"Molecular marker assisted selection as a potential tool for genetic improvement of crops, forest trees, livestock and fish in developing countries"
Having reached the landmark of 10 conferences in this FAO Biotechnology Forum, we are happy to dedicate an entire conference to biotechnology involving the use of DNA markers, in particular to their use in marker assisted selection (MAS) for genetic improvement of domestic plant and animal populations in developing countries.
The potential benefits of using markers linked to genes of interest in breeding programmes have been obvious for many decades. However, realisation of this potential has been limited by the lack of markers. With the advent of DNA-based genetic markers in the late 1970's, the situation changed and researchers could, for the first time, begin to identify large numbers of markers dispersed throughout the genetic material of any species of interest and use the markers to detect associations with traits of interest, thus allowing MAS to finally become a reality. This led to a whole new field of academic research, including the milestone paper by Paterson and co-workers in 1988 which showed, given the availability of large numbers of genetic markers for their species of interest (tomato), how the effects and location of marker-linked genes impacting a number of quantitative traits (fruit traits in their case) could be estimated, using an approach that could be applied to dissect the genetic make-up of any physiological, morphological and behavioural trait in plants and animals.
Most of the traits considered in animal and plant genetic improvement programmes are quantitative traits i.e. they are controlled by many genes, together with environmental factors, and the underlying genes have small effects on the observable phenotype. Milk yield and growth rate in animals or yield and seed size in plants are typical examples of quantitative traits. In classical genetic improvement programmes, selection is carried out based on observable phenotypes (of the candidates for selection and/or their relatives) but without knowing which genes are actually being selected. The development of molecular markers was therefore greeted with great enthusiasm as it was seen as a major breakthrough promising to overcome this key limitation. As Young wrote in a recent review, "Before the advent of DNA marker technology, the idea of rapidly uncovering the loci controlling complex, multigenic traits seemed like a dream. Suddenly, it was difficult to open a plant genetics journal without finding dozens of papers seeking to pinpoint many, if not most, agriculturally relevant genes".
However, despite the considerable resources that have been invested in this field and despite the enormous potential it still represents, MAS, with few exceptions, has not yet delivered its expected benefits in commercial breeding programmes for crops, animals, forest trees or farmed fish in the developed world. This is just one of the aspects that should be considered in this e-mail conference which aims to examine the appropriateness and potential of MAS as a tool for genetic improvement in developing countries.
This Background Document aims to provide information that participants will find useful for the debate. Firstly, a brief overview of the technical aspects of molecular markers and MAS is provided. Then, the current status of the application of MAS in crops, forest trees, livestock and fish is summarised. Section 4 then raises some important issues that might be relevant to applications of MAS in developing countries. Section 5 provides references to articles mentioned in the document. Finally, in Section 6, some of the topics that should be discussed throughout the conference are highlighted.
On 17-18 October 2003, the Fondazione per le Biotecnologie, the University of Turin and FAO organised an international workshop in Turin, Italy, entitled "Marker assisted selection: A fast track to increase genetic gain in plant and animal breeding?". The proceedings of the firstname.lastname@example.org, with 11 papers covering crops, livestock, fruit trees and farmed fish, provide an excellent overview of the current status of MAS and can be consulted by anyone looking for more detailed technical information on this subject.
In conferences hosted by the FAO Biotechnology Forum, clearly defined topics of relevance to agricultural biotechnology in developing countries are discussed for a limited amount of time. In defining the topic for this conference, it can be noted that although molecular markers may be used for a wide range of different tasks, such as to quantify the genetic diversity and relationships within and between agricultural populations (e.g. livestock breeds), to investigate biological processes (such as mating systems, pollen movement or seed dispersal in plants) or to identify specific genotypes (e.g. cloned forest trees), these applications will not be considered in the conference and we will instead focus on the use of molecular markers for genetic improvement of populations through marker assisted selection, including marker assisted introgression.
