Potential drawbacks of MAS

MAS is not universally advantageous. Some limitations of the technique are as follows:

Figure 7. Linkage maps of two chromosomes showing positions of two resistance genes and nearby markers. If a marker is located close to the gene of interest (as shown for resistance gene 1), then one could reliably select the desired allele of the resistance gene based on the allele pattern for m2. As the distance between the marker and the gene of interest increases (as for resistance gene 2), there is a greater chance of recombination between gene and marker. If recombination does occur, selecting for an allele at the marker would not result in the expected allele at the resistance gene. To reduce the occurrence of that problem, it would be safer to select for the desired allele pattern at both flanking markers, m6 and m7. If both flanking markers have the allele pattern of the resistant parent, this indicates that recombination probably has not occurred between them, and the target locus will most likely also have the resistant parent allele.

  • MAS may be more expensive than conventional techniques, especially for startup expenses and labor costs. Dreher et al. (2003) and Morris et al. (2003) discuss economic factors of MAS compared to phenotypic selection.
  • Recombination between the marker and the gene of interest may occur, leading to false positives. For example, if the marker and the gene of interest are separated by 5 cM and selection is based on the marker pattern, there is an approximately 5% chance of selecting the wrong plant. This is based on the general guideline that across short distances, 1 cM of genetic distance is approximately equal to 1% recombination. The breeder will need to decide the error rate that is acceptable in the MAS program, keeping in mind that errors are also usually involved in phenotypic evaluation.
  • To avoid this last problem it may be necessary to use flanking markers on either side of the locus of interest to increase the probability that the desired gene is selected (Fig. 7). Peng et al. (2000) discuss this issue in detail.
  • Sometimes markers that were used to detect a locus must be converted to 'breeder-friendly' markers that are more reliable and easier to use. Examples are:
  1. RFLP markers converted to STS markers (Ribaut et al., 1997). Evaluation of RFLPmarkers requires several steps and a large quantity of highly purified DNA. STS (sequence tagged site) markers can be detected via PCR using primers developed from RFLP probe sequences. Thus the same locus can be detected with the two types of marker, but the STS marker is far more efficient.
  2. RAPD markers converted to SCAR markers (Hernandez et al., 1999). Results of RAPDreactions may vary from lab to lab, and therefore, may be considered less reliable for MAS. Variable RAPD results are due largely to short (10-base) PCR primers, resulting in low binding specificity. SCAR markers are developed by sequencing RAPD bands and designing more specific 18-25 base PCR primers to amplify the same DNA segment more reliably.
  • Imprecise estimates of QTL locations and effects may result in slower progress than expected. Many QTLs have large confidence intervals of 20 cM or more or their relative importance in explaining trait inheritance has been over-estimated (Beavis, 1998; Kearsey and Farquhar, 1998).
  • Markers developed for MAS in one population may not be transferrable to other populations, either due to lack of marker polymorphism or the absence of a marker-trait association.