Circling in on plasmids: benchmarking plasmid detection and reconstruction tools for short-read data from diverse species

Supporting analysis code for Circling in on plasmids: benchmarking plasmid detection and reconstruction tools for short-read data from diverse species.

The predictions of plasmid detection and reconstruction tools, assembly statistics, other plasmid information (such as plasmid size, ARG presence/absence and the number of transposases), and ANI between chromosomes and plasmids are in ./data/.

Python scripts used for analysis are in ./scripts/ and include:

| Category | Script | Description | | -------- | ------- | ----------- | | Dataset | dataset.args.fragmentation.py | Compares the number of IS, SR contigs, and plasmid length between plasmids with ARGs and plasmids without ARGs | | Dataset | dataset.plasmidlength.py | Plots the distribution of plasmid lengths in samples with complete hybrid assemblies | | Dataset | dataset.ani.py | Plots heatmaps of ANI between chromosomes and plasmids. Also clusters plasmids based on their alignment fraction | | Dataset | dataset.plsdb.py | Compares the Mash identity to the closest PLSDB plasmids for different taxa | | Detection | detection.metrics.py | Calculates plasmid detection metrics from a set of predictions and plots the results | | Detection | detection.glm.py | Fits a Logistic Regression model to estimate the contribution of certain plasmid and assembly features to plasmid detection | | Detection | detection.srcontiglength.py | Compares the plasmid detection metrics for SR contigs of different lenghts | | Detection | detection.repcluster.py | Compares the plasmid detection metrics for plasmids of different Inc types/rep clusters | | Detection | detection.args.py | Compares the plasmid detection metrics for SR contigs with and without ARGs | | Detection | detection.plasmidsize.py | Compares the plasmid detection metrics for plasmids of different sizes (large vs. small) | | Reconstruction | reconstruction.metrics.py | Calculates plasmid reconstruction metrics from a set of predictions and plots the results | | Reconstruction | reconstruction.glm.py | Fits a Linear Regression model to estimate the contribution of certain assembly features to plasmid reconstruction | | Reconstruction | reconstruction.metricsbest_detector.py | Calculates plasmid reconstruction metrics from a set of predictions, using initial contig classifications from the best plasmid detection tools, and plots the results | | Reconstruction | reconstruction.args.py | Compares plasmid reconstruction metrics for plasmids with and without ARGs |

You can run each script individually or all at once with the script in ./data/run-everything.

The exact package versions in machine-readable format used to generate the results presented in the manuscript are in pkgs.versions.txt.

Usage

1. Clone this repository and navigate into it:

   cd plasmid-detection-benchmark

2. Create a new virtual environment and install the required packages. If you are using conda:

   conda create --name plasmid-env --file requirements.yaml conda activate plasmid-env

3. Run all scripts with: sh scripts/run-everything Or run individual scripts with: python scripts dataset.ani.py You can check the command line arguments for each script withpython script.name.py --help`.

Outputs

By default, the output files are written to ./outputs/. If running the run-everything script, the result of each individual script will be written to a separate directory within ./outputs/.

Authors

Marco Teixeira (mcarvalh@broadinstitute.org), Celia Souque, Colin J. Worby, Terrance Shea, Nicoletta Commins, Joshua T. Smith, Arjun M. Miklos, Thomas Abeel, Ashlee M. Earl, and Abigail L. Manson.