If we were going to do a live demo …

• install R and RStudio
  • see datacarpentry instructions
  • or next two slides
• install packages qtl, rrBLUP
• download or view https://github.com/byandell/thailand

R: powerful programming language for data exploration, visualization, and analysis

• main R site: https://cran.r-project.org/
• Thailand mirror: http://mirrors.psu.ac.th/pub/cran/ (use this)
• Want version 3.2.2 or higher
• click on machine version to "Download and Install R"
• for Windows, click on "base"
• click on current version to install
• follow instructions

RStudio: nice graphical interface for R projects

• https://www.rstudio.com/products/rstudio/download3/
• click "Download" under "RStudio Desktop Personal License"
• click on "Installer" for your machine
• follow instructions for easy install

• open up Rstudio, which opens R
• On console, type

install.packages("qtl", "rrBLUP")

This will install the two main packages we will use.

• http://rqtl.org (lots of useful stuff here)
  • see R/qtl Install steps
• A brief tour of R/qtl: [pdf | code]
• A shorter tour of R/qtl: [pdf | code]
• R/qtl sample data formats
• A Guide to QTL Mapping with R/qtl: book by Karl W. Broman & Śaunak Sen

• EMMA (Efficient Mixed Model Association)
• EMMAX (Efficient Mixed-Model Association eXpedited)
• Convert matrix operations to vector operations
  • with one clever eigendecomposition
  • 1000x faster than TASSEL/SAS implementation
  • O(n) per iteration
  • accurate and reliable
• Eleazar Eskin Lab (UCLA) standalone packages

• rrBLUP
• EMMA adapted to genotypic prediction
  • A.mat to compute \(K\) from markers
  • kin.blup to predict phenotypes
• http://potatobreeding.cals.wisc.edu/software/
• Amy Jacobson slides
• Jeff Endelman Notes

qtl2scan

• EMMA adapted to R/qtl2 package
• leave out one chromosome method
  • calc_kinship with type="loco" to compute \(K\)
  • scan1 with kinship option to fit mixed model
• Karl Broman (UW-Madison) R package suite

There is a very useful set of R packages that were recently developed, which are packaged together as the "Tidyverse" (http://tidyverse.org/). We will not use these directly, but they are worth knowing about. To install them type the following on the Rstudio console:

install.packages(tidyverse)

This will install a number of small packages on your system.

For the advanced R/qtl2 section, there are further packages. We will discuss that later.

Want to go further organizing your work?

• Karl Broman's initial steps to reproducible research
• Karl Broman's data organization tips

• many online resources
  • http://www.stat.wisc.edu/~yandell/statgen/reference/: Yandell Statistical Genomics References
  • http://datacarpentry.org: spreadsheet to R in 2 days
  • https://www.lynda.com/: online skills tutorials
  • Professional Skills for Data Science
• open source tools are quite good
  • R & Rstudio for data science
  • Github for version control
• Google Drive, DropBox, Box, …

Material for this course at https://github.com/byandell/PlantSysGen