TaxIDermy

a set of scripts compiled by Bryan Wee (1st Aug 2012)

This pipeline produces an artemis readable genbank file that colours CDSs according according to their taxonomic distance from the reference species.

It facilitates annotation of genomic islands or mobile elements by differentiating regions that have different phylogenetic distances from the organism of interest.

Updated: 28th Nov 2012

Updates

v0.1 - 28 Nov 2012 Released on GitHub and tidied up

v0.2 - 4 Apr 2014 Fixed bug. Crashes when CDS hit is from Viruses

Requirements

• BioPerl
• Biopython
• Blast+
• local copy of nr
• 1x .faa protein sequence multi fasta file from the genome that you want to analyse with locus tags
• 1x gbk file of the genomic island/region you want to analyse
• 1x gi_taxid_prot.dmp Note: this file is huge >1GB
• 1x nodes.dmp
• 1x names.dmp
• 1x scientific_names.dmp Note: this file is derived from names.dmp. See Note 2

Note:

1. The latest version of the .dmp files can be downloaded at ftp://ftp.ncbi.nih.gov/pub/taxonomy. nodes.dmp and names.dmp are found within taxdump.tar.gz
2. scientific_names.dmp is made using: grep "scientific" names.dmp > scientific_names.dmp
3. make sure the latest version of nodes.dmp, names.dmp and gi_taxid_prot.dmp are used so to ensure compatibility with blast hits from the latest version of NR.

Instructions

1. The first step is to create a list of gi numbers that correspond to sequences in NCBI that belong to the taxons that you do not want to include in the results. This is us This can be done by using the script retrieve_child_gis.pl "<Node to filter>" and using the generated file <Taxon>.gi_list in the blastp step in step 2:

   $ retrieve_child_gis.pl "Pseudomonas"
   # or
   $ retrieve_child_gis.pl "Pseudomonas aeruginosa PA14"
   

This script will output the gi numbers of associated with all taxies below the node "Pseudomonas" into a new file: <Taxon>.gi_list. The first line of the STD_OUT of this script will be <TaxonID> <Name of taxon> <Taxon level>. The rest of the STD_OUT of this script will be the list of Taxons that are below the Pseudomonas node.

2. Do a blastp using your the protein multi-fasta from the region you would like to colour against nr. Run blastp with the flag -negative_gilist <Pseudomonas>.gi_list:

   $ blastp  -query Pseudomonas_PA14.faa -db nr -outfmt 6 -out  Pseudomonas.v.nr.blastp.tab -num_threads 8 -negative_gilist Pseudomonas.gi_list
   

Note: nr must be in your path. if not, point the -db argument to nr on your computer. The -negative_gilist flag is incompatible with -remote which is why a local copy of nr is needed.

3. To create the final output, the following script needs to be run:

   $ python taxIDermy.py <blastp output> <Genome genbank> <TaxID of reference> <output.file.with.colour>
   Example: $ python taxIDermy.py Pseudomonas.v.nr.blastp.tab Pseudomonas_PA14.gbk 446 > bryanscript.colour.out.tab
   

Note: this script calls traverse_taxonomy.pl which must be in the working directory

For use on a parallel cluster

1. Same as step 1 above.
2. Prior to running blastp, split up the multi fasta (.faa) file (e.g. using genometools: gt splitfasta -numfiles 10 example.faa)
3. Use sample PBS script (included) to run both blastp followed by taxIDermy.py on a cluster