Testing for Selection

In this example, we describe how to perform a codon-based test of positive selection for five alleles from the human HLA-A locus (Nei and Hughes 1991).

The “HLA-3Seq.meg" data file, which is used in this tutorial, can be found in the MEGA/Examples folder (The default location for Windows users is C:\Program Files\MEGA\Examples. The default location for Mac users is $HOME/MEGA/Examples, where $HOME is the user’s home directory).

 

Computing Synonymous and Non-synonymous Distances

Example 7.1:

Activate the "HLA-3Seq.meg" file. If necessary, refer to Example 1.2 in the “MEGA Basics” tutorial.

From the main MEGA window launch bar, select Selection | Codon-based Z-Test of Selection.

An Analysis Preferences window appears. For the Model/Method, select the Nei-Gojobori method (Proportion) model.

In the Test Hypothesis (HA: alternative) row, select Positive Selection (HA: dN > dS) from the pull-down menu.

From the Scope row, select the Overall Average option.

For the Gaps/Missing Data Treatment option, select Pairwise Deletion.

Click on "Compute" to accept the default values for the remaining options. A progress indicator appears briefly, and then the computation results are displayed in a results window in grid format.

The column labeled "Prob" contains the probability computed (must be <0.05 for hypothesis rejection at 5% level). The column labeled "Stat" contains the statistic used to compute the probability. The difference in synonymous and non-synonymous substitutions should be significant at the 5% level.

Close the Test of Positive Selection window.