Basic Concepts of Selection in Plant Breeding

K.K.Vinod

Generally breeders in two ways can change variability in their working population. Firstly by allowing individuals to produce next generation within limits of breeders preference or by natural choice in which which some of these individuals may be allowed to produce many offsprings and some of which may be allowed to produce which few or no offsprings. This non-random reproduction in a breeding population is called selection. The two agencies involved in the selection, one is by the natural choice is called natural selection and the other that is done by the breeder is the artificial selection. Secondly, breeders can decide how the selected individuals will be mated to each other either by inbreeding and cross-breeding.

Following basic principles operates on selection:

1. Selection works only on heritable variation

Variation that is governed by genetic makeup of the individual alone can be recovered in the progenies, not the variation that is environmentally dependent.

2. Selection works only in existing variability ie., it cannot create new variation. 

Hence, selection can only shuffle the variation and also can combine variations of different individuals of the population, which is however, pre-exist within the population itself.

3. Selection works by providing preference to certain individuals to reproduce, therefore certain set of genes get preferential inheritance at the expense of the unselected genes. Asa consequence gene and genotypic frequency  get altered. 

4. Continuous selection leads to loss of variability, because unselected genes get eliminated in the course of generation advancement.

5. When a new variability is created in a population either by mutation or by movement of genes from outside (from another unrelated population), new selection pressure is generated on the introduced genes.

6. Purpose and consequence of selection varies from type of selection, individual genes, mating methods etc.

Plant breeding employs artificial selection to generate plant populations that are cultivated for the benefit of the mankind. Artificial selection is fundamentally aimed at the following facts:

1. To change the phenotype of the selected individual in the most desired way to the man

2. To bring down the variability in the selected population

3. To increase uniformity in the selected population

4. To increase number of populations with different level of variability between them

5. To increase the range of uniform populations

Various artificial selection methods used in plant breeding are, 

In self pollinated crops,

1. Mass selection

2. Pureline selection

3. Pedigree selection

In cross pollinated crops,

1. Mass selection

2. Progeny selection (ear to row methods, modified ear to row methods etc)

3. Family selection (half sib, full sib family selection and their modifications)

4. Recurrent selection

Clonally propagated crops,  

1. Clonal selection.

Genetic makeup and variability

K K Vinod

The genetic structure of a population determines its capacity for response to selection, both natural and artificial, and as such is of primary consid­eration in the formulation of strategies for the collection and conservation of biodiversity. The structure of the population is controlled by several fac­tors such as its life form, breeding system and effective population size. These factors, which often reflect past selection pressures, all influence the nature and maintenance of genetic variation both within and between populations and in some cases may themselves be subject to genetic deter­mination. In the conservation of biodiversity it is the underlying genetic control of variability that is of major importance in determining appropri­ate strategies.

The phenotype of an organism is controlled by a multitude of genes which act both individually and in concert upon the various stages of develop­ment and are influenced to varying degrees by the environment. Their action leads, in the majority of cases, to a quantitative expression of growth forms which are continuously distributed in nature. The genetic control of these quantitative traits is by sets of genes (polygenes or quantitative trait loci) each of small effect; although these may be difficult to identify individually, they are inherited in a Mendelian manner and show all the properties of major genes, i.e. linkage, dominance, epistasis and the effects of ploidy. The study of such traits has in the past required the application of biometrical procedures utilizing means and variances, but develop­ments in molecular techniques for genome analysis and genetic mapping offer the prospect of more precise identification of single quantitative trait loci.

The behaviour of genes determining quantitative traits in a population is the same as that of major genes. If random mating is the mode of repro­duction, at a single locus level the individual genotypes are to be found in the Hardy-Weinberg proportions, p²:2pq:q², where p and q represent the diallelic frequencies. When extended over many loci it can be seen that the extreme homozygous classes and thus phenotypes are rare in the popula­tion. These genotypes, which represent free variability and are directly fix­able by selection, have the capacity by hybridization and segregation to create all intermediate genotypic classes. In doing so, the majority of individuals produced will be of differing homozygous/het­erozygous genotypic combinations and, as such, will give rise to interme­diate phenotypes. These differing genotypic classes again have the potential, by hybridization and segregation, to release variation. Here, however, this hidden variability is in two states, the homozygotic and het­erozygotic (Mather, 1973). As the number of genes controlling a trait increases, the proportion of variability exposed to the rigours of selection in the homozygous state will decrease.

