Chapter 19: Technical, economic and policy considerations on marker-assisted selection in crops

Chia sẻ: Ngo Van Quang | Ngày: | Loại File: DOC | Số trang:47

Thêm vào BST

Báo xấu

92
lượt xem 10
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

Due to their usefulness in characterizing and manipulating genetic factors responsible for qualitative as well as quantitative traits,molecularmarkersareconsideredto be valuable tools for crop improvement. Theseusesofmolecularmarkershavebeen invaluable in helping researchers understand complex traits, dissect them into singleMendeliangeneticfactors,andestablishtheirchromosomallocationsviatheuse oflinkagemapsand/orcytogeneticstocks. Availability of well characterized genetic linkage maps is a prerequisite for tagging important agronomic or other traits with molecular markers, enabling their use in MAS related activities. To date, however, few practical applications have been published from these studies. ...

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Chapter 19: Technical, economic and policy considerations on marker-assisted selection in crops

Chapter 19 technical, economic and policy considerations on marker-assisted selection in crops: lessons from the experience at an international agricultural research centre H. Manilal William, Michael Morris, Marilyn Warburton and David A. Hoisington
382 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish Summary Molecular markers and related technologies have been used extensively in genetic characterization and identification of loci controlling traits of economic importance in manycropspecies.However,theapplicationofsuchtoolsforcropimprovementhasnot beenextensive,atleastinthepublicsector.Althoughthereareclearadvantagesinusing molecularmarkersastoolsforindirectselectionoftraitsofimportance,availableexamples indicate that their use is restricted to traits with monogenic inheritance or when the inheritanceisconditionedbyafewgeneswithlargeeffects.Anotherimportantlimitationof large-scalemarkerapplicationsisthecostinvolvedinmarkerassays,whichmaybebeyond thecapacitiesofmanypublicplantbreedingenterprises.Foraneffectivemarker-assisted selection(MAS)activitytofacilitateongoingcropimprovementprogrammes,especiallyin thecontextofthedevelopingcountries,laboratorieswithadequatecapacityandadequately trainedscientificpersonnelaswellasoperationalresourcesarerequired.Althoughrecent technological advances such as single nucleotide polymorphisms (SNPs) and associated assayprotocolsarelikelytoreduceassaycostssignificantly,formanyoftheseoperations, assayplatformswithsignificantcapitalinvestmentsincludingcomputationalcapacityare required. Coupled with these limitations, private sector domination of biotechnology research with proprietary rights to important products and processes with immediate benefits to developing countries may further constrain the benefits these technologies may offer to resource-poor farmers. Policy-makers in different national programmes andinternationaldevelopmentandresearchagencieshavearesponsibilitytosustainand augmentthecapacityofnationalpublicagriculturalresearchorganizationstoensurethat biotechnologytoolsandprocessesareinfusedappropriatelyintonationalresearchefforts. Theymustalsoensurethatanybiotechnologyeffortsundertakenarewellintegratedwith nationalcropimprovementactivities.
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 383 introduCtion anyMAS-relatedactivity.Thus,whileitis Due to their usefulness in characterizing possibletocarryoutMAStosomedegree and manipulating genetic factors respon- incerealssuchasrice,maizeandwheat,and sible for qualitative as well as quantitative inlegumessuchassoybean,forspeciessuch traits,molecularmarkersareconsideredto as cassava and sweet potato, the so-called be valuable tools for crop improvement. “orphancrops”,geneticimprovementwith Theseusesofmolecularmarkershavebeen MAS may not yet be feasible. These crop invaluable in helping researchers under- species may benefit more readily from stand complex traits, dissect them into genetic modification arising from direct singleMendeliangeneticfactors,andestab- introduction of genes isolated from other lishtheirchromosomallocationsviatheuse speciesororganisms,whichisnotthefocus oflinkagemapsand/orcytogeneticstocks. ofthischapter. Availability of well characterized genetic Citing practical lessons learned at linkage maps is a prerequisite for tagging the International Maize and Wheat important agronomic or other traits with ImprovementCenter(CIMMYT)aswellas molecular markers, enabling their use in findingsofstudiesconductedelsewhere,this MAS related activities. To date, however, chapterdescribessomeactualandpotential few practical applications have been pub- applications as well as the advantages and lished from these studies. This paucity of disadvantages of MAS, and outlines pos- published studies may indicate the long- sible applications of MAS in developing term nature of this research, or it might countryplantbreedingprogrammes. simply reflect the fact that marker tech- nologyhasbeenappliedtoplantbreeding leSSonS learned from CropS effortsmostlybyscientistsworkinginthe Numerous scientific reports describe privatesector(HoisingtonandMelchinger, molecular mapping and analysis of quan- 2004). titative trait loci (QTL) for nearly every Maizewasoneofthefirstcropspecies agronomic trait in a diverse array of crop for which molecular linkage maps were species. The traits covered include many developed, and Gardiner et al. (1993) parameters associated with tolerance to consolidatedseveralindividualmapsintoa droughtandotherabioticstresses,maturity, consensusmap.Riceisanotherspeciesfor plantheight,qualityparameters,qualitative whichhigh-densitylinkagemapshavebeen andquantitativefactorsofdiseaseandpest developed(reviewedinGowdaet al.,2003) resistance, and numerous seed traits and while,duetoitshighploidylevelandlarge yield.Althoughtheseeffortshaveresulted genome(21linkagegroups,asopposedto10 inavastamountofknowledgeandbetter inmaizeand12inrice),effortstodevelop understanding of the underlying genetic well characterized, saturated linkage maps factorsthatcontrolthesetraits,application with wheat have lagged behind. Other ofthisknowledgetomanipulategenesinan importantcerealsandlegumesareatvarious effective or simple manner for improving stages of linkage map development. The crop species has had limited success. The availability of well-defined linkage maps scientific community is faced with the andtheextentofgeneticstudiesconducted challenges of accurate and precise QTL on them therefore vary among different identificationandapplicationoftheinfor- crops,andthisinfluencesthefeasibilityof mation derived to successful MAS efforts.
