An example Snakemake workflow using ppx and ANN-SoLo

ppx provides programmatic access to files stored in proteomics data repositories, such as PRIDE and MassIVE. In doing so, ppx enhances the reproducibility of proteomics analyses. Here, we provide an example Snakemake workflow, which in part reproduces the open modification search results from ANN-SoLo presented by Bittremieux et al.

This workflow uses ppx to download the MGF mass spectrometry data files and the spectral library from the original publication's project on PRIDE Archive, PXD009861. These MGF files are then searched using ANN-SoLo. Additionally, the workflow downloads the original ANN-SoLo result files described in the original publication (*.mztab). Finally, the workflow creates figures displaying the modification mass shifts detected by the different ANN-SoLo versions.

Prerequisites

You can run this workflow on MacOS and Linux systems, due to the requirements for ANN-SoLo. You'll also need the following software installed:

• A conda-based Python distribution - We recommend miniconda. This provides the conda package manager.
• git - git is a tool for managing versions of code. You'll need git to grab this workflow from here on GitHub.

You can confirm that both requirements are met by opening your terminal and trying the following commands:

conda

$ conda --version
conda 4.10.1

git

$ git --version
git version 2.31.1

Running the workflow

Downloading the workflow

We'll use git to download this workflow from GitHub:

$ git clone https://github.com/Noble-Lab/ppx-workflow.git
$ cd ppx-workflow

Alternatively, you can use click Code -> Download ZIP in the GitHub repository, unpack the zip archive, and navigate to the new ppx-workflow directory in your terminal.

Create a conda environment

We'll create a conda environment, which will automatically install Snakemake, ppx, and their dependencies:

$ conda env create -f environment.yaml
$ conda activate ppx-workflow

Execute the workflow

Now we'll use Snakemake to execute our workflow. Snakemake will manage the installation of ANN-SoLo and other dependencies, then run all of the steps required for our analyses.

In this example, we'll use --configfile config/one.yaml, which will restrict our analysis to a single MGF file:

$ snakemake -j -d workflow --use-conda --configfile config/one.yaml

Here is what each parameter does:

• -j: Use the maximum number of cores. You can also specify a number, such as -j4 to use 4 cores.
• -d workflow: Tells Snakemake that our working directory should be the workflow directory. This is where our Snakefile is.
• --use-conda: Tells Snakemake that we want it to handle setting up a conda environment, such as for ANN-SoLo.
• --configfile config/one.yaml: Specifies a config file, other than the default one. The config/one.yaml file changes the workflow to only download and search a single MGF file, instead of all of them.

Alternatively, you could analyze all of the MGF files as we did in the paper. Note that this could take a while depending on your system:

$ snakemake -j -d workflow --use-conda

Using a GPU

If you have an NVIDIA GPU available with at least 8 Gb of memory, you can use it during the ANN-SoLo searches in this workflow. By default, the workflow will only use the CPU. Enabling the GPU requires an additional parameter:

$ snakemake -j -d workflow --use-conda --config GPU=True

Note that we've setup the work flow to use a CUDA Toolkit version of 11.2. If this is not suitable for your system, you should adjust it in the ANN-SoLo GPU conda environment.

Conclusion

We've described how to run this workflow that uses ppx, ANN-SoLo, and Snakemake to reproduce a previously published analysis with data shared in PRIDE Archive. Our intention with this workflow is that it would be an illustration of how ppx can help create reproducible and shareable proteomics analyses.

To find out more about ppx, see our documentation: https://ppx.readthedocs.io