2. Background to MAS
2.1 Molecular markers
To begin at the beginning, we should say that all living things are made up of cells that are programmed by genetic material called DNA. This molecule is made up of a long chain of nitrogen-containing bases (there are 4 different bases - A, C, G and T). Only a small fraction of the DNA sequence typically makes up genes, i.e. that code for proteins, while the remaining and major share of the DNA represents non-coding sequences whose role is not yet clearly understood. The genetic material is organised into sets of chromosomes (e.g. 5 pairs in Arabidopsis thaliana; 30 pairs in cattle), and the entire set is called the genome. In a diploid individual (i.e. where chromosomes are organised in pairs), there are two alleles of every gene - one from each parent.
Molecular markers should not be considered as normal genes, as they usually do not have any biological effect, and instead can be thought of as constant landmarks in the genome. They are identifiable DNA sequences, found at specific locations of the genome, and transmitted by the standard laws of inheritance from one generation to the next. They rely on a DNA assay, in contrast to morphological markers, based on visible traits, and biochemical markers, based on proteins produced by genes.
Different kinds of molecular markers exist, such as RFLPs, RAPDs, AFLPs, microsatellites and SNPs. They may differ in a variety of ways - such as their technical requirements (e.g. whether they can be automated or require use of radioactivity); the amount of time, money and labour needed; the number of genetic markers that can be detected throughout the genome; and the amount of genetic variation found at each marker in a given population. The information provided by the markers for the breeder will vary depending on the type of marker system used. Each one has its advantages and disadvantages and, in the future, other systems are also likely to be developed. A brief overview of the major marker systems follows:
Restriction Fragment Length Polymorphisms (RFLPs) are markers detected by treating DNA with restriction enzymes (enzymes that cut DNA at a specific sequence). For example, the EcoR1 restriction enzyme cuts DNA whenever the base sequence GAATTC is found. Differences in the lengths of DNA fragments will then be seen if, for example, the DNA of one individual contains that sequence at a specific part of the genome (e.g. tip of chromosome 3) whereas another individual has the sequence GAATTT (which is not cut by EcoR1). RFLPs were the first molecular markers to be widely used. Their use is, however, time-consuming and expensive and simpler marker systems have subsequently been developed.
Random amplified polymorphic DNA (RAPD) markers were first described in 1990. They are detected using the polymerase chain reaction (PCR), a widespread molecular biology procedure allowing the production of multiple copies (amplification) of specific DNA sequences. The analysis for RAPD markers is quick and simple, although results are sensitive to laboratory conditions.
In the mid 1990's, another PCR-based method of generating molecular markers was described, giving rise to amplified fragment length polymorphism (AFLP) markers. With this technique, DNA treated with restriction enzymes is amplified with PCR. It allows selective amplification of restriction fragments giving rise to large numbers of useful markers which can be located on the genome relatively quickly and reliably. Unlike other methods described here, the technique is patented.
These are simple DNA sequences (e.g. AC), usually 2 or 3 bases long, repeated a variable number of times in tandem. They are easy to detect with PCR and a typical microsatellite marker has more variants than those from other marker systems. Initial identification of microsatellites is time-consuming.
In recent years, single nucleotide polymorphisms (SNPs), i.e. single base changes in DNA sequence, have become an increasingly important class of molecular marker. The potential number of SNP markers is very high, meaning that it should be possible to find them in all parts of the genome, and micro-array procedures have been developed for automatically scoring hundreds of SNP loci simultaneously at a low cost per sample.
Korzun (2003), considering the case of cereals, provided a nice comparison of these marker systems:
1 = Amount of DNA required (in micrograms) 2 = Quality of DNA required 3 = PCR-based 4 = Number of polymorphic loci analysed per analysis 5 = Ease of use 6 = Amenable to automation 7 = Reproducibility 8 = Development cost 9 = Cost per analysis
2.2 From markers to MAS
The molecular marker systems described above allow high-density DNA marker maps (i.e. with many markers of known location, interspersed at relatively short intervals throughout the genome) to be constructed for a range of economically important agricultural species, thus providing the framework needed for eventual applications of MAS.
