MOODA: Multi-Objective Optimization for DNA design and assembly

Current version: 0.11.0

MOODA is a multi-objective optimisation algorithm for DNA sequence design and assembly.

It takes in input an annotated sequence in GenBank format, and optimize it with respect to user-defined objectives.

Currently, some of the most common common operations in synthetic biology are built-in, including:

• The GCOptimizationOperator introduces silent mutation in coding regions to obtain DNA constructs with a user-defined GC content.

• The CodonUsageOperator probabilistically recodes coding regions by probabilistically selecting the most frequent codon for an aminoacid in a host organism.

• The BlockJoin and BlockSplit operators allow the division of a sequence into fragments (or blocks). After the optimisation, each block is then adapted to the assembly method selected by the user. Currently, only the Gibson assembly is supported.

New operators, objective functions or assembly method can be added by extending the Operator, ObjectiveFunction and Assembly classes.

Installation

The easiest and fastest method to use mooda is using Docker:

    docker pull ghcr.io/stracquadaniolab/mooda

You can also install mooda using conda:

    $ conda install -c stracquadaniolab -c bioconda -c conda-forge mooda

or using pip:

    $ pip install mooda-dna

Please note, that pip will not install non Python requirements.

Getting started

A typical mooda analysis consists of 3 steps:

1. Select a DNA sequence in Genbank format.

2. Write a MOODA configuration file. A .yaml file defining operators, objective functions, assemblies strategy and their parameters.

3. Run MOODA.

Example: optimizing GC content, E. coli codon usage, number of fragments and the variance of their length

Create a test directory as follows:

    $ mkdir example-run

Move to your test directory as follows:

    $ cd example-run

Download test data from Github as follows:

    $ curl -LO https://github.com/stracquadaniolab/mooda/raw/master/examples/mooda-example1.tar.gz

Extract test data as follows:

    $ tar xvzf mooda-example1.tar.gz

Run mooda as follows:

    $ docker run -it --rm -v $PWD:$PWD -w $PWD ghcr.io/stracquadaniolab/mooda -i seq_5_5.gb  -c config.yaml -p 10 -it 20 -a 100 -mns 200 -mxs 2000 -bss 50 -js 40 -dir example-opt -gf True

Results will be available in the example-opt directory, where you will find:

• Genbank files of the Pareto optimal sequence.
• FASTA files with the fragments for Gibson assembly for each Pareto optimal sequence.
• _logfile.yaml file with information about the analysis.
• _mooda_result.csv file with objective function value information for each sequence.

Command line options

• -i: Input DNA sequence to process.

• -c: Configuration file to set operators, objective functions and their parameters.

• -p: Pool size. Number of candidate solutions sampled at each iteration. The pool size should increase with the length and complexity of the input sequence.

• -it: Number of iterations. The number of iterations should increase with the length and complexity of the input sequence, although it will take longer to run.

• -a: Archive size. The number of non-dominated solutions to store at each iteration, which allows to use smaller pools for improved efficiency.

• -mns: Block minimum size.

• -mxs: Block maximum size.

• -bss: Sequence block step size, define the minimum variance between block size. Default: 50.

• -js: Sequence block assembly overlap size, define the amount of overlap between blocks. Default: 40.

• -dir: Output directory for MOODA results.

• -gf: Allow the writing of FASTA and GenBank files. Default: False.

Authors

• Angelo Gaeta, a.gaeta@sms.ed.ac.uk
• Giovanni Stracquadanio, giovanni.stracquadanio@ed.ac.uk

Citation

Design and assembly of DNA molecules using multi-objective optimization. A Gaeta, V Zulkower, G Stracquadanio - Synthetic Biology, 2021

@article{10.1093/synbio/ysab026,
    author = {Gaeta, Angelo and Zulkower, Valentin and Stracquadanio, Giovanni},
    title = "{Design and assembly of DNA molecules using multi-objective optimization}",
    journal = {Synthetic Biology},
    volume = {6},
    number = {1},
    year = {2021},
    month = {10},
    issn = {2397-7000},
    doi = {10.1093/synbio/ysab026},
    url = {https://doi.org/10.1093/synbio/ysab026},
    note = {ysab026},
    eprint = {https://academic.oup.com/synbio/article-pdf/6/1/ysab026/40977182/ysab026.pdf},
}

Issues

Please post an issue to report a bug or request new features.