384 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish Scientificadvanceshavebeeninstrumental traitofinterest.Comparedwithtraditional in increasing the power and accuracy backcrossing,theuseofDNAmarkersena- of computational parameters as well as blesfasterrecoveryoftherecurrentparent designing ways of combining the infor- genotypealongwiththeintrogressedtarget mation generated from molecular genetics trait in line conversion activities. Ribaut withtraditionalcropimprovementefforts. andHoisington(1998)reportedthatMAS Numerous simulation studies have been should enable the recovery of the target undertakentoevaluatetheeffectivenessof genotypeafterthreecyclesofbackcrossing, MAS,takingintoaccounttheinfluenceof compared with a minimum of six cycles heritability, population size, linkage dis- with traditional approaches (Tanksley et tance, etc. (Xie and Xu, 1998; Moreau et al.,1989). al., 1998; Ribaut, Jiang and Hoisington, CIMMYT has a long history of using 2002), and MAS procedures have been molecular markers for certain traits in used to incorporate traits of interest from maize improvement. Although maize is exoticspeciesincludingwildrelativesinto widelyusedforbothfoodandfeed,maize elitecultivarsthroughadvancedbackcross kernels do not provide sufficient quanti- QTLanalysis(TanksleyandNelson,1996; tiesoftwoessentialaminoacids,lysineand Fultonet al.,2000). tryptophan.Theopaque2 mutation,identi- fiedatPurdueUniversity(UnitedStatesof manipulation of qualitative traits America)inthemid-1950s,conferselevated Molecular markers that are tightly linked levels of these two amino acids. Although to genes having a strong effect on the initial efforts to introduce the opaque2 expression of a trait can be used to intro- mutationintobreedingmaterialswerenot gress the genes (and thus the trait) into successful(Villegas,1994),researcherseven- different backgrounds through backcross tuallysucceededinproducingnutritionally breeding schemes that rapidly and effi- enhanced maize lines. These came to be ciently improve the recurrent parent for known as quality protein maize (QPM). the target trait. In conventional backcross CIMMYT breeders have used traditional breeding schemes and line conversion backcrossingtotransfertheopaque2 muta- activities, the donor parent containing the tion and associated modifiers into elite traitofinterestiscrossedwiththerecurrent lines. To perform phenotypic selection in parent, normally a well-adapted variety segregatingprogeniesforlinescarryingthe lacking the trait of interest. The resulting opaque2 mutation,itisnecessaryeitherto progeny are screened to identify the trait wait until the plants produce mature ears, of interest, and individuals exhibiting the or to do random pollination on a large trait are crossed to the recurrent parent. number of plants. Although reliable labo- Theentireprocessisrepeatedseveraltimes. ratory screening techniques are available, For traits that are conditioned by reces- co-dominantmicrosatellitemarkerspresent sive gene action, a cycle of selfing is also within the opaque2 mutation can be used required after each crossing cycle. After earlier in the growing season. Using these several cycles of backcrossing and a final markersinbackcrossprogenies,plantshet- self-pollination, plant breeders are often erozygousfortheopaque2 mutationcanbe abletorecoverlinesthatarenearlyidentical selectively identified as a qualitative trait totherecipientparentbutalsocontainthe foruseinthenextcrossingcycle.Markers
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 385 are not used to select for the background line. After screening the F progeny and 2 recurrent parent genotypes, but only to F families, lines identified by markers 3 selectlinescarryingtheopaque2 mutation were sent to Africa, where MSV is preva- allele. Although CIMMYT uses markers lent. By phenotypic screening of the lines for detecting the presence of the opaque2 selected by MAS, it was established that mutation,markersarenotavailabletoselect MAS-selectedlinesweresignificantlymore for the modifiers, which are important in resistant to MSV (J-M. Ribaut, personal determining seed texture and quality and communication). forwhichothertraditionalscreeningtech- In legumes, resistance to soybean cyst niquesarebeingused. nematode (SCN) is one example of an A well known example of marker- effective MAS approach. Routinely used assisted backcrossing of a qualitative trait phenotypicassaysforSCNscreeningtake involves the introgression of the Bt trans- approximately five weeks and extensive gene into different maize lines (Ragot et greenhouse space and labour. Successful al., 1994). Whenever plant transformation identification of closely linked microsat- techniquesareusedtoproducegenetically ellite markers has enabled transfer of the modifiedorganisms(GMOs),usuallythere resistance gene rhg1 with about 99per- are some cultivars that are more receptive cent accuracy (Cregan et al., 1999; Young to transformation procedures than others. 