randmag 1.2.0

Construct artifical MAGs from complete genomes according to distribution

• Eric Hugoson ([e-mail] / [e-mail] / @EricHugo)

Introduction

With an ever increasing rate of sequencing comes challenges in analysing the data. In particular, single-amplified genomes and metagenome assembled genomes (SAGs & MAGs) may prove challenging due to their propensity to be incomplete or contaminated with contigs from other genomes. Many tools exist to tackle the challenges at various stages of the analysis, but none are perfect. In order to devise new tools methods for dealing with SAGs and MAGs the ability to simulate such data sets is paramount.

randMAG is a slim software aimed at producing artificial MAGs or SAGs from existent complete genomes. randMAG produces randomised MAGs based on whole genomes provided, a distribution of contig lengths given, as well as desired completeness- and contamination levels.

Packaged with randMAG is a file containing the lengths of 192243 contigs of 2284 bacterial MAGs [1] from the Tara Oceans metagenomic survey.

1. [TullyEtAl_2018]

Dependencies

Python (>=3.6)

Python libraries

If built from the PyPI package these will be installed automatically, otherwise can easily be installed using pip (Install pip).

Required

• Biopython (>= 1.70)

• Numpy (>= 1.13.1)

• Matplotlib (>= 2.0.2)

• SciPy (>= 1.1.0)

• seaborn (>= 0.9.0)

• pandas (>= 0.23.4)

Positional arguments

genome_tab            Genomes in .fna in list format
distribution          Set of lengths to base the distribution on

genome_tab should be a simple file with a list of genomes (nucleotide fasta files) that will be used to built the artificial MAGs, e.g.:

GCF_000007265.1_ASM726v1_genomic.fna
GCF_000007325.1_ASM732v1_genomic.fna
...

distribution should also be a simple file containing a list of integers representing a distribution of contig sizes by which the artificial contigs will be split, e.g.:

23967
15981
149609
...

Optionally, distribution can be filled with Tara_bact to use the set of contig lengths from Tara Oceans described in the introduction.

Optional arguments

-h, --help

show this help message and exit

-c COMPLETENESS, --completeness COMPLETENESS

Range of completeness levels to be produced. Default=1.0

-r CONTAMINATION, --contamination CONTAMINATION

Range of contamination to be included in produced MAGs. Default=1.0

-n NUM, --num NUM

Number of randomised MAGs to produce. Produces one randomised per provided genome by default.

Examples

Initially, create a file with a list of the genomic fasta files which are to be the basis for the artificial MAGs:

$ head -n5 reference_fna.list
/nobackup/genomic_sources/genbank_reference_bacteria_fna_20181128/ncbi-genomes-2018-11-28/GCF_000005845.2_ASM584v2_genomic.fna
/nobackup/genomic_sources/genbank_reference_bacteria_fna_20181128/ncbi-genomes-2018-11-28/GCF_000006745.1_ASM674v1_genomic.fna
/nobackup/genomic_sources/genbank_reference_bacteria_fna_20181128/ncbi-genomes-2018-11-28/GCF_000006765.1_ASM676v1_genomic.fna
/nobackup/genomic_sources/genbank_reference_bacteria_fna_20181128/ncbi-genomes-2018-11-28/GCF_000006785.2_ASM678v2_genomic.fna
/nobackup/genomic_sources/genbank_reference_bacteria_fna_20181128/ncbi-genomes-2018-11-28/GCF_000006845.1_ASM684v1_genomic.fna

Then a second file which is a list of lengths representing the distribution of contig sizes:

$ head -n5 Tara_bact
23967
15981
149609
37636
121320

Example 1 - Split into contigs

This example splits up all .fna files in reference_fna.list into contigs according to the distribution of contig lengths in Tara_bact:

$ randMAG reference_fna.list Tara_bact

The result is all a single .fna file created in the current working directory for each input .fna. Also produced is simulated_MAGs.tab, a tabular file containing input file, the unique 8-char suffix assigned, completeness-, and redundancy fractions:

$ head -n5 simulated_MAGs.tab
GCF_000005845.2_ASM584v2_genomic    mpmwoutk        1.0     1.0
GCF_000006745.1_ASM674v1_genomic    mmqehxfx        1.0     1.0
GCF_000006765.1_ASM676v1_genomic    engiyxoz        1.0     1.0
GCF_000006785.2_ASM678v2_genomic    rtvayhqu        1.0     1.0
GCF_000006845.1_ASM684v1_genomic    bzpsqnfq        1.0     1.0

The content of the fasta files in this example remained unchanged apart from being split into contigs.

Example 2 - Alter completeness/contamination

To change the completeness and contamination of the fasta files the -c and -r arguments need to be used. The -n argument can be used to get precisely the desired number of unqiue MAGs:

$ randMAG reference_fna.list Tara_bact -c 0.7 -r 1.2 -n 10000

This will produce files that are at most 70% complete and at least 20% contaminated:

$ head -n5 simulated_MAGs.tab
GCF_000005845.2_ASM584v2_genomic    kfcckaxy        0.6956260400391929      1.2092498368299183
GCF_000006745.1_ASM674v1_genomic    xqnerzfy        0.6927500102156292      1.202911467089386
GCF_000006765.1_ASM676v1_genomic    kiubfyau        0.6988059837775469      1.200510423558475
GCF_000006785.2_ASM678v2_genomic    xxltcsmv        0.6849106013550827      1.2144702270571384
GCF_000006845.1_ASM684v1_genomic    rcxezdoq        0.6927952822804169      1.2028640216997193

As well as 10 000 unique MAGs as requested with -n:

$ wc -l simulated_MAGs.tab
10000 simulated_MAGs.tab

References

[TullyEtAl_2018]

Tully,B.J. et al. (2018) The reconstruction of 2,631 draft metagenome-assembled genomes from the global oceans. Sci. Data, 5, 170203