beast-skygrid 0.3.3

No project description provided

Skygrid

Skygrid is a Snakemake-based workflow tool designed for performing demographic analyses using BEAST’s skygrid model. It streamlines the entire pipeline—from sequence alignment parsing and outlier detection through BEAST MCMC sampling to post-analysis visualizations—so that you can focus on interpreting your results rather than managing complex workflows.

Features

• End-to-End Workflow:
  Automatically parses sequence alignments, performs root-to-tip regressions to detect outliers, executes BEAST analyses, and renders demographic plots and trees.

• Customizable Configuration:
  Easily override workflow defaults by supplying a configuration file or specifying options on the command line. Customize clock models (strict/relaxed), chain lengths, sample sizes, and more.

• Robust Resource Management:
  Leverages Snakemake’s advanced resource handling including multi-core execution, workflow locking, and DAG generation.

• Seamless Integration:
  Integrates tools such as IQ-TREE, treetime, and R (via ggtree) to offer a comprehensive analysis suite.

Installation

Skygrid is available on PyPI. Install it using pip:

pip install beast-skygrid

Quick Start

1. Prepare Your Data

Ensure you have a FASTA alignment file. This file is used to extract taxon names and sampling dates for your analysis.

2. Run the Workflow

Invoke the skygrid workflow from the command line. For example:

skygrid run --alignment your_alignment.fasta --output-dir results

This command will:

• Extract Taxa & Dates: Parse your alignment to determine the number of taxa and the sampling date range.
• Root-to-Tip Regression & Outlier Detection: Use root-to-tip regression to estimate the skygrid length and transition points. Optionally, filter out outlier sequences.
• BEAST Analysis: Generate a BEAST XML file from a Jinja2 template and run BEAST with the skygrid model.
• Visualization: Create skygrid plots and render the Maximum Clade Credibility (MCC) tree in SVG format.

3. Explore the Results

After completion, check the specified output directory (here, results). You will find:

• A skygrid plot (SVG, PNG, and PDF formats)
• BEAST logs and tree files
• Rendered MCC tree visualizations

Command Line Options

To display the complete list of options, run:

skygrid run -h

A summary of some key options:

• Workflow Configuration Options:
  • --alignment (-a PATH) (required)
    Path to your sequence alignment file.
  • --output-dir (-o PATH)
    Directory to save the output (default: ./skygrid).
  • --constant-sites TEXT
    Provide constant site counts in the format 'A,C,G,T'.
  • --discard-outliers
    Use a root-to-tip regression to identify and discard outlier sequences.
  • --clock [strict|relaxed]
    Choose the clock model for the analysis (default: strict).
  • --relaxed-gamma-shape FLOAT and --relaxed-gamma-scale FLOAT
    Parameters for the relaxed clock’s gamma distribution (default: shape=0.3, scale=0.001).
  • --transition-points-per-year FLOAT
    Set the number of transition points per year for the skygrid (default: 2).
  • --cutoff INTEGER
    Specify the skygrid length in years (if not provided, it is estimated from the data).
  • --chain-length INTEGER
    Length of the MCMC chain (default: 10,000,000).
  • --samples INTEGER
    Number of samples to extract from the chain (default: 10,000).
  • --sample-from-prior
    Run the analysis by sampling from the prior distribution only.
  • --beast-params TEXT
    Additional parameters to pass to BEAST (default: -overwrite).

Workflow Overview

The skygrid workflow is divided into several key stages:

1. Taxa Extraction & Date Parsing

The workflow begins by parsing your FASTA alignment to extract taxon information and sampling dates. Dates must be in the format >SampleName|YYYY-MM-DD or . Uncertain dates can be specified as >SampleName|YYYY-MM-DD/uncertainty where uncertainty is a number of years e.g. >SampleName|2020-01-01/1 includes 2020 to 2021.

2. Root-to-Tip Regression & Outlier Detection

The workflow uses root-to-tip regression (via treetime) to estimate the skygrid length (cutoff) and transition points (dimensions). Skygrid optionally uses a root-to-tip regression to identify and filter out outlier sequences.

3. BEAST Analysis

The workflow automatically generates a BEAST XML configuration file from a Jinja2 template. This XML file is used to run the BEAST analysis with parameters such as clock model, chain length, and sampling frequency. The BEAST run is executed as follows:

rule beast:
    input:
        beast_XML_file = OUTDIR / "beast" / "skygrid.xml",
    output:
        log_file = OUTDIR / "beast" / "skygrid.log",
        trees_file = OUTDIR / "beast" / "skygrid.trees",
    shell:
        """
        beast {params.beast} -working {input.beast_XML_file} > {log} 2>&1
        """

4. Post-Processing & Visualization

After BEAST completes, the workflow:

• Computes the Maximum Clade Credibility (MCC) tree using treeannotator.
• Renders the MCC tree in SVG format with R (using scripts that leverage the ggtree package).
• Generates skygrid plots (SVG, PNG, PDF) to visualize changes in population size over time.

Contributing

Contributions to skygrid are welcome! If you would like to propose improvements or bug fixes, please:

1. Fork the repository.
2. Create a feature branch for your changes.
3. Submit a pull request with a detailed description of your changes.

For major changes, please open an issue first to discuss what you would like to change.

License

Skygrid is distributed under the terms of the MIT License.