1999). Many public and commercial soy- Whenthecultivarwiththebestagronomic bean cultivar improvement efforts use typeisnotthemostreceptivetotransfor- these markers to screen for SCN resist- mation, it is often possible to transform ance (Young, 1999). Another example of another cultivar that is receptive and then successful MAS in common beans was use the diagnostic marker that detects the reported by Yu, Park and Poysa (2000) transgene to introgress it into more desir- whousedmarkersassociatedwithcommon ablebackgrounds.ThistypeofMAS-aided bacterialblight.Thesemarkersidentifieda line conversion can be accomplished for locusthatexplainedabout62percentofthe anycropspecies.Thepresenceofmarkers phenotypicvariationandhavebeenusedin to detect the transgene enables the detec- MASexperiments. tion of converted progeny with a high As described earlier, linkage map con- degreeofaccuracy. struction in wheat is more challenging Another MAS-related CIMMYT expe- than in species such as rice or maize. rience involves the case of maize streak The allohexaploid nature allows wheat to virus (MSV) resistance, for which a major withstand chromosomal imbalances as the QTLwasidentifiedonmaizechromosome loss of one chromosome can be compen- 1thatexplains50–70percentoftotalphe- sated by the presence of a homologous notypic variation (Pernet et al., 1999a, b). chromosome. As a result, wheat can be As maize has a well-saturated molecular crossedwitharangeofwildrelatives(both linkagemap,severalmicrosatellitemarkers intergeneric and interspecific), enabling associated with this QTL were identified introgressionofgeneticmaterialpossessing inthespecificchromosomalregion.These resistances to different biotic and abiotic markers were tested in three populations stresses. When translocations (especially generated using three different MSV tol- intergeneric translocations) are present in erant lines crossed with one susceptible wheat, markers can be readily developed
386 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish forthetranslocatedchromosomesegments. locatedchromosomesegmentallowsthe If a translocated segment carries a trait of alien-derived resistance to be combined importance, markers can then be used to with the BYDV tolerance available in transferitintodifferentwheats.Diagnostic wheat. orperfectmarkers(i.e.markerswithcom- • Marker for Aegilops ventricosa-derived plete linkage to the genes of interest with resistance to stripe rust (Yr17), leaf rust nopossibilityofrecombination)havebeen (Lr37)andstemrust(Sr38 )(O.Robert, developed for genes conferring resist- personalcommunication).Thetransloca- ance to different biotic stresses in wheat. tion from Ae. ventricosa is present on CIMMYT’swheatimprovementeffortsuse chromosome2AS.Thediagnosticmarker a set of markers routinely on a seasonal for the translocation is used mainly in basisforintrogressionofasetofgenesinto bread wheat x durum wheat crosses, to high-yielding backgrounds. Examples of identify the durum derivatives carrying the perfect markers that are currently in thetranslocation. useare: In addition, CIMMYT uses a set of • Cereal cyst nematode (CCN) resistance linkedmarkersfortransferringalocuswith gene Cre1 (2BL), identified in wheat major effects for boron tolerance (Bo-1), landrace AUS10894 and Cre3 (2DL), crown rot resistance, scab resistance and derivedfromTriticum tauschii (Lagudah, stem rust resistance in its MAS efforts. MoulletandAppels, 1997).Thesemark- These efforts with linked genes are con- ersareusedroutinelyinsegregatingpop- ductedwiththeobjectiveofincreasingthe ulationstoenableselectiveadvancement allelefrequencyfordesirableallelesinseg- oflinescontainingtheCre genestargeted regating populations (William, Trethowan to all environments, but particularly to andCrosby-Galvan,2007). marginalones,wherehealthyrootarchi- tectureisessentialtoallowplantstotake gene pyramiding/stacking advantageofminimalsoilmoisture.Phe- MAS lends itself well to gene pyramiding notypic evaluation for CCN resistance efforts for disease resistance. When a cul- is labour intensive as well as expensive. tivar is protected by one gene with major Given that it is impossible to screen effectsagainstaspecificdisease,itisoften for CCN resistance in Mexico (where notpossibletointrogressadditionalgenes CIMMYTheadquartersarelocated)due conferring resistance to the same disease to the lack of required screening facili- because of the difficulty of phenotypic ties, the use of markers is essential for screening for the presence of additional improvingthistrait. genes(astheplantalreadyshowsresistance • Barleyyellowdwarfvirus(BYDV)resis- to the disease). However, if several genes tance, derived from a chromosome seg- canbetaggedwithcloselylinkedmolecular ment introgressed from Thinopyrum markers, MAS strategies can be used to intermedium, on chromosome 7DL developlineswithstackedgenes,givingthe (Ayala et al.,2001).BYDVisanimpor- cultivarmoredurableprotectionthanthat tantviraldiseaseincertainwheatgrow- affordedbyasingleresistancegene. ingregionsoftheworld.Environmental Resistance to bacterial blight provides influencemakesfieldscreeninglessreli- anexcellentexampleofusingMASforgene able.Thediagnosticmarkerforthetrans- pyramiding. Bacterial blight is caused by
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 387 Xanthomonas oryzae andis oneofthemost several genomic regions simultaneously, important diseases of rice. Several genes each having different effects on the same that confer resistance to bacterial blight trait of interest, MAS-based approaches havebeentaggedwithmolecularmarkers. become more complicated and present Huang et al. (1997) and Hittalmani et al. formidablechallenges.Mappingstudiescon- (2000) developed strategies for combining ductedatCIMMYTidentifiedfivegenomic four resistance genes, namely Xa-4, Xa-5, regionsassociatedwiththeanthesis-silking Xa-13 andXa-21,inasinglecultivarusing interval which is a parameter associated pairwise combinations of the genes. Due with drought tolerance in maize (Ribaut to the co-dominant nature of the markers et al., 1996, 1997). The drought tolerant used,theauthorswereabletoselectfrom parent was used in MAS experiments as F generations without having to perform 2 thedonorparenttotransferthefiveQTL progeny testing. The derived lines con- toCML247,aneliteinbredlinewithgood taining pyramided genes showed higher combining ability that was drought-sus- levelofresistanceand/orawidespectrum ceptiblebuthigh-yieldingunderfavourable of resistance compared with the parental conditions.Markerswereusedtogenerate material.Anothergenepyramidingexample 70 BC F lines containing the favourable 2 3 usingMASinvolvesstackingoftheresist- alleles