This repository is under active development, many backwards incompatible changes are likely.

taxa-assign-benchmarking

This repo is a Snakemake workflow for benchmarking taxonomy assignments. The goal of this project is to reduce the amount of work put into setting up and formatting results from taxonomic profilers.

Currently, the input depends on having CAMISIM formatted reads folders as input.

Install

There are no releases of this software currently. It can be installed via git and bash in the following way:

git clone https://github.com/gwarmstrong/taxa-assign-benchmarking.git
cd taxa-assign-benchmarking
conda env create -n taxa-benchmarking -f ci/requirements.txt
conda activate taxa-benchmarking
pip install -e .

Usage

TODO: better explanations of file hierarchies

This tool makes use of snakemake's CLI. You can begin by directly copying a simulation's output from CAMISIM into the data/simulations/ directory. Additionally, if there are any methods that you have pre-computed the profile for, you can put them into the data/profiles/ directory. Once the input files are configured appropriately, you can run the snakemake workflow with a command similar to the following on a cluster:

snakemake -p --use-conda \
    --cluster-config cluster.json \
    --cluster "qsub -l mem={cluster.mem_gb}gb -l walltime={cluster.time} -o cluster -e cluster" \
    -j 10 \
    --latency-wait 120 \
    --configfile config.yaml

The command following --cluster may differ depending upon your specific cluster.

Example config.yaml and cluster.json files are included in the base directory of this repo.