The next step is that, using the marker map, putative genes affecting traits of interest can be detected by testing for statistical associations between markers variants and any trait of interest. These traits might be genetically simple - for example, many disease resistance traits in plants are controlled by one or a few genes (Young, 1999). Alternatively, they could be genetically complex quantitative traits, involving many genes (i.e. so-called quantitative trait loci (QTLs)) and environmental effects. (Most economically important agronomic traits tend to fall into the second category). For example, Babu and co-workers in a recent paper (2003), using 280 molecular markers (comprising 134 RFLPs, 131 AFLPs and 15 microsatellites) and recording populations of rice lines for various plant water stress indicators, phenology, plant biomass, yield and yield components under irrigated and water stress conditions, detected a number of putative QTLs for drought resistance traits.
Having identified markers physically located beside (or, even, within) genes of interest, it is now possible, in the next step, to carry out MAS, i.e. to select identifiable marker variants (alleles) in order to select for non-identifiable favourable variants of the genes of interest. For example, consider a hypothetical situation where a molecular marker M (with two alleles M1 and M2), that we can identify using a DNA assay, is known to be located on a chromosome close to a gene of interest Q (with a variant Q1 that increases yield and a variant Q2 that decreases yield), that is, as yet, unknown. Then, if a given individual in the population has the alleles M1 and Q1 on one chromosome and M2 and Q2 on the other chromosome we know that any of its progeny receiving the M1 allele will have a high probability (how high it is depends on how close M and Q are to each other on the chromosome) of also carrying the favourable Q1 allele, and thus would be preferred for selection purposes, while those that inherit the M2 allele will tend to have inherited the unfavourable Q2 allele, and so would not be preferred for selection. With conventional selection, relying on phenotypic values, it is not possible to use this kind of information.
The success of MAS is influenced by the relationship between the markers and the genes of interest. Dekkers (2003) distinguished three kinds of relationship:
1) The molecular marker is located within the gene of interest (i.e. within the gene Q, using the example above). In this situation, we can refer to gene assisted selection (GAS). This is the most favourable situation for MAS since, by following inheritance of the M alleles, we directly follow inheritance of the Q alleles. On the other hand, it is most difficult to find these kinds of markers.
2) The marker is in linkage disequilibrium (LD) with Q throughout the whole population. LD is the tendency of certain combinations of alleles (e.g. M1 and Q1) to be inherited together. Population-wide LD can be found when markers and genes of interest are physically very close to each other and/or when lines or breeds have been crossed in recent generations. Selection using these markers can be called LD-MAS.
3) The marker is not in linkage disequilibrium (i.e. it is in linkage equilibrium (LE)) with Q throughout the whole population. Selection using these markers can be called LE-MAS. This is the most difficult situation for applying MAS.
Because of the universal nature of DNA, molecular markers and genes, MAS can, in theory, be applied to any agriculturally important species and active research programmes have been devoted to building molecular marker maps and to detecting QTLs for potential use in MAS programmes in a whole range of crop, livestock, forest tree and fish species. In addition, MAS can be applied to support existing conventional breeding programmes. These programmes use strategies such as: recurrent selection (i.e. using within-breed or within-line selection, important in livestock); development of crossbreds or hybrids (by crossing several improved lines or breeds) and introgression (where a target gene is introduced from a low-productive line or breed (donor) into a productive line (recipient) that lacks the target gene (a strategy especially important in plants)). See Dekkers and Hospital (2002) for more details. MAS can be incorporated into any one of these strategies (e.g. for marker assisted introgression, by using markers to accelerate introduction of the target gene). Alternatively, novel breeding strategies can be developed to harness the new possibilities that MAS raises.
