synthaser 1.1.8

Automatic prediction and classification of protein domain architectures

synthaser

Process

synthaser parses the results of a batch NCBI conserved domain search and determines the domain architecture of secondary metabolite synthases.

Installation

Install from PyPI using pip:

$ pip install --user synthaser

or clone the repo and install locally:

$ git clone https://www.github.com/gamcil/synthaser
$ cd synthaser
$ pip install .

Dependencies

synthaser is written in pure Python (3.6+), and requires only the following dependencies for remote searches:

• requests, for interaction with the NCBI's CD-Search API
• biopython, for retrieving sequences from NCBI Entrez

If you want to do local searches, you'll need:

• RPS-BLAST, for performing local domain searches
• rpsbproc, for formatting RPS-BLAST results like CD-Search

These can be obtained from the NCBI FTP.

Usage

A full synthaser search can be performed as simply as:

$ synthaser -qf sequences.fasta

Where sequences.fasta is a FASTA format file containing the protein sequences that you would like to search.

For a full listing of available arguments, enter:

$ synthaser -h

Visualising your results

synthaser is capable of generating fully-interactive, annotated visualisations so you can easily explore your results. All that is required is one extra argument:

$ synthaser -qf sequences.fasta -p

This will generate a figure like so:

Click here to play around with the full version of this example.

Saving your search session

synthaser allows you to save your search results such that they can be easily reloaded for further visualisation or exploration without having to fully re-do the search.

To do this, use the --json_file command:

$ synthaser -qf sequences.fasta --json_file sequences.json

This will save all of your results, in JSON format, to the file sequences.json. Then, loading this session back into synthaser, is as easy as:

$ synthaser --json_file sequences.json ...

Using your own rules

Though synthaser was originally designed to analyse secondary metabolite synthases, it can easily be repurposed to analyse the domain architectures of any type of protein sequence.

Under the hood, synthaser uses a central rule file which contains:

1. Domain types, containing specific families to save in CD-Search results, corresponding to domain 'islands';
2. Rules for classifying the sequences based on domain architecture predictions; and
3. A hierarchy which determines the order of evaluation for the rules.

We distribute our fungal megasynthase rule file as the default, but providing your own rule file is as simple as:

$ synthaser -qf sequences.fasta --rule_file my_rules.json

We also provide a web application for assembling your own rule files, which can be found here.

For a detailed explanation of how the rule file works, as well as API documentation, please refer to the documentation.

Citations

If you found synthaser helpful, please cite:

1. <pending>