Analyses workflow for POCP benchmark manuscript

popcbenchmark_manuscript contains the workflow for analyzing data produced by our benchmark ClavelLab/pocpbenchmark. We set out to compare proteins alignment tools for improved genus delineation using the Percentage Of Conserved Proteins (POCP).

Graphical overview

Schematic overview of POCP and POCPu calculation and the overall benchmarking workflow. (A) Three simplified examples of bacterial genomes (two from the same genus and one from a different genus) to illustrate how the Percentage Of Conserved Proteins (POCP) or Percentage Of Conserved Proteins using unique matches (POCPu) were calculated. Details on the formula for POCP and POCPu calculation are provided in the Methods. (B) Overview of the benchmarking workflow. Shortlisted, high-quality genomes from the Genome Taxonomy Database (GTDB) were used to compare different protein alignment methods (top-right) and to evaluate the potential of POCP and POCPu for genus delineation (bottom-right).

Our work is published in PeerJ

Pauvert C, Hitch TCA, Clavel T. 2025. Fast and robust estimate of bacterial genus novelty using the percentage of conserved proteins with unique matches (POCPu) PeerJ 13:e20259 https://doi.org/10.7717/peerj.20259

Setup the environment for the workflow

These analyses were conducted in R 4.3.1 and in Rstudio. We recommend setting up R and specific versions using rig, and getting Rstudio from Posit. We also use renv for reproducible environment, which can be installed in R with install.packages("renv").

1. Open Rstudio and create a new project via "File > New Project..."
2. Select "Version Control" and then "Git"
   1. Type https://github.com/ClavelLab/pocpbenchmark_manuscript in Repository URL.
   2. Make sure the project is going to be created in the correct subdirectory on your computer, or else edit accordingly
   3. Click on "Create project"

If you comfortable with the command line and git, clone the repository either with SSH or HTTPS in a suitable location.

3. Rstudio warns you that One or more packages recorded in the lockfile are not installed because a couple of R packages and dependencies are needed.
   1. Install the dependencies by typing renv::restore() in the Console and agree to the installation of the packages.
   2. Check that all dependencies are set by typing renv::status() in the Console where you should have No issues found

Our analysis workflow is orchestrated by targets and is composed of two subworkflows.

Prepare the data for the analysis

Note

You can skip to the next section if you want to start the workflow from already prepared files!

1. Download the raw output files from the workflow using the "Download all" button: https://doi.org/10.5281/zenodo.14974869
2. Uncompress the zip archive within your project
3. Create a data_benchmark folder within your project.
4. Move all the zip files downloaded from zenodo (benchmark-gtdb-f__*.zip) to data_benchmark.
5. Ensure the two csv files are at the root of your project.
6. Run the workflow with the following command:

Sys.setenv(TAR_PROJECT = "prepare_pocpbenchmark_data")
targets::tar_make()

Analyze the data and build the manuscript

If you skipped the first workflow, you need to download the cleaned and formatted POCP/POCPu values and metadata tables for analysis from https://doi.org/10.5281/zenodo.14975029. These are the files you would have generated with the previous section.

1. Run the workflow with the following command:

Sys.setenv(TAR_PROJECT = "analyze_pocpbenchmark_data")
targets::tar_make()

The manuscript is then available in the _manuscript folder, both as a HTML document (index.html) and a docx document. The figures are generated in the figures folder.