from the drought-resistant parent ance genes rym4, rym5, rym9 and rym11 after two backcrosses and two self pol- for the barley yellow mosaic virus com- linations. These lines were crossed with plexusingmolecularmarkersanddoubled twotestersforfieldevaluation.Fieldtests haploids(Werner,FriedtandOrdon,2005). indicated that under severe drought stress Other examples include pyramiding for conditions,the70MAS-derivedlineswere barley stripe rust resistance (Castroet al., significantly better yielding than the con- 2003), and powdery mildew resistance in trols. The differences were less prominent wheat(Liuet al.,2000)and,inMASappli- under reduced drought stress (Ribaut and cations at CIMMYT, crosses have been Ragot,2007). madetocombinetwogenesforcerealcyst Other CIMMYT experiments aimed at nematode resistance and three different comparingMASwithphenotypicselection genes for stem rust resistance (Sr24, Sr26 have been conducted for stem borers in andSr25)intargetedwheatgermplasm. tropicalmaize(Willcoxet al.,2002).Inthe case of maize stem borer resistance, three manipulation of quantitative traits QTL identified through mapping experi- Quantitatively inherited traits are geneti- ments were used in MAS. Three BC S 2 2 callycomplex,areconditionedbyanumber families that carried all three target QTL ofgeneseachhavingrelativelysmalleffects, fromthedonorparentinhomozygousstate andtheirexpressionoftendependsoninter- weredeveloped.Comparativestudieswith actionsamongdifferentgeneticcomponents MAS and traditional phenotypic selection (epistasis).Theenvironmentalsohasahigh didnotestablishaclearadvantageforMAS, degreeofinfluenceontheexpressionofthe but both approaches yielded significant trait,whichconfoundstheinterpretationof geneticgainsinreducingleafdamage.MAS QTL identification and often renders the is not being used currently on a routine results obtained from QTL studies cross- basis at CIMMYT for drought and stem specific.Whenitisnecessarytomanipulate borerresistance.
388 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish Other reports describing the manip- markershavealsobeenusedforidentifying ulation of quantitatively inherited traits redundanciesinexistinggermplasmcollec- include those of Bouchez et al. (2002) tions in rice (Xu, Beachell and McCouch, for introgressing favourable alleles at 2004)andsorghum(Dean et al.,1999).In three QTL for earliness and grain yield cassava, Chavarriaga-Aquirre et al. (1999) in maize, and by Yousef and Juvik (2001) usedmorphologicaltraits,isozymeprofiles who reported on MAS for seedling emer- andagronomiccriteriatoidentifyacoreset genceandeatingqualitycharactersinsweet of630accessionsfromabasecollectionof corn.Also,Han et al.(1997)attemptedto approximately5500accessions. selectforbarleymaltingtraitsusingMAS. Modern farming in advanced countries Additional scientific reports are available is based on high performing, genetically thatdescribeMAS-relatedeffortsforquan- uniform new cultivars, which are gener- titativelyinheritedtraits. allyderivedfromwelladapted,genetically In general, manipulating several QTL related parental material. Tanksley and associatedwithmultiplegenomicregionsin McCouch(1997)haveconcludedthatmost segregatingprogeniesisconsiderablymore modern soybean cultivars grown in the challenging. Often the success in genetic UnitedStatescanbetracedbacktoavery gainsdependsonthestabilityoftheseQTL limitednumberofstrainsfromasmallarea aswellasthecostefficiencyoflarge-scale ofnortheasternChina,whileamajorityof MASapplications. hardredwinterwheatsisderivedfromafew linesoriginatedinPolandandtheRussian genetic diversity studies Federation. The genetic basis of modern In addition to being used in MAS activi- ricevarietiesgrownintheUnitedStatesis ties, molecular markers have been used alsoconsiderednarrow(Dilday,1990). extensively for genetic diversity studies. Another application in the area of Numerousscientificpublicationsareavail- genetic diversity is the use of markers able that describe the use of molecular in identifying heterotic groups. Molecular markersinestimatingthedegreeofrelated- markershavebeenusedextensivelyinthe nessofasetofcultivarsinmanycultivated construction of heterotic groups since the cropspecies.Incommonwiththeirusein 1990s in many different crop species of traitmanipulations,thepracticaloutcomes economicimportance.Heteroticgroupsare of the numerous genetic diversity studies clustersofgermplasmusuallywithsimilar using molecular markers are not clear. characteristicsandahighdegreeofrelated- Evaluation of genetic relatedness using nessthat,whencrossedwithmaterialsfrom molecular markers will have implications another heterotic group, tend to give rise on understanding the genetic structure of to progeny with high levels of heterosis. existingpopulations,enablingthedesignof Althoughmarkersrandomlydistributedin strategies for proper acquisition of germ- the genome can be used to develop heter- plasm for conservation purposes. The oticgroups,theirusefulnessindetermining genetic uniqueness of accessions or popu- hybridperformanceisnotclear.Whileitis lations in germplasm collections can be reasonabletoassumethatheterosisdepends accurately estimated by the use of DNA ontheinteractionsamongfavourablealleles profiling (Brown and Kresovich, 1996; belongingtothetwoparents,unlessmolec- Smith and Helentjaris, 1996). Molecular ularmarkersthatareknowntobelinkedto
