A comparison of formal and participatory breeding approaches using selection theory

Abstract

Progress from plant breeding has been slow in some marginal environments. Conventional or formal plant breeding (FPB) programs conducted by international agricultural research centres or national programs in developing countries have been criticized for ignoring indigenous germplasm, failing to breed for conditions facing poor farmers, and emphasizing selection for broad versus local adaptation. A suite of techniques, referred to as participatory plant breeding (PPB) and including farmer-participatory or farmer-led selection, on-farm evaluation, and use of local landraces, has been advocated in response to this critique. PPB programs are diverse in scope and approach, but often rely heavily on farmer visual evaluation or phenotypic mass selection to select for simply-inherited traits, with limited replicated yield testing in multiple-environment trials (MET), one of the main tools of FPB. Prediction equations derived from selection theory can be used to examine the conditions under which idealized versions of FPB and PPB may be expected to achieve genetic progress for traits such as yield. The effectiveness of any selection environment is determined by both the genetic correlation between genotype performance in it and the target environment (r G) and the heritability of genotypic differences in the selection environment (H s). r is a measure of the accuracy with which performance of genotypes in the selection environment predicts performance in the target environment; H s is a measure of the precision with which performance differences among genotypes can be measured in the selection environment. We compare FPB and PPB with respect to these determinants of selection response, using examples from self-pollinated species. Particular areas examined include: (i) selection for broad versus specific adaptation; (ii) on-station versus on-farm selection; and (iii) selection under high-yield versus low-yield conditions. In general, PPB systems attempt to maximize gains through the use of on-farm evaluation and the skills of farmer-selectors to maximize r G. FPB exploits METs to maximize H s. PPB is most likely to develop cultivars that out-perform the products of FPB when it is applied in low-yield cropping systems, because it is in such situations that r G between high-yield breeding nurseries and low-yield target environments is likely to be low or negative. To make continued gains, and to compete with internationally-supported FPB programs, PPB systems will need to counter the obscuring effects of uncontrollable within-field, site-to-site, and year-to-year heterogeneity. Simple and robust designs for on-farm METs are needed for this purpose.