Understanding crop genetic diversity under modern plant breeding

Abstract

Maximizing crop yield while at the same time minimizing crop failure for sustainable agriculture requires a better understanding of the impacts of plant breeding on crop genetic diversity. This review identifies knowledge gaps and shows the need for more research into genetic diversity changes under plant breeding. Modern plant breeding has made a profound impact on food production and will continue to play a vital role in world food security. For sustainable agriculture, a compromise should be sought between maximizing crop yield under changing climate and minimizing crop failure under unfavorable conditions. Such a compromise requires better understanding of the impacts of plant breeding on crop genetic diversity. Efforts have been made over the last three decades to assess crop genetic diversity using molecular marker technologies. However, these assessments have revealed some temporal diversity patterns that are largely inconsistent with our perception that modern plant breeding reduces crop genetic diversity. An attempt was made in this review to explain such discrepancies by examining empirical assessments of crop genetic diversity and theoretical investigations of genetic diversity changes over time under artificial selection. It was found that many crop genetic diversity assessments were not designed to assess diversity impacts from specific plant breeding programs, while others were experimentally inadequate and contained technical biases from the sampling of cultivars and genomes. Little attention has been paid to theoretical investigations on crop genetic diversity changes from plant breeding. A computer simulation of five simplified breeding schemes showed the substantial effects of plant breeding on the retention of heterozygosity over generations. It is clear that more efforts are needed to investigate crop genetic diversity in space and time under plant breeding to achieve sustainable crop production.

Fig. 1

Illustration of the spatial and temporal changes (solid line and some highlight in red) in crop genetic diversity generated by modern plant breeding with variable goals and methods (a) and how they are obscured (broken line) under diversity assessments of variable nature (b) (color figure online)

Fig. 2

Predicted heterozygosity (H _t) over generations for an allele of various characteristics (a neutral alleles; b recessive allele; c additive allele; and d dominant allele) under genetic drift and/or selection in a finite population of size N. The prediction for neutral allele (a) was obtained from the Eq. (1) and the predictions for b–d from the Eq. (2). The predictions in b–d are specified at a selective locus with two allele frequencies (q = 0.1, 0.4), two population sizes (N = 20, 50), two selection coefficients (s = 0.1, 0.2), and three levels of dominance [h = 0 (recessive), 0.5 (additive), and 1 (dominant)]

Fig. 3

Breeding schemes (a) and parental heterozygosity (H _o) changes (b) over 20 generations in simulated breeding programs to improve a quantitative trait of interest. The simulation considered five breeding schemes (Self = selfing; Half-sib = half-sib; SH = selfing + half-sib; HS = half-sib + selfing; HSH = half-sib + selfing + half-sib) and generated 50 diploid progeny in each generation with 5000 loci. The first 20 loci control the trait with four genetic models considered [neutral (s = 0, h = 0), recessive (s = 0.2, h = 0), additive (s = 0.2, h = 0.5), and dominant (s = 0.2, h = 1)]. The progeny with the largest genetic values were selected as parents for crossing and the parental heterozygotes were counted over 5000 loci. Environmental variation and its interactions with genotypes were not considered