StructureTutorial

Structure Tutorial

[there are many little details that are difficult to explain in here -- I will try to point them out during the tutorial] A problem will be that the installation of the software will take time and will need troubleshooting. I suggest that we will use a show-and-tell mode for this tutorial (I show and tell and you ask questions and try to follow, but do not bother to much if there are problems).

Tutorial material: download STRUCTURE_TUTORIAL.ZIP

If you want to use this tutorial with your own data you will need to convert your data into the STRUCTURE format, I have used PGDSpider to do this conversion from my MIGRATE file to STRUCTURE. The file supplied in the structure_tutorial was sampled at four locations/populations with (known migration structure -- which I will reveal once report and review are written, I intend to use the same dataset for other programs if appropriate.

  1. Install the graphical version of Structure from http://pritchardlab.stanford.edu/structure_software/release_versions/v2.3.4/html/structure.html
  2. Open the program Structure [on macs see note at the end of file]
  3. Open a new project
  4. Follow the project wizard and give data path, output directory and files; it is important that you understand the structure of your data file, for example if you want plot the results in distruct you will need to specify the population association. For the example data I specified that we have a column for the individual_label and a column for the population_label.
  5. Create a parameter set (there are a couple of issues with this because it will ask several questions and some options will be not allowed, but you only will find out later, for example do not estimate POPINFO if there is no POPINFO information.
  6. Use "Start Job" to run the parameter set you just created, try to run a few replicates, because if you do not run at least 3 replicates you will not be able to use Evanno's method to estimate the number of populations.
  7. Look at the outputs (there is a folder with the parameters set and in it is a Results folder)
  8. Zip up the Results folder (zip -r filename.zip Results)
  9. Goto Structure Harvester at http://taylor0.biology.ucla.edu/structureHarvester/#
  10. run the Harvester, it will generate a plot that may help to figure the number of populations you should use for your analysis
  11. Download the archive generated by Harvester

Optional 12. Run these files optionally through CLUMPP (https://rosenberglab.stanford.edu/clumpp.html) CLUMPP reevaluates the population assignment and generates an average output that can be used to plot using the program distruct

  1. Get the program Distruct1.1 (https://rosenberglab.stanford.edu/distruct.html) If you use the harvester results directly you will need to make sure that only one replicate is in the popfile and the indfile (CLUMPP generates averages of the original Harvester files). Distruct uses a options file called drawparam, you will need to edit that file. In the distribution there is an example.

======================================
Problem with Mac operation systems: - You need Java installed (from here: [1]) - Structure is not signed by authors and the default Mac configuration will not allow to run. You will need to go to the preferences in the security options and allow "all programs to run", then start Structure.app, once it starts you save can return the previous option in the security setting.