Marker validation
Markers that are found to be linked to a gene only appear if the marker is polymorphic (in fact di-morphic) between both parents. However, even if they are closely linked to the gene of interest they have to be tested ('validated') in other populations, such as breeding material, in order to be useful in marker-assisted breeding. Many markers found to be associated with a gene of interest in a mapping population are not useful in breeding material. This can be due to several reasons, e.g.:
- The marker (and maybe the linked gene!) is not polymorphic in other breeding material:
- the contrasting undesirable allele and its associated marker allele that were detected in the mapping work may not occur in breeding lines. Mapping studies may be done on a mapping population derived from an exotic or low-domesticated parental accession that has many undesirable traits: crosses with an economically elite cultivar results in populations in which many markers and QTL alleles are found for many agronomically relevant characteristics. However, most of the markers and negative genes do not occur in offspring from crosses among elite cultivars.
- the marker might have been polymorphic in the mapping population, but the marker allele linked with the desirable allele of the gene of interest may occur in high allele frequency in the breeding population. So, in many crossing populations both parents carry the same (common) marker allele, making it uncertain whether or not the parents carry different alleles for the gene of interest. This is especially a problem in the case of AFLP (or other) markers where a null allele (absence of amplification product) does not necessarily imply a same DNA sequence.
- In different sources, traits may be under the influence of different genes, e.g. a high protein content may be the result of a QTL located on chromosome 1 (gene A) in one source and on chromosome 2 (gene B) in another source. Marker-assisted selection with markers linked to gene A in a population that segregates for gene B is not effective! If gene A is discovered in a mapping population, one may not necessarily presume that the variation in other breeding material is due to the same gene.
- A marker allele that in a mapping study has been indicated to be associated with a gene of interest, may in another breeding line be linked to less desirable allele of the trait gene, due to recombination between the marker in the trait gene in the past. This chance is higher if the other breeding line is more distantly related to the line used in the mapping study.
- Marker alleles for selection in complex crosses should be diagnostic. If material has been generated from crosses in which the parent with the desirable gene was only a great-parent (or more remote relative) the offspring can only be selected for the gene of interest if the associated marker allele is diagnostic (unique). This means that the marker allele to be selected for should not also occur in one of the other ancestors. If one or more of the other ancestors has the same marker allele, it may not have the same gene of interest, so selection for the marker allele makes it uncertain which trait gene is being selected. This is in particular a limiting factor in the case:
- the marker allele to be selected for is a null allele. In for example AFLP a null allele may be due to different causes (see above).
- the marker is an SSR marker. SSR markers are based on differences in numbers of repeats, which might not pair correctly during meiosis. As a result, offspring may have a different number of repeats, and thus a different SSR allele than expected. This also may happen during separate events in different individuals, resulting in alleles that look the same, but do not indicate a common origin of that allele.
Reliability of a marker can be predicted from its distance to a QTL. To see to what extent a marker is capable in predicting the desired phenotype, the marker should be tested in other populations. This can be done by crossing one of the parents used in the initial mapping population with a different cultivar, to develop another mapping population. In the resulting population, the ability of a marker to predict the phenotype can be evaluated and compared to the original mapping result. However, this is quite an investment, and implies that marker-assisted selection is not just a quick and easy strategy!
The best possible markers are markers that are part of the actual functional gene sequence or so tightly linked that they show no recombination with the target gene. Also markers that are consistently found to be absolutely linked to the target gene and unique in the germplasm. Such markers are highly predictive for the target gene and have a good change to be 'diagnostic' as well.
Summary
→ Markers need to be validated to make sure they are indeed tightly linked to genes of interest in breeding material
→ Markers need to be diagnostic to be followed if the donor of the gene is crossed to elite breeding lines of various origin
Show/hide comprehension question...
Show/hide comprehension question...