3. Current Status of Applications of MAS in Agriculture
Here, we provide a brief summary of the current status regarding application of MAS in the different agricultural sectors.
The promise of MAS has possibly been greeted with most enthusiasm and expectation in this particular agricultural sector, stimulating tremendous investments in the development of molecular marker maps and research to detect associations between phenotypes and markers. Molecular marker maps have been constructed for a wide range of crop species. Information on major plant projects (such as the sequencing of the entire rice genome) can be found at http://www.ncbi.nlm.nih.gov/genomes/PLANTS/PlantList.html
Dekkers and Hospital, however, in a recent review (2002) noted that "as theoretical and experimental results of QTL detection have accumulated, the initial enthusiasm for the potential genetic gains allowed by molecular genetics has been tempered by evidence for limits to the precision of the estimates of QTL effects" and that "overall, there are still few reports of successful MAS experiments or applications". They reported that marker assisted introgression of known genes was widely used in plants, particularly by private breeding companies, whereas marker assisted introgression of unknown genes had often proved to be less useful in practice than expected. As Young (1999) wrote: "even though marker-assisted selection now plays a prominent role in the field of plant breeding, examples of successful, practical outcomes are rare. It is clear that DNA markers hold great promise, but realizing that promise remains elusive".
There is also considerable divergence between different crop species with respect to their applications of MAS. For example, Koebner (2003) highlights the relatively fast uptake of MAS in maize compared with wheat and barley, arguing that it largely reflects the breeding structure, where maize breeding is dominated by a small number of large private companies that produce F1 hybrids, a system allowing protection from farm-saved seed and competitor use, while for the other major cereal species breeding is primarily by public sector organisations and most varieties are inbred pure breeding lines, a system allowing less protection over the released varieties. Progress in arable crops is nevertheless quite advanced compared to horticultural crop species, such as apples and pears, where development of molecular marker maps has been slow and only few QTLs have been detected (Tartarini, 2003), even if MAS can potentially be very useful for genetic improvement of such long-cycle plants.
As for crops, extensive efforts have been devoted to construction of molecular marker maps for the major commercial genera, such as eucalypts, pines and acacia. RFLPs, RAPDs, microsatellites and AFLPs have been extensively used. The website http://dendrome.ucdavis.edu/index.php provides updated information on the status regarding molecular marker maps in forestry.
The molecular maps have been used to locate markers associated with variation in forestry traits of commercial interest, such as growth, frost tolerance, wood properties, vegetative propagation, leaf oil composition and disease resistance. A major incentive for using molecular techniques in tree breeding is to improve the rate of genetic gain by reducing the long generation interval since MAS allows early selection before the traits of interest (e.g. wood quality) are expressed. However, Butcher (2002) noted that "MAS has yet to be incorporated in operational breeding programs for plantation species" and she referred to the high costs of genotyping, the large family sizes required to detect QTLs and the lack of knowledge of QTL interactions with genetic background, tree age and environment as explanatory factors.
In a recent review of biotechnology in forestry, Yanchuk (2002) also highlighted the potential advantage of early selection using MAS, but again pointed out that MAS is not yet being routinely applied in tree breeding programmes, largely "because of economic constraints (i.e. the additional genetic gains are generally not large enough to offset the costs of applying the technology). Thus it is likely that MAS will only be applied for a handful of species and situations, e.g. a few of the major commercially used pine and Eucalyptus species. Molecular markers are therefore primarily an information tool and are used to locate DNA/genes that can be of interest for genetic transformation, or information on population structure, mating systems and pedigree confirmation".
Again, much effort has been put into the development of molecular marker maps in this sector. The first reported map in livestock was for the chicken in 1992 which was quickly followed by publication of maps for cattle, pigs and sheep. Since then, the search for useful markers has continued and further species have been targeted, including the goat, horse, rabbit and turkey (see http://www.thearkdb.org/ for the current status regarding some major farm animal species). Microsatellite markers have been of major importance.