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 389 thesefavourableallelesareusedinheterotic scientists have been conducted to assist in studies,thepredictivepowerofmarkersin regeneratinggenebankaccessionswithout estimating heterosis for practical applica- losing genetic diversity, measuring the tionsmaynotbeveryhigh. contribution of wild ancestors and exotic At CIMMYT, large-scale, rapid char- species in advanced backcross progenies acterization methods for inbred lines and ofsyntheticbreadwheat,andtotrackthe populations have been optimized using changes over time in diversity levels of up to 120 microsatellite markers spread CIMMYTwheatcultivarsfromtheoriginal throughoutthemaizegenome.Inthepast, Green Revolution varieties to modern characterizingmaizepopulationswascostly breedinglines. andtime-consuming,giventhatasmanyas 22individualshadtobeanalysedindividu- marker implementation allytocalculateallelefrequenciesforeach To facilitate the use of MAS activities in marker. Currently, a bulking method in wheat and maize improvement efforts, which 15 individuals from a population CIMMYT has recently established a areamplifiedinthesamepolymerasechain marker implementation laboratory. This reaction (PCR) and run on an automatic provides the facilities and technical exper- DNA sequencer, provides a reliable esti- tisetoprovideCIMMYTwheatandmaize mate of the allele frequencies within that breeders with access to biotechnology particular population. Between one and tools,includingMAS.Thelaboratorycar- two bulks can now be used to fingerprint ries out two main MAS-related activities, populations with considerable savings in markeradoptionandresearchsupport.The timeandresources.Otherstudiesofmaize first includes constantly reviewing the lit- genetic diversity have been conducted for erature to identify markers developed by CIMMYTmaizebreedersaswellasoutside thirdpartiesandverifyingthatthesecanbe collaborators with objectives that include: usedtodetecttraitsorgenesofinterestin determining how maize inbred lines from CIMMYTgermplasmimprovementefforts, different national breeding programmes anddevelopingefficientprotocolsfortheir arerelatedtoeachother(andtodetermine in-houseuse.Thesecondconsistsofarange the possibility of sharing among regions ofroutinetasksthatincludegrowthand/or or using lines from one region to expand sampling of plant tissue, DNA extraction, diversity in another); establishing heter- markerdetection,dataanalysisanddissem- otic groups; determining levels of genetic inationofresultstobreeders. diversity present in synthetic varieties; Close cooperation between field and determining how landraces and farmers’ laboratory staff is important to be able to varieties from different regions are related applymolecularmarkersincropimprove- to each other; monitoring homozygosity mentefforts.Ideally,laboratorystaffshould levelsininbredlines;andtrackingchanges haveanunderstandingoffieldactivitiesand inlinesthathavebeenintensivelyselected field workers should have basic knowl- foragiventrait. edgeofdifferentaspectsofMAS-associated Acoresetof100microsatellitemarkers laboratoryprocedures.MASisusedwhen hasbeenselectedforwheatgeneticdiversity thereisahighprobabilitythatmarkerswill studies. Recent fingerprinting studies help plant breeders achieve genetic gains by CIMMYT and national programme faster and more economically than field
390 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish or laboratory-based phenotypic selection MASiswhethertoinvestinbiotechnology methods. When perfect markers are avail- researchcapacity. able to screen for a particular trait, such Economictheorysuggeststhatthemost markers are preferred. However, for traits efficient level of research investment can thatcannotbescreenedconvenientlyusing be determined with the help of a research traditionalapproachesandevenwhenper- production function that relates research fectmarkersarenotavailable,ifmarkersare inputstoresearchoutputs.Atthenational availablewithcloselinkagestothetrait(s) level,theresearchproductionfunctioncan of interest, these can be used to increase be thought of as a meta-function encom- thedesirableallelefrequencyforthetarget passing the frontiers of many smaller gene.MAS-relatedactivitiesinbothwheat functions, each representing a different and maize at CIMMYT are conducted levelofresearchcapacitydistinguishedby as collaborative projects involving both complexityandscope(Figure1)(Brennan breedersandbiotechnologists.Thebreeders 1989; Byerlee and Traxler, 2001; Maredia, use information coming from wheat MAS Byerlee and Maredia, 1999; Morris et al., activities to define better their parental 2001). Movement outwards along the crossingblockmaterialsandtomakeselec- meta-function, accomplished by adding tive crosses using parents identified by subprogrammes and thereby increasing markers. Moreover, segregating early gen- the number of researchers and the extent eration progenies in certain crosses are ofavailableresearchinfrastructure,isasso- selected in the field based on whole plant ciatedwithchangesinfocusandincreases phenotype,whicharethenfurtherrefined in the capacity of the national research by sampling leaf tissue from field-tagged programme. plantsandprocessingforMASassaysinthe Foraplantbreedingprogramme,adding laboratory.Onlythoseentriesthatcontain new biotechnology-based subprogrammes the target genes identified with MAS are is equivalent to taking a series of dis- advancedtothenextgeneration.Thisena- crete steps involving increased complexity blesbreederstoreducepopulationsizesfor and cost. These steps have the effect of thetraitsunderevaluationandaccumulate moving the programme from one level of certain gene combinations in elite back- research capacity to the next. These levels grounds. The material thus generated is ofresearchcapacitycanbebroadlycharac- advanced through several cycles of selfing terizedasfollows: and eventually used in field screening to • Biotechnology product user. Here, the identifythebestperforminglines. researchprogrammeimportsgermplasm productsdevelopedusingbiotechnology eConomiCS of maS and incorporates them into its conven- establishing the capacity to conduct tionalcropimprovementschemes,either maS by backcrossing them into local germ- ForMAStobeaviableoptionforaplant plasm or by testing them for potential breedingprogramme,adequatelyequipped immediaterelease. laboratory facilities must be in place and • Biotechnology tools user where the appropriately trained scientists must be research programme imports bio- available. Therefore, one of the first deci- technology tools and uses them, if sionsfacingresearchmanagersconsidering necessary, after adapting them to local
