A pipeline to perform phylogenesis from protein sequences

The pipeline is designed on Ubuntu 14.04 LTS, and was only tested on this Linux distribution.

phylo_pipe is a pipeline to perform phylogenesis from protein sequences.

It combines softwares and scripts in the order below:

• Orthofinder

Standalone software, find orthologus from protein, more info here: https://github.com/davidemms/OrthoFinder

• read_ortho_out.py

Create p/a matrix, multifastas of single copy orthologs (SCO)

• muscle_batchrun.py

Aligns single copy orthogroups with Muscle

• phi_batchrun.py [optional]

Perform test for recombination with Phi; recombining SCO are discarded

• gblock_batchrun.py

Perform Gblocks on each alignment to remove not informative sites

• orthos2cat.py

Concatenate gblocked multialigned SCO into one file

The final output is obtained from orthos2cat.py and is a concatenated Multiple Single Alignment (MSA) combining all not recombinant single copy orthologs gblocked MSAs. This file can be used as input for a phylogenetic software such as FastTree or RaxML. A temp folder includes every output for each software 1-5.

The file name of each protein mutifasta will be the one appearing in the phylogenetic tree; change it accordingly. Anyway, the file name must be no longer than 60 characters and it mustn't end with "a" (ex: "Midichloria.faa" it's not fine, it should be converted in Midichlori.faa or something else without "a" as the last character)

When performing the pipeline on a reduced number of taxa (~less than 3/4) and/or on taxa with short proteins, Phi may not be able to work; the pipeline will continue skipping this step and printing a WARNING.

The following softwares must be installed for the pipeline to work: Blast+, MCL And the following python modules: python-dev, pandas, numpy, biopython.

• How to install from the terminal

Blast+, MCL:

sudo apt-get install [software name]

python-dev, pandas, numpy, biopython:

sudo apt-get install python-[name of module]

python-dev is a requirement for all other modules, install it first

• Full manual from the terminal:

python phylo_pipe.py -h

The pipeline must be run inside its folder. Once you're inside the folder, from the terminal:

python phylo_pipe.py -i [input_folder] -t [number_of_threads] {-r} -p [Phi_test]

see the description below:

• -i

Mandatory, folder containing one protein multifasta for each organism, you have to specifies the folder's path

• -t

Mandatory, number of threads depending of your PC, more is faster

• -r

Optional, remove temp folder: if -r is specificed it will remove the temp folder. If it is not present it will keep it.

• -p

Optional, if -p True is specified it will run Phi test. If it is not present it will skip this step.

Examples:

• Run the pipeline with 11 threads, does NOT remove the temp folder, perform Phi test

python phylo_pipe.py -i fasta_folder -t 11 -p True

• Run the pipeline with 11 threads, remove the temp folder, perform Phi test

python phylo_pipe.py -i fasta_folder -t 11 -r -p True

• Run the pipeline with 8 threads, remove the temp folder, do not perform Phi test

python phylo_pipe.py -i fasta_folder -t 8 -r

output files will be in the multifastas folder (input_folder as described in 4.) and will be the following:

• orthos2cat_out

The main output of the pipeline (described in 1.)

• [OPTIONAL] temp

Temporary files folder (if multiple analyses are run; each one will have a temp folder with a progressive number -e.g. temp, temp_1, temp_2 etc. - ) containing the output of all softwares and scripts componing the pipeline.

The process may require some time, depending of your hardware specifics and number of taxa/number of proteins per taxon.

You can test the pipeline with the "test" folder, which contains a trial dataset of multifastas.

• Run the pipeline with "test" folder

python phylo_pipe.py -i test -t 11 -p True