Configuring a parameter set.

Once you've successfully loaded a data file, you are ready to start running structure. You will create one or more ``parameter sets''; these represent a whole list of choices that you make about how to analyze the data. We have entered a series of default settings, and these are good place to start. You will probably want to run structure multiple times for each parameter set, at different values of $K$, and the front end is set up to facilitate this. Go to the pull-down menu under Parameter Set. You can create a new parameter set, modify an existing one, or delete one. Click on ``New''. You now see a dialogue box with four tabs (Figure 5). Click on each of these:

Run Length. See section 4.1 for discussion of this. Note that the front end provides time series plots of some key parameters to help you assess whether the run length seems to be sufficient.

Ancestry Model. See section AncestryModels. The admixture model is a good place to start for most data sets. The linkage model will be disabled unless you entered linkage information about the markers. Section 7 provides some extra details about some of the detailed options, including GENSBACK and MIGRPRIOR under the ``Use Population Information'' option. Note that the linkage model is relatively computationally intensive.

Allele Frequency Model. See section 4.3. We recommend applying both the correlated frequencies model and the independent frequencies model. The correlated frequencies model has better power to detect subtle population structure, but the posterior probabilities for $K$ may be biased upwards somewhat. The correlated frequencies model is parameterized in terms of $F_{ST}$, with a separate parameter for each population (details in section 4.3). Inferring $\lambda$ is probably not necessary during initial investigation.

Advanced. Turning off the function that computes the posterior probabilities (for estimating $K$) speeds the program up significantly. You can also get the program to output posterior credible regions for the ancestry of each individual (see ANCESTDIST, section 7.3). ``Initialize at POPINFO'' is described in more detail under STARTATPOPINFO, section 7.5.

Figure 5: Specifying a new parameter set: setting the run length.