Contents - Index


Mean Performance vs. Stability of entries

From Biplot Tools, click Mean vs. Stability Coordination. The following will appear on the biplot:

Interpretations on Mean Performance:

Entry7 > Entry5 > Entry2 > Entry1 > Entry4 > Entry6 > Entry3.

Entry7 > Entry5 > Entry2 > Grand Mean >= Entry1 > Entry4 > Entry6 > Entry3.

Caution: The vector length of the average tester (the distance from biplot origin and the average tester marker), relative to the biplot size, is a measure of the relative importance of the entry main effect (G) vs. the entry by tester interaction (GE). The longer it is, the more important is G and the more meaningful the selection based on mean performance. At extremes, a zero average tester vector means G = 0 and therefore the selection based on mean performance is meaningless.

Interpretations on Stability:

 

Caution: When the biplot explains only a small portion of the total variation, some entries may not be as stable as they appear, because they may have greater values in PC3 or PC4.

In Depth: Here "stability" is defined by GE alone. Such a stability index can only be used as a modification of the genotype main effect in genotype evaluation. Stability is meaningless if it is detached from the genotype main effects. Some researchers devoted their research solely on GE and tried to identify genes for stability. This is very misleading because there is no such thing as stability genes: high stability means performing consistenly well when associated with high mean performance and consistely poorly when associated with low mean performance. How could the two opposite things be controlled by the same genes  

 

Also see Mean performance AND stability of entries