Page 131 - Breeding and regulatory opportunities, Renaud

Broccoli crop improvement

To study correlations between conventional and organic conditions at the ine

grained level of location by trialing period combinations, we used Spearman’s

rank correlations, because we were especially interested in rank changes.

A further comparison of conventional and organic conditions was performed

by evaluating stability of performance versus mean performance for the set of

varieties. Genotypic (in)stabilities under organic and conventional conditions

were calculated as the variance for individual genotypes across all trials in the

system.

Correlated Response theory (Falconer and Mackay, 1996) was used to assess the

feasibility of selection for organic conditions (the target environment) under

conventional conditions (the selection environment). The ratio of correlated

response (for organic conditions using conventional conditions), CR, to direct

response (for organic conditions in organic conditions), DR, can be used to

decide whether it is possible to use selection under conventional conditions for

improvement in organic systems; it was calculated as the product of the genetic

correlation between organic and conventional systems () and the ratio of the
r
G
roots of conventional and organic repeatabilities ( and respectively):

CR/DR=rH⁄ H.
G C O

Ratios smaller than 1 indicate that it is better to select under organic conditions

when the aim is to improve the performance in organic conditions.

4.3 Results

4.3.1 Partitioning of variance components

We itted variance components models for all traits, where we report these

variance components as coeicients of variation (CV’s). We do not report the

signiicance of the variance components as almost all components were found

to be signiicant by likelihood ratio tests, even for components that were

close to zero. The information on the variation is best considered through the

magnitude of the variance components and not through signiicance tests.

113