PoSeiDon

Please note that the code of PoSeiDon was just transfered to Nextflow so there might be still some bugs. Please feel free to report issues!

Here we present PoSeiDon, a pipeline to detect significantly positive selected sites and possible recombination events in an alignment of multiple protein-coding sequences. Sites that undergo positive selection provide insights in the evolutionary history of your sequences, for example showing important mutation hot spots, accumulated as results of virus-host arms races during evolution.

PoSeiDon relies on a variety of different third-party tools (see below). But don't worry, we encapsulated each tool in its own Docker container and connected them in the Workflow Management System Nextflow.

Go directly to a small example of PoSeiDon output for the SARS-CoV-2 spike protein in comparison to a recent study of Zhou et al. 2020.

Installation

You only need Nextflow (version 20.+) and Docker installed to run the pipeline. All dependencies will be pulled automatically.

Either run PoSeiDon by cloning this repository:

git clone https://github.com/hoelzer/poseidon.git
cd poseidon
nextflow run poseidon.nf --help

or let Nextflow do the pull

nextflow pull hoelzer/poseidon

We recommend using a specific release of PoSeiDon via

#pull
nextflow pull hoelzer/poseidon -r v1.0.1

#run
nextflow run hoelzer/poseidon -r v1.0.1 --help

Update

Depending on you installation procedure, update the pipeline via git pull or nextflow pull hoelzer/poseidon.

Run

Important: PoSeiDon needs nucleotide sequences with a correct open reading frame as input. In addition, the results heavily depend on your selection of sequences, thus, you might consider running the pipeline multiple times with different samplings of your input sequences. Also the pipeline can't work with too many sequences because in its core PoSeiDon uses CODEML from the PAML suite that is not inteded for >100 sequences. Please find a detailed description of the input parameters and settings below.

Profiles

Nextflow can be easily executed on different environments like your local machine, a high-performance cluster or the cloud. Different -profile are used to tell Nextflow which system should be used. For local execution -profile local,docker should be used (and is also the default). You can also run PoSeiDon on a HPC using Singularity via -profile lsf,singularity or -profile slurm,singularity. In such cases, please also consider to adjust --cachedir to point where to store Singularity images on your cluster. The parameter --workdir might be also helpful to adjust where to store temporary working directories (e.g. use /scratch instead of /tmp depending on your HPC configuration.)

Examples

Now let's assume you used Nextflow to pull the PoSeiDon code and you execute the pipeline on a local machine using the default profile -profile local,docker.

# show help 
nextflow run hoelzer/poseidon --help

# run small example on a local machine with 
# (first time this will need some more time because the Docker containers are downloaded)
nextflow run hoelzer/poseidon -r v1.0.1 --fasta ~/.nextflow/assets/hoelzer/poseidon/test_data/bats_mx1_small.fasta \
--cores 4

# resume a broken run
nextflow run hoelzer/poseidon -r v1.0.1 --fasta ~/.nextflow/assets/hoelzer/poseidon/test_data/bats_mx1_small.fasta \
--cores 4 -resume

# instead of using all available cores only use a maximum amount on the local machine
nextflow run hoelzer/poseidon -r v1.0.1 --fasta ~/.nextflow/assets/hoelzer/poseidon/test_data/bats_mx1_small.fasta \
--max_cores 8
--cores 4

To reproduce the positive selection results reported in Fuchs et al. (2017), Journal of Virology run:

nextflow run hoelzer/poseidon -r v1.0.0 --fasta ~/.nextflow/assets/hoelzer/poseidon/test_data/bats_mx1.fasta \
--cores 4 --kh --outgroup "Pteropus_alecto,Eidolon_helvum,Rousettus_aegyptiacus,Hypsignatus_monstrosus" \
--reference "Myotis_daubentonii"

Workflow of the PoSeiDon pipeline and example output

The PoSeiDon pipeline comprises in-frame alignment of homologous protein-coding sequences, detection of putative recombination events and evolutionary breakpoints, phylogenetic reconstructions and detection of positively selected sites in the full alignment and all possible fragments. Finally, all results are combined and visualized in a user-friendly and clear HTML web page. The resulting alignment fragments are indicated with colored bars in the HTML output.