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 391 FiGURe 1 Biotechnology research production meta-function PM C Biotechnology innovator ts u Biotechnology tp ou B tools user Research Biotechnology product user A P1 P2 P3 0 Research inputs Source: Morris et al. (2001). circumstances, to improve current crop research capacity can be expected to gen- improvementpractices. erate enough additional benefits to justify • Biotechnology methods innovator, in theadditionalexpenditure.Formostplant which the research programme estab- breeding programmes, benefits consist of lishes the full capacity needed to devel- value added to crop production enter- op innovative biotechnology tools and prises. Therefore, the incentive to invest products. inadditionalresearchcapacitywilltendto Movingfromonelevelofbiotechnology increase with the size of the area planted researchcapacitytothenextusuallyrequires and/or the value of the crops expected to significantinvestmentsinlaboratoryfacili- benefitfromtheresearch. tiesandstafftraining.Thepracticaldecision There are few published estimates of facingresearchmanagersisnottodetermine the cost of moving from one level of bio- theoptimallevelofresearchinvestment,but technology research capacity to the next, rather to select from among the different and new estimates are not provided here. levels of biotechnology research capacity Empirical estimates would quickly be characterizedbyincreasingcomplexityand outdated, as the cost of biotechnology cost(AorBorCinFigure1).Thechoice laboratory equipment and materials con- shouldbebasedonwhetheragivenlevelof tinues to change very rapidly. However,
392 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish forthepurposesofthischapteritisimpor- Althoughthecostsoffieldoperationsand tanttopointoutthatalthoughestablishing laboratoryproceduresrequiredformolec- capacitytodevelopnewmolecularmarkers ularmarkeranalysismayremainrelatively requiressubstantialinvestment,establishing constantacrossapplications,everybreeding thecapacitytousefreelyavailableexisting project is likely to involve unique pheno- molecularmarkersrequiresonlyamodest typic evaluation procedures whose costs investment. willfrequentlydiffer. Second, direct comparisons of unit variable cost of maS costs for phenotypic and genotypic anal- At CIMMYT the capacity to carry out ysisprovideusefulinformationtoresearch MASonareasonablescalehasbeendevel- managers, but in many cases technology oped, but the need now is to make the decisionsarenotmadesolelyonthebasis technology work on a high-throughput ofcost.Factorsotherthancostofteninflu- scale to reduce the cost per data point, encethechoiceofscreeningmethods.Time whilebeingabletohandlelargequantities considerations are often critical, as geno- ofassayspergrowingseason.Inthisregard, typic and phenotypic screening methods there are several challenges to consider maydifferintheirtimerequirements.Even as markers are not always cost-effective when labour requirements are similar, for even when they improve the precision of applicationsinwhichphenotypicscreening selection. Depending on the nature of the requiressamplesofmaturegrain,genotypic targettrait(quantitativeorqualitative),the screening can often be completed much typeofgene(majororminor),theformof earlierintheplantgrowthcycle. geneactionthatcontrolsexpressionofthe Third, conventional and MAS methods trait(dominantorrecessiveeffect),andthe are not always direct substitutes. Using easewithwhichthetraitcanbemeasured molecular markers, breeders may be able (visually detected or more expensive field to obtain more information about what is or laboratory analysis required), conven- goingonatthegenotypiclevelthantheycan tionalselectionmaybecheaperthanMAS. obtainusingphenotypicscreeningmethods. The desirability of using genetic markers For example, in conventional backcross therefore depends in part on the costs breeding or line conversion projects (see of genotypic versus phenotypic screening, sectionManipulation of qualitative traits), whichvaryamongapplications. backgroundmolecularmarkerscanbeused Information about the cost of using to identify those plants among a set of MAS at CIMMYT for specific breeding progeny that not only possess a desirable projects is available from case studies. allele but also closely resemble the recur- For example, Dreher et al. (2002, 2003) rent parent at the genetic level. Based on examined the costs and benefits of using this additional information, breeders are MAS for a common application in maize often able to modify their entire breeding breeding.Thisstudygeneratedthreenote- strategy, with potentially significant worthyconclusions. implications in terms of cost and/or time First, for any given breeding project, requirements(thisissueisdiscussedinthe detailed budget analysis is needed to nextsection). determine the cost-effectiveness of MAS The CIMMYT case study thus con- relativetoconventionalselectionmethods. firmedwhatmanypractisingplantbreeders
