Conventional breeding, marker-assisted selection, genomic selection and inbreeding in clonally propagated crops: a case study for cassava

Abstract

Consolidates relevant molecular and phenotypic information on cassava to demonstrate relevance of heterosis, and alternatives to exploit it by integrating different tools. Ideas are useful to other asexually reproduced crops. Asexually propagated crops offer the advantage that all genetic effects can be exploited in farmers' production fields. However, non-additive effects complicate selection because, while influencing the performance of the materials under evaluation, they cannot be transmitted efficiently to the following cycle of selection. Cassava can be used as a model crop for asexually propagated crops because of its diploid nature and the absence of (known) incompatibility effects. New technologies such as genomic selection (GS), use of inbred progenitors based on doubled haploids and induction of flowering can be employed for accelerating genetic gains in cassava. Available information suggests that heterosis, non-additive genetic effects and within-family variation are relatively large for complex traits such as fresh root yield, moderate for dry matter or starch content in the roots, and low for defensive traits (pest and disease resistance) and plant architecture. The present article considers the potential impact of different technologies for maximizing gains for key traits in cassava, and highlights the advantages of integrating them. Exploiting heterosis would be optimized through the implementation of reciprocal recurrent selection. The advantages of using inbred progenitors would allow shifting the current cassava phenotypic recurrent selection method into line improvement, which in turn would allow designing outstanding hybrids rather than finding them by trial and error.

Fig. 1

Illustration of a truncated phenotypic recurrent selection scheme such as the one currently used in cassava breeding. C1, C2, C3 and C4 are the successive cycles of selection. Shifts in allelic frequencies gradually occur in the different versions of the population represented by the successive cycles. This genetic progress is achieved mostly exploiting additive genetic effects. The selection of a successful clone, however, is affected by all genetic effects as well as experimental errors and the ever confounding effect of genotype-by-environment interaction

Fig. 2

Illustration of a typical scheme of reciprocal recurrent selection of two heterotic populations. Hybrids are made through crosses of selected gentoypes from each population with a tester from the reciprocal population. Progenitors of the best hybrids are combined (within each population) to start a new cycle of selection

Fig. 3

A breeding scheme for cassava based on the use of inbred progenitors from two heterotic populations. Solid black arrows indicate the between heterotic group crosses for production and evaluation of experimental hybrids. White arrows indicate within-population variation. Line A is gradually improved for its heterotic response when crossed with line B. On the other hand, line B is improved for resistance or quality traits

Fig. 4

Photograph of S₅ cassava genotypes produced and grown at CTCRI in Kerala State, India. Notice the uniformity of the two plants shown. These plants come from botanical seed. The uniformity is indicative of the high degree of inbreeding in their parental genotype. Also notice that they have already flowered at least once (as branching has already taken place)