Dekkers (2003) recently reviewed commercial applications of MAS in livestock and showed that several gene or marker tests are available on a commercial basis, in different species and for different traits, and that the majority of uses involve GAS, where an important gene (e.g. responsible for a congenital defect) has been identified or, to a lesser degree, LD-MAS. He pointed out that documentation is poor since "although several genetic tests are available, the extent to which they are used in commercial applications is unclear, as is the manner in which they are used and whether their use leads to greater response to selection". He concluded that "opportunities for the application of MAS exist, in particular for GAS and LD-MAS and, to a lesser degree, for LE-MAS because of greater implementation requirements. Regardless of the strategy used, successful application of MAS requires a comprehensive integrated approach with continued emphasis on phenotypic recording programs to enable QTL detection, estimation and confirmation of effects, and utilization of estimates in selection. Whereas initial expectations for the use of MAS were high, the current attitude is one of cautious optimism".
Molecular marker maps have been constructed for a number of aquaculture species e.g. tilapia, catfish, giant tiger prawn, kuruma prawn, Japanese flounder and Atlantic salmon, although their density is generally low. Density is highest for the rainbow trout, where the map published in 2003 has over 1300 markers spread throughout the genome - the vast majority are AFLPs but also including over 200 microsatellite markers. Some QTLs of interest have been detected (e.g. for cold and salinity tolerance in tilapia; for specific diseases in rainbow trout and salmon). Sonesson (2003), in a recent review of MAS in fish breeding schemes, suggested that MAS would be especially valuable for traits that are impossible to record on the candidates for selection, such as disease resistance, fillet quality, feed efficiency and sexual maturation and concluded that MAS is not used in fish breeding schemes today and that the lack of dense molecular maps is the limiting factor.
Molecular marker maps, the necessary framework for any MAS programme, have been constructed for the majority of agriculturally important species. Density of the maps varies considerably between species. Currently, MAS does not play a major role in genetic improvement programmes in any of the agricultural sectors. The enthusiasm and optimism concerning the potential contributions that MAS offers for genetic improvement still remains. However, they seem to be tempered by the realisation that it may take longer than originally thought and that genetic improvement of quantitative traits using MAS may be more difficult than previously considered. The conclusions from the review by Dekkers and Hospital (2002) are a good reflection of this: "Further advances in molecular technology and genome programmes will soon create a wealth of information that can be exploited for the genetic improvement of plants and animals. High-throughput genotyping, for example, will allow direct selection on marker information based on population-wide LD. Methods to effectively analyse and use this information in selection are still to be developed. The eventual application of these technologies in practical breeding programmes will be on the basis of economic grounds, which, along with cost-effective technology, will require further evidence of predictable and sustainable genetic advances using MAS. Until complex traits can be fully dissected, the application of MAS will be limited to genes of moderate-to-large effect and to applications that do not endanger the response to conventional selection. Until then, observable phenotype will remain an important component of genetic improvement programmes, because it takes account of the collective effect of all genes".
4. Some Factors Relevant to Applying MAS in Developing Countries
In the debate on the role or value of MAS as a potential tool for genetic improvement in developing countries, some of the potential factors that should be considered are briefly described below, as they may influence applications of the technology.
4.1 Economic factors
As with any new technology promising increased benefits, the costs of application must also be considered. According to Dekkers and Hospital (2002), "economics is the key determinant for the application of molecular genetics in genetic improvement programmes. The use of markers in selection incurs the costs that are inherent to molecular techniques. Apart from the cost of QTL detection, which can be substantial, costs for MAS include the costs of DNA collection, genotyping and analysis". For example, Koebner (2003) suggested that the current costs of MAS would need to fall considerably before it would be used widely in wheat and barley breeding. In practice, therefore, although MAS may lead to increased genetic responses, decision-makers need to consider whether it may be cost-effective or whether the money and resources spent on developing and applying MAS might instead be more efficiently used on adopting other new technologies or on improving existing conventional breeding programmes.