The PoSeiDon pipeline is based on the following tools and scripts:

• TranslatorX (v1.1), Abascal et al. (2010); 20435676
• Muscle (v3.8.31), Edgar (2004); 15034147
• RAxML (v8.0.25), Stamatakis (2014); 24451623
• Newick Utilities (v1.6), Junier and Zdobnov (2010); 20472542
• MODELTEST, Posada and Crandall (1998); 9918953
• HyPhy (v2.2), Pond et al. (2005); 15509596
• GARD, Pond et al. (2006); 17110367
• PaML/CodeML (v4.8), Yang (2007); 17483113
• Ruby (v2.3.1)
• Inkscape (v1.0)
• pdfTeX (v3.14)

Parameters

Most of the PoSeiDon parameters are optional and are explained below in detail.

Input

• --fasta

Mandatory. Your input FASTA file must follow the format:

>Myotis_lucifugus Mx1 Gene
ATGGCGATCGAGATACGATACGTA...
>Myotis_davidii Mx1 Gene
ATGGCGGTCGAGATAAGATACGTT...

All sequences must have a correct open reading frame, are only allowed to contain nucleotide characters [A|C|G|T] and no internal stop codon.

Sequence IDs must be unique until the first occurrence of a space.

Reference

• --reference

Optional. Default: use first sequence ID as reference. You can define one species ID from your multiple FASTA file as a reference species. Positively selected sites and corresponding amino acids will be drawn in respect to this species. The ID must match the FASTA header until the occurence of the first space. For example, if you want Myotis lucifugus as your reference species and your FASTA file contains:

>Myotis_lucifugus Mx1 Gene
ATGGCGATCGAGATACGATACGTA...

use

--reference "Myotis_lucifugus"

as parameter to set the reference species. Per default the first ID occurring in the multiple FASTA file will be used.

Outgroup(s)

• --outgroup

Optional. Default: trees are unrooted. You can define one or multiple (comma separated) species IDs as outgroup. All phylogenetic trees will be rooted according to this species. For example, if your multiple FASTA file contains

ATGGCGATCGAGATACGATACGTA...
>Myotis_davidii Mx1 Gene
ATGGCGGTCGAGATAAGATACGTT...
>Pteropus_vampyrus Mx1 Gene
ATGGCCGTAGAGATTAGATACTTT...
>Eidolon_helvum Mx1 Gene
ATGCCCGTAGAGAATAGATACTTT...

you can define:

--outgroup "Pteropus_vampyrus,Eidolon_helvum"

to root all trees in relation to this two species.

Use also insignificant breakpoints

• --kh

Optional. Default: false. With this parameter you can decide if insignificant breakpoints should be taken into account. All breakpoints are tested for significant topological incongruence using a Kashino Hasegawa (KH) test Kishino, H. and Hasegawa, M. (1989). KH-insignificant breakpoints most frequently arise from variation in branch lengths between segments. Nevertheless, taking KH-insignificant breakpoints into account could be interesting, because we already observed putative positively selected sites in fragments without any significant topological incongruence. KH-insignificant fragments are marked in the final output, as they might not occur from real recombination events.

Per default only significant breakpoints are used for further calculations.

Please also keep in mind that using also insignificant breakpoints can extend the run time of PoSeiDon from minutes to hours, depending on the number of detected breakpoints.

Other parameters

Please see the --help for other parameters (GARD, RAxML, ...) and let us know if you need more customization!

Citation

If PoSeiDon helps you please cite:

Martin Hölzer and Manja Marz, "PoSeiDon: a Nextflow pipeline for the detection of evolutionary recombination events and positive selection", OUP Bioinformatics (2020)