HINTS

You can select traits to include in the analysis in three ways:

a.

Set the trait to analyze at 0, so that no traits except those beginning with a [+] (plus sign) are analyzed. You would need to edit the .cro file first to prepend a + to all traits you wanted in the analysis.

b.

Set the trait to a value in the range [1-t] inclusive, where t is the number of traits in the .cro file. You will then get single trait results.

c.

Set the trait to a value greater than t. Then all traits will be put in the analysis, unless they begin with a minus sign [-]. As in a. above, you would need to edit the .cro file to minus out some traits.

You need to set the hypothesis test for SFx and RFx crosses. The default of 10 is ok for crosses in which there are only two marker genotypic classes (BCx, RIx). To test GxE, use 14. For SFx and RFx, values of 30, 31 or 32 are valid, and a 34 invokes the GxE test. Recall that we have the following hypotheses:

1. H0: a = d = 0

2. H1: a !=0 , d = 0

3. H2: a = 0 , d != 0

4. H3: a != 0, d != 0

For 30, we test H3:H0. For 31, we test H3:H0, H3:H1 and H1:H0. For 32, we test H3:H0, H3:H2 and H2:H0. 30 is probably fine for initial scans. Hypothesis 34 does a test for H3:H0 as well as the GxE.

For Model 6, be sure to run SRmapqtl first. Once done, JZmapqtl will use all markers that are significant for any of the traits in the analysis. We need to work out a better way to select the cofactors. Presently we use any markers that are significant for any trait. Also, be sure to use FB regression (Model 2 in SRmapqtl), or else you will end up using all markers as cofactors.