The mode of gene action will also influence the state and proportion of exposed variability. The effects of dominance will be two-fold depend­ing on the direction of dominance at the individual loci concerned. Firstly, if all dominant alleles are acting in the same direction the effect will be to reduce the number of phenotypic classes observed as the heterozygous classes will be indistinguishable from the dominant homozygotes. As a consequence, selection will be more difficult as potential variability will still be present in the heterozygotes and can only be revealed by progeny testing. The distribution of individuals will be very much skewed in the direction of the dominant expression. Secondly, if dominance is ambidirec­tional, the effect will be to increase the proportion of intermediate pheno­types in the distribution and with it the release of variability and the potential for response to selection. Nonallelic interaction will to some extent reinforce the effects of dominance in leading to a reduction in class frequencies and the mean expression of a trait in a population.

The evolution of the genetic architecture of a trait is governed by the components of the genetic system (Darlington, 1958), namely the creation of new variation by mutation, recombination and the breeding system. These, when coupled with selection and/ or genetic drift, are the major determinants controlling the manner in which variability is organized within a population.

Genetic variation due to mutation

Mutational change at the DNA level of the genome is the basis of new variation and can take several forms such as base pair deletion, duplica­tion or rearrangement. Its effect may be detectable at the molecular level, as for example in the changes that lead to differing forms of an enzyme (allozyme), which in most cases would be neutral in its action, or it may have a gross effect on the phenotype such as in flower form or colour. Most mutational changes are considered to be deleterious in that they generally disrupt the hitherto integrated structure of the gene. However, some may be advantageous, with their subsequent survival and spread in the population being dependent on such factors as selective advantage, population size and genetic drift. If the mutation is recessive, as in most cases, its frequency in the homozygous state will initially be very rare in outbreeding species, hence the likelihood of exposure to the rigours of selection is very low. However, mutations that affect the breeding system can be at an immediate advantage. The occurrence of a mutant incompat­ibility allele in a single-locus gametophytic system, such as occurs in Trifolium repens, would be advantageous in that it is directly exposed in the haploid phase in the pollen grain and can be effective in promoting fertilization. In that it provides a further option for cross-pollination to occur its survival in the population/ species is ensured. This may well account for the very high number of incompatibility alleles that can be found in species with gametophytic systems.

Adaptive change may arise in a population through alterations at the chromosomal level. This may take the form of structural or numerical change such as gross deletions, inversions, interchanges, aneuploidy and polyploidy. The mechanisms and origin of these types of change are well documented (Darlington, 1956). It is their influence on the maintenance and release of variability and the opportunity they provide for new adap­tive forms to arise which are of importance from a conservation aspect. Polyploidy, for example, which may arise by the direct doubling of a chro­mosome set or be coupled with wide hybridization, is well known as a mechanism for maintaining heterozygous combinations of genes (Stebbins, 1950).

Genetic variation due to recombination

The role of recombination in controlling the release and distribution of variation within a population is of fundamental concern in the develop­ment of strategies for the conservation of genetic resources. It has long been established that the mechanisms controlling chromosome pairing, and the frequency and position of crossing over in the genome are under genetic control. The evolution of the Ph (pairing) gene on chromosome 5B of wheat has led to the regu­lar diploid pairing that takes place, and with it the stability and fertility of a diploid as opposed to the instability and sterility of an allopolyploid. Selection, irrespective of whether it be natural or artificial, can lead to marked differences in the rate of recombination between populations. In the outbreeding species Lolium perenne and Festuca pratensis bred cultivars have a higher chiasma frequency than their wild counterparts. This has arisen as a correlated response to selection for varia­bility by the breeder. Similarly the presence of B chromosomes can influ­ence chiasma frequency. The fine-scale collinearity of cereal and grass genomes should enable strategies to be developed for the positional cloning of the gene(s) controlling chromosome pairing in wheat and the forage grasses, opening up the prospect of genetically manipulating the processes of recombination at will and, with it, the range of variation that may be extracted from a population.