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 393 intuitively know: namely, the costs and the only one. Plant breeders worry about benefitsofMASprojectsarelikelytovary controllingcosts,buttheyalsoworryabout dependingonthecropbeingimproved,the gettingproductsoutquickly.Therefore,it breedingobjectivebeingpursued,theskill is not sufficient to consider potential cost ofthebreeder,thecapacityoftheresearch savings alone. The time requirements of organization, the location of the work alternativebreedingstrategiesmustalsobe being carried out, the cost of key inputs, takenintoaccount,becauseevenwhenMAS andmanyotherfactors. costsmorethanconventionalselection(as it does in some, although not all, cases), economic trade-offs breeders who use it may be able to gen- While caution is required when extrapo- erateadesiredoutputquicker.Accelerated lating from the results of a case study, release of improved varieties can translate general conclusions regarding the cost- intolargebenefits,especiallyfortheprivate effectivenessofmolecularmarkersincrop seedindustry,sotimeisanimportantcon- genetic improvement work can be drawn siderationinadditiontocost. based on the findings of the CIMMYT ForbreedingapplicationsinwhichMAS studyandanumberofotherstudiescarried offers cost and time savings, the advan- outelsewhere.Broadlyspeaking,twotypes tagesofMAScomparedwithconventional ofbenefitsassociatedwithMAScanbedis- breedingareclear.Moreproblematic,how- tinguished:costsavingsandtimesavings. ever, are the many applications in which MAS methods cost more to implement Cost savings than conventional selection methods but Forcertainapplications,MASmethodscan also reduce the time needed to accom- substitutedirectlyforconventionalselection plishabreedingobjective.Thiscommonly methods,andfortheseapplicationstherel- happens, for example, with inbred line ativecost-effectivenessofthetwomethods conversionschemesbasedonbackcrossing caneasilybedeterminedbycomparingthe procedures.Insuchschemes,MASmethods screeningcostpersample.Generally,asthe can often be used to derive converted costofphenotypicscreeningrises,markers inbredlinescontainingoneormoreincor- are more likely to represent a cost-effec- poratedgenesinmuchlesstimethanwould tive alternative. For applications in which be possible using conventional selection phenotypic screening is easy and cheap methodsalone. (e.g. visual scoring of plant colour), MAS In applications that involve a trade- will not offer any obvious advantages in off between time and money, under what terms of cost. However, for applications circumstances is the higher cost of MAS in which phenotypic screening is difficult relative to conventional breeding justi- or expensive (e.g.assessing root damage fied? The choice of the plant breeding causedbynematodesorforadiseasethatis method can be viewed as an investment notpresentinthefieldsite),MASwilloften decision and evaluated using conventional bepreferable. investment criteria (Sanders and Lynam, 1982).UsingdatafromtheCIMMYTcase Time savings study, Morris et al. (2003) explored the Cost is an important factor affecting the relationship between time and money as choiceofbreedingtechnology,butitisnot it relates to crop improvement research
394 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish FiGURe 2 Stylized economic model of a plant breeding programme 500 000 Varietal adoption stage 400 000 300 000 Net annual bene?ts (US$) 200 000 Research Varietal 100 000 investment release stage stage 0 -100 000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Time (years) by developing a simple model of a plant efits turn positive as the variety is taken breedingprogrammeandusingittocom- up and grown by farmers; they continue pare the returns with alternative inbred toincreaseuntilthepeakadoptionlevelis lineconversionschemesbasedonconven- achievedandthendeclinewhenthevariety tionalselectionandMAS.Twomeasuresof isreplacedbynewervarieties. project worth were used: the net present The model was used to estimate the value (NPV) of the discounted streams of NPVandIRRofconventionalandmarker- costsandbenefits,andtheinternalrateof assisted inbred line conversion schemes. return(IRR)totheinvestment. Research cost data were taken from the Figure2depictsthestylized“varietylife CIMMYTcasestudy.Plausiblevalueswere cycle”assumedbythemodel.Thestreamof usedforkeyparametersrelatingtothevari- costsandbenefitsassociatedwiththedevel- etalreleaseandadoptionstages(fordetails, opment,releaseandadoptionbyfarmersof seeMorriset al.,2003).Figure3showsthe an improved variety can be divided into streamsofannualnetbenefitsgeneratedby threestages:aresearchstageduringwhich eachofthetwobreedingschemes.Annual the variety is developed; a release stage netbenefitsarecalculatedasfollows: during which the variety is evaluated and registeredforrelease,andcommercialseed NB =(GB -VR -RC ) t t t t isproduced;andanadoptionstageduring which the variety is taken up and grown where: by farmers. During the first two stages, NB= netbenefits net benefits are negative, because costs GB= gross benefits (calculated as area are incurred without any benefits being plantedtothevarietyxincremental realized. During the third stage, net ben- benefitsassociatedwithadoption)
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 395 FiGURe 3 annual net benefits, conventional vs. marker-assisted line conversion scheme 400 000 Conventional 350 000 selection scheme 300 000 MAS scheme 250 000 200 000 150 000 100 000 50 000 0 1 2 3 4 5 -50 000 6 7 8 9 1011 12 13 1415 16 17 1819 20 Time (years) breeding schemes, MAS and conven- tional, were found to differ depending VR= varietal release expenses (cost of evaluation trials, registration pro- cedures, seed multiplication, advertisingandpromotion,etc.) RC= researchinvestmentcosts t= year(1…n) NPVswerecalculatedbyaddingthedis- counted stream of net benefits associated witheachbreedingschemeoverthelifeof thevariety(nyears): where: n NPV = Σ ( GB t - VR t - RC t ) / (1 + r)t t=1 NPV= netpresentvalue r= discountrate IRRswerecalculatedconventionallyby solving the discount rate that drives the NPVto0. The profitability rankings of the two
