• Running phylotaR with restez
  • " />

    In this tutorial we will showcase how a restez database can be used to speed up a phylotaR run. phylotaR runs an automated pipeline for identifying ortholgous gene clusters as the first step in a phylogenetic analysis. A user provides a taxonomic identity and the pipeline downloads all relevant sequences and identifies clusters using a local-alignment search tool. For more information on phylotaR see its published article.

    By using restez in conjunction with phylotaR, we will not only being saving time, but also improving the chances of a successful phylotaR run – often NCBI Entrez limits the number of requests or even rejects requests from IP addresses that are making too many. Note, however, that the gains in using restez with phylotaR only make sense if you make use of the restez database multiple times or if you wish to radically increase the maximum number of sequences to download per taxon (by default it is only 3,000). Also, note that using a restez database does not currently eliminate the need for an internet connection. phylotaR still needs to look up taxonomic information and must also identify relevant sequence IDs using Entrez (this may change in the future as restez develops).

    We will run a phylotaR run for the rodent subfamily, Dipodinae, the clade that contains jerboas. Because this clade is so small, it should not take too long to run it. We will, however, limit to showcasing only the download stage of the phylotaR pipeline, where restez is required.

    Install phylotaR

    # Currently only the development version of phylotaR can work with restez
    # It must be installed from GitHub using devtools
    devtools::install_github(repo = 'ropensci/phylotaR')

    Setup

    Since we will be running the phylotaR pipeline for Dipodinae, we can use the rodents database we created before. We do not need to set-up the phylotaR pipeline any differently with a restez database, except to ensure we have set the restez path to the rodent database. (Connections to the restez database will be made automatically, restez_connect or restez_disconnect do not need to be run.)

    library(phylotaR)
    
    # Restez (no need to call package)
    restez::restez_path_set(filepath = rodents_path)
    
    # Vars
    wd <- 'dipodinae'
    dir.create(wd)
    txid <- 35737  # Dipodinae
    mxsql <- 500
    ncbi_dr <- '[PATH/TO/BLAST]' # e.g. '/usr/local/ncbi/blast/bin'
    
    # setup
    setup(wd = wd, txid = txid, ncbi_dr = ncbi_dr, mxsql = 500)

    Run

    # run just the first two stages for this demonstration
    taxise_run(wd)
    #> --------------------------------------------
    #> Starting stage TAXISE: [2018-11-27 16:14:21]
    #> --------------------------------------------
    #> Searching taxonomic IDs ...
    #> Downloading taxonomic records ...
    #> . [1-25]
    #> Generating taxonomic dictionary ...
    #> ---------------------------------------------
    #> Completed stage TAXISE: [2018-11-27 16:14:25]
    #> ---------------------------------------------
    download_run(wd)
    #> ----------------------------------------------
    #> Starting stage DOWNLOAD: [2018-11-27 16:14:25]
    #> ----------------------------------------------
    #> Identifying suitable clades ...
    #> Identified [1] suitable clades.
    #> Downloading hierarchically ...
    #> Working on parent [id 35737]: [1/1] ...
    #> . + whole subtree ...
    #> . . Getting [100 sqs] ...
    #> . . . [1-100]
    #> Successfully downloaded [99 sqs] in total.
    #> -----------------------------------------------
    #> Completed stage DOWNLOAD: [2018-11-27 16:14:38]
    #> -----------------------------------------------

    Sequences that cannot be found locally will be downloaded via Entrez. Sequences may not be found locally for three sets of reasons, 1. phylotaR may have identified non-GenBank sequences (e.g. RefSeq), 2. the local database may have size limits (ours has a min of 100 and a max 1000), and 3. the local database is out of date, new GenBank releases are only made once a month.

    Note: If the phylotaR messages do not explicit state that they are using a restez database, then, you may not have set the restez path.

    Compared to running without a restez database, phylotaR download can run 2x faster.

    Next up

    Tips and tricks