Genetic variation due to selection, drift and gene flow

The differentiation of populations depends on the three processes of selec­tion, drift and gene flow. The forces of selection, reflecting the environmen­tal pressures acting upon the population, are instrumental in defining the underlying genetic structure. The differing modes which it may take, such as stabilizing, disruptive or directional, each have their own effect upon the subsequent gene action and organization of the variation (Mather, 1973). Random drift, particularly in small populations, can lead to arbit­rary fixation of genes. The immigration of new genes from distinctive neighbouring subpopulations can increase the extent of both selective response and drift by introducing new alleles; or it can reduce it by repeatedly introducing genes adapted to a different microenvironment and by so doing retard micro-adaptation to local patches. In addition, life form and the persistence of seed banks may all influence the capacity for selection to lead to local adaptation.

Variability in breeding systems

The flow of variability within a species is dependent on its mode of repro­duction. Sexual species, which may be either inbreeders or out-breeders, have the capacity for recombination and as a consequence variability may be exposed to selection. Asexual forms, which reproduce either by apo­mixis or vegetative means, maintain a uniform genotype, which may be well adapted to present selective forces, but lack the ability to respond to changing conditions. The breeding system is often under genetic control and may be associated with specific life forms. Inbreeding, which is pre­dominantly found in annual life forms, often at the limits of a species dis­tribution (Stebbins, 1950), is generally achieved by mechanisms that ensure self-pollination. Pollen may be shed within closed florets (cleistogamy), as in wheat and barley, or flowers may open and be receptive when anther dehiscence occurs (chasmogamy), as in tomato. Although these mecha­nisms are under precise genetic control, breakdown may occur, allowing outcrossing to take place. In barley, for example, Allard and Hansche (1965) showed that up to 1% outcrossing may be found under some envi­ronmental conditions. Novel recombinants will appear offering scope for further selection and evolutionary change.

Outbreeding is generally found in the more perennial species and is often promoted by one or more genetically controlled mechanisms. These may range from timing differences in anther dehiscence and the receptiv­ity of the stigma through to precisely controlled incompatibility systems. The consequence of such processes is the mainten­ance of a high level of heterozygosity within the individual and variability both between individuals within the population and between populations. This aspect of population structure will be considered in more detail in a later section.

The apomictic mode of reproduction, which involves the production of seed by asexual means, is found in many genera, predominantly of the Gramineae and Rosaceae. It is generally associated with polyploidy and can be obligate or facultative. In those cases where sexual relatives are to be found, which allow genetic analysis, it has been shown to be under simple genetic control. For example, in Panicum maximum it appears to be under the control of a single dominant gene whilst in Citrus several genes are involved. Apomixis has the attribute of maintaining well-adapted combinations of genes together but has the disadvantage that there is no flow of variability and as such the species may well be at an evolutionary dead end (Stebbins, 1950).

Truly vegetative reproduction is rare but like apomixis can lead to the widespread distribution of a species. Spartina anglica is now to be found all around the shores of Great Britain having spread from its origins in Southampton Water by the continual breakup of its rhizomes. It is a repro­ductive mode that is often exploited by humans to maintain and distribute a crop species, as in the potato.

Each of these reproductive modes can be under genetic control and thus subjected to the forces of natural selection in the same manner as the genes responsible for other traits of adaptive significance. An insight into their effect on the state of variability and structure of populations can be obtained from the numerous studies of molecular markers in plant popu­lations.

References:

Allard, R.W. and Hansche, P. E. (1965) Population and biometrical genetics in plant breeding. In: Geerts, S.J. (ed.) Genetics Today, vol. 3, Proceedings of XIth International Congress of Genetics, The Hague, Netherlands, 1963. Pergamon Press, Oxford, pp. 665-668.

Darlington, C.D. (1956) Chromosome Botany. Allen & Unwin, London, 186 pp.

Darlington, C.D. (1958) The Evolution of Genetic Systems. Oliver & Boyd, London, 265 pp.

Mather, K. (1973) Genetical Structure of Populations. Chapman & Hall, London, 197pp.

Stebbins, G.L (1950) Variation and Evolution in Plants. Columbia University Press, New York.