Prerequisites for an efficient marker-assisted selection program

Listed below are the most important requirements for implementing a MAS program. See Xu (2003) for more details and discussion.

A. High throughput DNA extraction. Most breeding programs would need to screen hundreds to thousands of plants for desired marker patterns. In many cases, the results will be needed quickly to allow the breeder to make selections in a timely manner. Both of these considerations demand a DNA extraction system that can handle a large number of samples in a streamlined operation. Many labs conducting MAS extract DNA from small tissue samples in 96- or even 384-well plates (Fig. 2).

Figure 2. 96-well plates and multi-channel pipetters are tools to streamline DNA extraction and marker analysis.

B. Genetic markers. The species of interest should have available a technology that allows detection of markers along all of the plant’s chromosomes. Although DNA markers have received the most attention, other types of markers (protein, morphological, cytological) can also be used in MAS programs. DNA polymorphisms as markers have the advantages that large numbers are available throughout the genome, their presence or absence is unaffected by environment, and they usually do not directly affect the phenotype.

Commonly used DNA markers are referred to by their acronyms: RFLPSSRRAPDAFLPSCAR, and SNP, all defined in the glossary. Each type of marker has advantages and disadvantages for specific purposes. For efficient MAS, important attributes of markers include:

  • Ease of use
  • Small amount of DNA required
  • Low cost
  • Repeatability of results
  • High rate of polymorphism
  • Occurrence throughout the genome
  • Codominance

This last term (codominance) is the ability to detect both parental forms of a marker in heterozygotes. It is an advantage when heterozygous individuals are screened, such as in backcross breeding programs or in an F2 population. Figures 3 and 4 show examples of a codominant marker (SSR) and a dominant marker (AFLP).

Figure 3. Segregation pattern of the codominant SSR marker Xgwm11 in a wheat F2 mapping population, shown as a print-out from the computer program FRAGMENT MANAGER (Peng, 2003). Lanes 3 and 4 show the patterns of the two parents of the population, and lanes 5-20 correspond to 16 F2 individuals. As expected, about half of the F2’s are heterozygous (e.g., lane 6). The horizontal scale indicates fragment size in base pairs.

Figure 4. AFLP markers are usually dominant, scored as present or absent, as seen in this segregating wheat population. Some of the scorable bands are marked with arrows, and heterozygotes cannot be distinguished. Each vertical lane corresponds to a different DNA sample.

SSRs, also known as microsatellites, combine the desirable features listed above and are the current marker of choice for many crop species. SNPs require more detailed knowledge of the specific, single nucleotide DNA changes responsible for genetic variation among individuals. Only a small number of SNPs are currently available for MAS in plants, but within a few years many more are expected to be developed and may become an important marker type for MAS. Practical examples of SNPs are the markers for Rht dwarfing alleles in wheat published by Ellis et al. (2002).

C. Genetic maps. Linkage maps provide a framework for detecting marker-trait associations and for choosing markers to employ in MAS. Once a marker is found to be associated with a trait in a given population, a dense molecular marker map in a standard reference population will help identify markers that are closer to, or that flank, the target gene. Figure 5 shows a marker linkage map in barley.


Figure 5. Linkage map of the Oregon Wolfe Barley Population, showing the seven chromosomes of barley with centromeres, morphological markers, DNA markers, and linkage distances (  

D. Knowledge of associations between molecular markers and traits of interest. The most crucial ingredient for MAS is knowledge of markers that are associated with traits important to a breeding program. This information might come from gene mapping or QTL studies, bulked segregant analysis (Michelmore et al., 1991), classical mutant analysis, or some other means.

E. Data management system. Large numbers of samples are handled in a MAS program, with each sample potentially evaluated for multiple markers. This situation requires an efficient system for labeling, storing, retrieving, and analyzing large data sets, and producing reports useful to the breeder.