How to run Structure, Structure Harvester, CLUMPP, and Distruct

How to run Structure :

1. Format your data so it similarly looks like the one from the example files (I even used that example file to create my file by deleting the allele one by one because Structure just didn't want to load my file although I have formatted it correctly. Note: I use Mac OS X v 10.7.5)

2. Open Structure; create a directory where you want your Structure files will be put (for example: My Doc/Structure)

3. Open your data file

4. Create a parameter set by click on menu: Parameter Set, New...

5. Fill in the length of burnin period (100.000 is usually more than enough, for more info please consult Structure manual by Pritchard et al.), and number of MCMC reps that you want (people use variable reps for this, from 100.000 to 2.000.000 or even more).

6.I usually follow the instruction to use admixture model for first run, leave the others as default.

7. Allele frequencies correlated.

8. Compute probability of the data (for estimating K)

9. Click ok and name your parameter

10. On menu file: Project, start a job, click on the parameter that you want to run, specify how many k you want to test for (k=1 to k=the max number of cluster you initially thought your data might possibly have)

11. Number of iterations: 5-10 or more

12. Click ok and let structure runs.

13. After complete, you can continue with Structure Harvester.


How to run Structure Harvester:

1. Go to the Results folder from your Structure results (for example: My Doc/Structure/ParameterName/Results)

2. Zip the Results folder. (Results.zip)

3. Upload that to Structure Harvester http://taylor0.biology.ucla.edu/structureHarvester/

4. Harvest!

5. Download the harvester output files.


How to run CLUMPP:

1. Based on Evanno et al. (2005) delta K formulation, you can identify your k.

2. Based on that k (for example k=2), take the specific k file for the indfile (then you will take K2.indfile; if your k is 5, take K5.indfile), move it to a new folder called (for example) FolderA

3. Edit the paramfile from the example files downloaded with CLUMPP package,
Datatype 0, revise everything else accordingly (for example change the indfile name to K2.indfile and so on). Move/copy this file to FolderA

4. Copy/move also CLUMPP into your FolderA so in your folder, there are: CLUMPP, k2.indfile and the edited paramfile

5. Open Terminal, change the directory to where your FolderA is located

6. then type ./clumpp paramfile

7. CLUMPP will produce several files in the folder. Take the output file (for example arabid.outfile) and change it into arabid.indivq

8. Repeat number 2-7 for K2.popfile by changing the paramfile to Datatype 1 and put those 3 files to FolderB; take the output file (for example: arabid.outfile) change the name to arabid.popq


How to run Distruct:

1. Take your indvq and popq files (for example: arabid.indivq and arabid.popq) and move it to a Folder C

2. Edit the drawparams file from the Distruct package accordingly. Create your arabid.names and arabid.perm files

3. Put those 5 files (at least): arabid.indivq, arabid.popq, arabid.names, arabid.perm files and drawparams in Folder C with distruct in it.

4. Run distruct (for me, distruct does not work in my Mac, so I have to use PC to run it. i just click on the windows executable file and distruct produce the output.ps file.

Yay!