Contents - Index


Biplot based on grand-mean centered data

 

From Models, select Centering, then Grand-mean Centered

 

For genotype-by-environment data, biplot based on this model would contain E+G+GE:

 

 

Note that PC1 explains mostly the variation of tester main effect (E), and PC2 mostly that of the entry main effect (G). Thus, Testers 5, 6, and 8 had obviously higher mean values than testers Testers 9, 3, and 5. Entries 5 and 7 had obviously higher mean values than entries 3, 4, and 6. WhyBecause all testers are on the side of entries 5 and 7! In other words, Entries 5 and 7, in contrast with entries 3, 4 and 6, 'interacted' positively with all the testers.

 

Although 88% of the total E+G+GE is explained by this biplot, not much GE can be visualized from this biplot. This biplot is therefore not very effective in displaying the GE patterns. Therefore, this biplot is not effective in genotype evaluation, environment evaluation, or revealing genotype by environment interaction patterns. It may be more effective than model 5, however, as the grand mean is at least removed, which represents no variation. Compare with other types of biplots.