Alternative Experimental Designs for Plant Breeding

Le résumé fourni par la source

With limited resources, plant breeders must make trade-offs in resource allocation. While the randomized complete block design (RCBD) has been popular for agricultural research, it has been recognized that alternative experimental designs may better meet the goals of plant breeding. These alternative designs primarily address the issues with the RCBD that arise when blocks do not sufficiently control field variation. This occurs when the entry numbers are large (leading to large block sizes and heterogeneity within blocks), when field variation exists in more than one direction, or when there is insufficient seed or other resources available to replicate all entries. While these alternative designs have been widely adopted in large-scale breeding programs of field or row crops (where they provide a clear and immediate benefit), they have been less widely used in small- to mid-sized programs, where it is more difficult to evaluate the trade-offs in accuracy and resource allocation between different field layouts of experimental designs. However, these programs also face the limitations of the RCBD and may benefit from the use of alternative designs. This chapter will review these designs as used in plant breeding, including incomplete block designs such as lattice designs and α-designs, row–column designs such as Latin square and Latinized α-designs, and partially replicated designs such as augmented designs and p-rep designs. It will also cover variations of these designs, methods of design and analysis. We focus on considerations when using these designs in smaller trials, particularly when conducting organic, low-input, or participatory plant breeding.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.