In thierrygosselin/stackr: Run stacks pipeline for RADseq analysis inside R

stackr

description <- readLines("DESCRIPTION")
rvers <- stringr::str_match(grep("R \\(", description, value = TRUE), "[0-9]{1,4}\\.[0-9]{1,4}\\.[0-9]{1,4}")[1,1]
version <- gsub(" ", "", gsub("Version:", "", grep("Version:", description, value = TRUE)))

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

stackr: an R package to run stacks pipeline

This is the development page of stackr.

stackr package provides wrapper functions to run STACKS (process_radtags, ustacks, cstacks, sstacks, tsv2bam, gstacks and populations) inside R.

Who is it for ?

stackr is currently developed for my projects at CSIRO and to help colleagues and collaborators to get the most out of stacks.
I make stackr available for others, but I don't offer help, use it at your own risk.
STACKS related issues should be highlighted in stacks google group.
It's not for R and/or stacks beginners. Train with STACKS in the Terminal first.

Overview of the differences from running stacks

I work with thousands of samples, my STACKS experience is more reproducible and organized inside RStudio
stackr requires you to think ahead by using a project info file that outlines the sample names and it's metadata (think sampling sites, barcodes, sequencing lanes, etc).
The outputs are organized by folders, it's more tidy and it's automatic. e.g. logs generated by STACKS are summarized in human-readable tables/tibbles, making problem detection easier.
Most functions that runs STACKS can be restarted if the computer/cluster/server crash. Just restart the same function. stackr takes care of the rest.

Made it this far, here's more details:

process_radtags -> run_process_radtags (demultiplexing) : works better with multiple lanes and technical replicates inside or across chip/lanes are managed and accounted for.
ustacks -> run_ustacks (de novo assembly): manages STACKS unique integer (previously called SQL IDs) and can also do mismatch testing! Integrated inside run_ustacks is a de novo mismatch threshold series function that generates tables and figures automatically to help you pick the appropriate thresholds.
cstacks -> run_cstacks (catalog)
With more than 1000 samples, cstacks is problematic: time consuming and vulnerable to computer crashes and other interruptions.
stackr generates catalogs incrementally, this way the pipeline is more rigorous if your computer/cluster/server crash, because the previous catalog will already be saved and available to start over.
Different catalogs with increasing sample number = better testing

Installation

To try out the dev version of stackr, copy/paste the code below:

if (!require("devtools")) install.packages("devtools")
devtools::install_github("thierrygosselin/stackr")
library(stackr)

Stacks modules and RADseq typical workflow

stackr package provides wrapper functions to run STACKS process_radtags, ustacks, cstacks, sstacks, tsv2bam, gstacks and populations inside R.

Below, a flow chart showing the corresponding stacks modules and stackr corresponding functions.

Vignette

Citation:

To get the citation, inside R:

citation("stackr")

Life cycle

stackr is quite mature, i've used it for almost 8 years with dozens of projects, but changes are inevitable.

For reproducibility, if you use it, keep a copy safe.
I follow STACKS development closely, when a new version comes out, stackr will likely change as well.
Because of it's intrinsic nature, stackr will not be compatible with previous STACKS version.
Philosophy, major changes and deprecated functions/arguments are documented in life cycle section of functions.
The latest changes are documented in changelog, versions, new features and bug history
issues and contributions