Little consideration has been given to this issue. Some results have, however, been recently published from studies at the International Maize and Wheat Improvement Center (CIMMYT) in Mexico on the relative cost-effectiveness of conventional selection and MAS for different maize breeding applications. One application, considered by Morris and co-workers (2003), was the transfer of an elite allele at a single dominant gene from a donor line to a recipient line. Here, conventional breeding is less expensive but MAS is quicker. For situations like this, where the choice between conventional breeding and MAS involves a trade-off between time and money, they suggested that the cost-effectiveness of using MAS depends on four parameters: the relative cost of phenotypic versus marker screening; the time saved by MAS; the size and temporal distribution of benefits associated with accelerated release of improved germplasm and, finally, the availability to the breeding program of operating capital. They conclude that "all four of these parameters can vary significantly between breeding projects, suggesting that detailed economic analysis may be needed to predict in advance which selection technology will be optimal for a given breeding project".
In the different applications considered by CIMMYT, the costs of developing molecular markers associated with the trait of interest were not considered, as it was assumed that they were already available. There is a distinction between development costs (e.g. identifying molecular markers on the genome, detecting associations between markers and the traits of interest) and running costs (typing individuals for the appropriate markers in the selection programme) of MAS. Development costs can be quite expensive, so developing countries need to consider whether to develop their own technology or, alternatively, to import the technology developed elsewhere, if available.
Another aspect to be considered here is how to evaluate the economic benefits of MAS. For a publicly funded breeding programme, it should include economic benefits to farmers from genetic improvement of their plants or animals. For private companies, instead, the impacts of using MAS on their market share, and not on rates of genetic improvement, would be of greatest interest.
4.2 MAS versus conventional methods
Although conventional breeding programmes, relying on phenotypic records, have their limitations, they have shown over time that they can be highly successful. Application of MAS will not occur in a vacuum and the potential benefits (genetic, economic etc.) of using MAS need to be compared to those achieved or expected from any existing conventional breeding programmes.
In the different agricultural sectors, this question has received much attention from researchers. There seems to be general consensus that the relative success of MAS compared to conventional breeding may depend on the kind of trait (or traits) to be genetically improved. If the trait is difficult to record or is not routinely recorded in conventional programmes, MAS will offer more advantages than if it is routinely recorded. Similarly, if the trait is sex-limited or can only be measured late in life then MAS is favoured, as marker information can be used in both sexes and at any age.
4.3 MAS versus other biotechnologies for genetic improvement
The relative costs and benefits of applying MAS should be compared not only with conventional breeding but also with potential use of other new technologies that can genetically improve agricultural populations. These include tissue culture in crops and forest trees; reproductive technologies (e.g. embryo transfer or cloning) in livestock and triploidisation or sex-reversal in farmed fish. These also include genetic modification (GM), a technology that can be applied to all sectors. Compared to GM, regulation of MAS, be it at the level of research and development, field testing, commercial release or import/export of developed products, is more relaxed and, in addition, acceptance of the technology by the public is not an issue.
4.4 IPR issues
As discussed in Conference 6 of this Forum, the issue of intellectual property rights (IPRs) is playing an ever greater role on food and agriculture in developing countries. Participants in that conference, among other things, suggested that it was influencing, generally in the negative sense, the quality of agricultural research carried out and the nature of research collaborations between the public and private sector and between developing and developed countries.
It is therefore obvious that IPRs may also impact MAS in developing countries. The impact may be felt at a number of steps involving development and application of markers for genetic improvement. For example, the AFLP molecular marker mapping technique is patented. Molecular markers can be patented, although this can often be overcome by using other markers near the gene of interest. Individual genes can also be patented. With IPRs, however, there is nevertheless public disclosure of the invention or information. Non-disclosure of information, where patents are not sought but the information on markers or detected QTLs is nevertheless kept secret, can also have negative impacts, by denying developing countries access to potentially useful information.
