to/out/catalogue.json

or

BuildCatalogue --samples path/to/samples.csv --mutations path/to/mutations.csv --to_piezo --outfile path/to/out/catalogue.csv --genbank_ref '...' --catalogue_name '...' --version '...' --drug '...' --wildcards path/to/wildcards.json

Python/Jupyter notebook

Should you wish to run catomatic in a notebook, for example, you can do so simply by calling BuildCatalogue after import.

from catomatic.CatalogueBuilder import BuildCatalogue

#instantiate a catalogue object - this will build the catalogue
catalogue = BuildCatalogue(samples = samples_df, mutations = mutations_df)

#return the catalogue as a dictionary in order of variant addition
catalogue.return_catalogue()

#return the catalogue as a piezo-structured dataframe
catalogue.build_piezo(genbank_ref='...', catalogue_name='...', version='...', drug='...', wildcards='path/to/wildcards.json')

More detailed examples on running catomatic can be found in examples/demo.ipynb

CLI Parameters

Parameter	Type	Description
--samples	str	Path to the samples file. Required.
--mutations	str	Path to the mutations file. Required.
--FRS	float	Fraction Read Support threshold. Optional.
--seed	list	List of seed mutations using GARC grammar. Optional.
--test	str	Type of statistical test to run: None, Binomial, or Fisher. Optional.
--background	float	Background mutation rate for the binomial test. Required if using test = Binomial. Optional.
--p	float	Significance level for statistical testing. Optional. Defaults to 0.95.
--strict_unlock	bool	Enforce strict unlocking for classifications, which requires p < 0.05. Optional.
--to_json	bool	Export the catalogue to JSON format. Optional.
--outfile	str	Path to output file for exporting the catalogue. Used with --to_json or --to_piezo. Optional.
--to_piezo	bool	Export catalogue to Piezo format. Optional.
--genbank_ref	str	GenBank reference for the catalogue. Required if to_piezo = True. Optional.
--catalogue_name	str	Name of the catalogue. Required if to_piezo = True. Optional.
--version	str	Version of the catalogue. Required if to_piezo = True. Optional.
--drug	str	Drug associated with the mutations. Required if to_piezo = True. Optional.
--wildcards	str	JSON file with wildcard rules. Required if to_piezo = True. Optional.
--grammar	str	Grammar used in the catalogue. Optional. Defaults to GARC1.
--values	str	Values used for predictions in the catalogue. Optional. Defaults to RUS.