configure-dms-viz 1.5.0

Configure your data for visualization with dms-viz.github.io

configure_dms_viz

---

Overview

configure_dms_viz is a python command line utility that formats your data for the visualization tool dms-viz. dms-viz is a tool that helps you take quantitative data associated with mutations to a protein and analyze that data using intuitive visual summaries and an interactive 3D protein structure. Visualizations created with dms-viz are flexible, customizable, and shareable.

For more information on getting started with dms-viz, check out the documentation.

Prerequisites

To use configure-dms-viz, you must ensure that you have the correct version of Python (>= 3.9) installed on your operating system. You can check this by running:

python --version

The version number displayed should >= 3.9.x.

Installation

configure-dms-viz is distributed on PyPI, allowing you to install configure-dms-viz using pip. To install the latest version of configure-dms-viz, run the following command:

pip install configure-dms-viz

configure-dms-viz should now be installed. You can double-check that the installation worked by running the following command:

configure-dms-viz --help

You should see a help message printed to the terminal.

Basic Usage

configure_dms_viz takes input data consisting of a quantitative metric associated with mutations to a protein sequence and returns a .json specification file that is uploaded to dms-viz to create an interactive visualization. Below is a simple tutorial on configure-dms-viz; however, for a detailed guide to the configure-dms-viz API, check out the documentation.

configure-dms-viz has two commands, format and join. To format a single dataset for dms-viz, you execute the configure-dms-viz format command with the required and optional arguments as needed:

configure-dms-viz format \
    --name <experiment_name> \
    --input <input_csv> \
    --metric <metric_column> \
    --structure <pdb_structure> \
    --output <output_json> \
    [optional_arguments]

The information that is required to make a visualization file for dms-viz is as follows:

1. --name: The name of your dataset as you'd like it to appear in the visualization.
2. --input: The file path to your input data.
3. --metric: The name of the column that contains the metric you want to visualize.
4. --structure: The protein structure that you want to use as a 3D model.
5. --output: The file path of the output .json file.

The remaining arguments are optional and configure the protein, appearance, and data included in your final visualization.

Now, let's use configure-dms-viz with a minimal example. The example data is included in this GitHub repository under tests/. If you want to follow along, clone the repository and run the following command from the root of the directory.

Input

configure-dms-viz format \
   --name "REGN mAb Cocktail" \
   --input tests/SARS2-RBD-REGN-DMS/input/REGN_escape.csv \
   --metric "mut_escape" \
   --metric-name "Escape" \
   --sitemap tests/SARS2-RBD-REGN-DMS/sitemap/sitemap.csv \
   --structure "6XDG" \
   --included-chains "E" \
   --condition "condition" \
   --condition-name "Antibody" \
   --output ./REGN_escape.json

First, we've specified that we want the name of the dataset as it appears in dms-viz to be REGN mAb Cocktail (named after the Regeneron Antibody cocktail therapuetic for SARS-CoV-2). This isn't so crucial when there is only a single dataset; however, when combining multiple datasets with the join command, it's necessary to have unique and descriptive names.

Next, we've pointed to the input data containing quantitative scores that measure the degree of antibody escape from the REGN mAb Cocktail. For details on the specific requirements for input data, check out the Data Requirements guide in the documentation. In addition to specifying the input data, we told configure-dms-viz which column contains the escape scores (mut_escape) and what to call that column in the plots (Escape).

Then, we've specified a sitemap. This is optional information that describes how the sites in your input data correspond to your 3D protein structure. If you do not provide a sitemap, the sites in the input data are assumed to correspond one-to-one with the sites in the protein structure.

After that, we specified a protein structure. In this case, we're fetching 6XDG from the RSCB PDB and only showing our data on chain E of that structure.

Finally, in this particular dataset, we have multiple 'conditions' for each mutation; this means there are multiple measurements (mut_escape) for each mutation/position (corresponding to escape from different antibodies). We need to specify the column that contains these conditions. In dms-viz, an interactive legend will let you toggle between conditions.

The result of this command should be a message printed to the terminal providing some basic information from the configure-dms-viz format command that looks like this:

Output

Formatting data for visualization using the 'mut_escape' column from 'tests/SARS2-RBD-REGN-DMS/input/REGN_escape.csv'...

Using sitemap from 'tests/SARS2-RBD-REGN-DMS/sitemap/sitemap.csv'.

'protein_site' column is not present in the sitemap. Assuming that the reference sites correspond to protein sites.

About 95.98% of the wildtype residues in the data match the corresponding residues in the structure.
About 4.02% of the data sites are missing from the structure.

Success! The visualization JSON was written to './REGN_escape.json'

That's how you use configure-dms-viz to format a single dataset! You can also combine multiple datasets into a single .json specification file using the configure-dms-viz join command. For more details on combining datasets to jointly visualize with dms-viz, check out the API.

Developing

configure-dms-viz was developed using Python (>=3.9) and the click library.

To contribute to configure-dms-viz, follow the instructions here for setting up a development environment.

Testing

pytest is the testing framework for configure-dms-viz.

The command line interface (CLI) of configure-dms-viz is tested using four example datasets from different projects and labs that cover 100% of its flags and features. These four examples are:

1. Deep mutational scanning of the SARS-CoV-2 Spike protein Authors: Bernadeta Dadonaite, Katharine H D Crawford, Caelan E Radford, Ariana G Farrell, Timothy C Yu, William W Hannon, Panpan Zhou, Raiees Andrabi, Dennis R Burton, Lihong Liu, David D. Ho, Richard A. Neher, Jesse D Bloom Manuscript: https://www.sciencedirect.com/science/article/pii/S0092867423001034?via%3Dihub

2. Deep mutational scanning of the HIV BF520 strain Envelope protein Authors: Caelan E. Radford, Philipp Schommers, Lutz Gieselmann, Katharine H. D. Crawford, Bernadeta Dadonaite, Timothy C. Yu, Adam S. Dingens, Julie Overbaugh, Florian Klein, Jesse D. Bloom Manuscript: https://www.sciencedirect.com/science/article/pii/S1931312823002184?via%3Dihub

3. Phylogenetic fitness estimates of every SARS-CoV-2 protein Authors: Jesse D. Bloom, Richard A. Neher Manuscript: https://www.biorxiv.org/content/10.1101/2023.01.30.526314v2

4. Deep mutational scanning of the Influenza PB1 polymerse subunit Authors: Yuan Li, Sarah Arcos, Kimberly R. Sabsay, Aartjan J.W. te Velthuis, Adam S. Lauring Manuscript: https://www.biorxiv.org/content/10.1101/2023.08.27.554986v1.full

In addition to these test datasets, there are specific tests using dummy data for the key formatting functions. To run the tests, execute the following command from the root of the directory:

poetry run pytest