Babu, R.C. et al. 2003. Genetic analysis of drought resistance in rice by molecular markers: Association between secondary traits and field performance. Crop Science 43:1457-1469.
Butcher, P.A. 2002. Molecular breeding of tropical trees. Paper presented at the conference on "Advances in Genetic Improvement of Tropical Tree Species", Yogyakarta, Indonesia, 1-3 October 2002.
Dekkers, J.C.M. and F. Hospital. 2002. The use of molecular genetics in the improvement of agricultural populations. Nature Reviews: Genetics 3:22-32.
Dekkers, J.C.M. 2003. Commercial application of marker- and gene-assisted selection in livestock: strategies and lessons. Paper presented at the 54th annual meeting of the European Association for Animal Production, Rome, Italy, 31 August - 3 September 2003.
Koebner, R. 2003. MAS in cereals: Green for maize, amber for rice, still red for wheat and barley. *
Korzun, V. 2003. Molecular markers and their applications in cereals breeding. *
Morris, M. et al. 2003. Money matters (II): costs of maize inbred line conversion schemes at CIMMYT using conventional and marker-assisted selection. Molecular Breeding 11: 235-247.
Paterson, A.H. et al. 1988. Resolution of quantitative traits into Mendelian factors by using a complete linkage map of restriction fragment length polymorphisms. Nature 335: 721-726.
Sonesson, A.K. 2003. Possibilities for marker-assisted selection in fish breeding schemes. *
Tartarini, S. 2003. Marker-assisted selection in pome fruit breeding. *
Yanchuk, A. 2002. The role and implications of biotechnology in forestry. Forest Genetic Resources 30: 18-22.
Young, N.D. 1999. A cautiously optimistic vision for marker-assisted breeding. Molecular Breeding 5: 505-510.
* Paper presented at an international workshop on "Marker assisted selection: A fast track to increase genetic gain in plant and animal breeding?" in Turin, Italy, 17-18 October 2003http://www.fao.org/biotech/Torino.htm.
- Appreciation is expressed to Oene Dolstra and Dirk-Jan de Koning for their comments on this document.
6. Topics to be Discussed in the Conference
This conference considers the subject of molecular marker assisted selection as a potential tool for genetic improvement of crops, forest trees, livestock and fish in developing countries. In particular, some items we would like to see discussed here are:
- How useful is MAS as a tool for genetic improvement in developing countries? - For which traits and types of species is it most appropriate?
- For which agricultural sectors (crops, forestry, livestock, aquaculture) is it most appropriate?
- For which production systems is it most appropriate?
- What are the current limiting factors to its successful application in developing countries?
- How can these limiting factors be overcome?
- What impacts are IPRs having on development of MAS tools and applications of MAS in developing countries?
- When should developing countries play an active role in the development of MAS technology (construction of molecular marker maps, detection of association between molecular markers and traits of interest etc.) or when, instead, should they aim to import the technology developed elsewhere?
- When is it appropriate for developing countries to use MAS? Should the previous establishment of a successful conventional breeding programme be a prerequisite?
- How appropriate is MAS as a tool for genetic improvement in developing countries compared to other biotechnologies?
- What role should international organisations, like FAO or the World Bank, or the Consultative Group on International Agricultural Research (CGIAR) centres have in this area?
Abbreviations: AFLP = Amplified fragment length polymorphism; CIMMYT = International Maize and Wheat Improvement Center; FAO = Food and Agriculture Organization of the United Nations; GAS = Gene assisted selection; GM = Genetic modification; IPRs = Intellectual property rights; LD = Linkage disequilibrium; LE = Linkage equilibrium; MAS = Marker assisted selection; PCR = Polymerase chain reaction; QTLs = Quantitative trait loci; RAPD = Random amplified polymorphic DNA; RFLPs = Restriction fragment length polymorphisms; SNPs = Single nucleotide polymorphisms
FAO, 10 November 2003.