tions involving trade-offs between time andmoney.Fromaneconomicperspective, on the measure of project worth that was the relative attractiveness of conventional used. The MAS scheme generated the versus MAS methods will depend on the highest NPV, whereas the conventional availability of research investment capital. breeding scheme generated the highest Ifinvestmentcapitalisabundant(meaning IRR on investment. These results, gen- thatthebreedingprogrammecanaffordto erated using a stylized model of a plant absorbthehigherup-frontcostsassociated breeding programme and plausible values withMASwithoutcurtailingotherongoing for varietal release and adoption param- breeding projects), MAS may become a eters,provideanimportantinsightintothe desirable option, because it generates the relative cost-effectiveness of conventional largestNPV.Ontheotherhand,ifinvest- selection methods and MAS in applica- mentcapitalisconstrained(i.e.thebreeding Annual net bene?ts (US$)
396 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish programme cannot absorb the higher up- become prime candidates to apply MAS- front costs associated with MAS, or that basedintrogressionoftheintroducedgene itcanabsorbthemonlybyforgoingother construct/stootherwelladaptedcultivars potentiallyprofitablebreedingprojects),it indifferentagro-ecologicalregions. makessensetochooseconventionalselec- Reports indicate that two rice varieties tion,becauseitgeneratesthelargestIRR. with improved bacterial blight resistance havebeendevelopedwithMASapproaches impliCationS for developing and deployed in Indonesia (Toenniessen, CountrieS O’Toole and DeVries, 2003). Moreover, When discussing policy implications of rice varieties carrying multiple disease MAS efforts in developing country resistance genes are being developed by scenarios, it is appropriateto consider the severalnationalprogrammeswithtechnical experience gained over the past several backstopping by the International Rice decades,mainlyinindustrializedcountries. Research Institute (IRRI) (Hittalmani et In advanced countries, the private sector al.,2000).Therearealsoreportsdescribing has made significant investments in MAS theuseofMASinChinaforimprovingcer- efforts while there are a few publicly- tainqualitytraitsinrice(Zhou,P.H.et al., funded research groups using MAS in 2003)andwheat(Zhou,W-C. et al.,2003) breedingroutinelyandthesearerestricted andfibrerelatedtraitsincotton(Zhang et to a few target crops (Eagles et al., 2001; al.,2003),butitisnotclearwhetherthese Dubcovsky,2004;William,Trethowanand are one-time research efforts or there is Crosby-Galvan, 2007). Information about continuedactivityusingMAS. the traits and the breeding strategies used Although it is not possible to obtain in MAS applications in large agribusiness entirely reliable estimates of the costs, enterprisesarenotpubliclyavailablefreely. benefits and cost-effectiveness of MAS To date, significant investments have been applications,thecostsassociatedwithMAS made in biotechnology applications only arefrequentlyconsideredasthemaincon- forwidelygrowncropspeciessuchasrice, strainttotheireffectiveusebymanyplant maize,wheat,soybean,cottonandcanola. breeders, especially in small- to medium- WhileGMcropsandtheirimplicationsare scale breeding enterprises. However, new notthefocusofthischapter,itisreasonable marker technologies, especially those toassumethattechnologiesassociatedwith basedonsinglenucleotidepolymorphisms GM crops offer significant potential for (SNPs) and associated ongoing large-scale addressingbioticandabioticstresstolerance genomesequencingprojects,shouldenable inwidelygrowncerealsandlegumesaswell thedevelopmentanddeploymentofgene- as species that are important but thus far basedmarkersinthenearfuture(Rafalski, neglectedsuchastef,millets,yamsandother 2002).SNPsaredefinedassinglebasediffer- tubercropsinthedevelopingcountries.For enceswithinadefinedsegmentofDNAat example, GM technologies that can make correspondingpositions.TheseSNP-based one crop species perform better are likely polymorphisms are known to be abun- to be valuable with slight modifications dantlypresentinhumanaswellasinplant toenhancetheperformanceofaneglected genomes.Consequently,thepotentialexists cropspecies.WhenusefulGMvarietiesofa to develop SNP markers associated with particularcroparemadeavailable,theyalso manyimportanttraitsinadiversearrayof
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 397 economically important crop species. For traits to be detected early, well before species such as maize, rice and soybeans, thetraitisexpressedandcanbedetected robustSNP-basedassayplatformsalready phenotypically. This benefit can be par- existintheprivatesectoraswellasinsome ticularly important in species that grow publicsectorenterprises.Theaddedadvan- slowly,forexample,treecrops. tage of SNP-based marker systems is that • Recessive genes.MASofferspotentialsav- they avoid gel-based allele separations for ings compared with conventional selec- visualization and have the potential for tion when it allows breeders to identify automationinhigh-throughputassayplat- heterozygousplantsthatcarryarecessive forms.Theseongoingresearcheffortswill allele of interest whose presence cannot inevitablyleadtothedevelopmentofmore bedetectedphenotypically.Intraditional robust, high-throughput assays that are breeding approaches, an extra step of bothsimpleandcosteffective(Jenkinsand selfing is required to detect phenotypes Gibson,2002). associatedwithrecessivegenes. • Heritability of traits.Uptoapoint,gains when is it advantageous to use maS? fromMASincreasewithdecreasingheri- In addition to the cost and time savings tability.However,duetothedifficulties describedabove,foranumberofbreeding encounteredinQTLdetection,thegains scenarios,MASmethodsarelikelytooffer are likely to decline beyond a certain significantadvantagescomparedwithcon- thresholdheritabilityestimate. ventionalselectionmethods.Thesescenarios • Seasonal considerations. MAS offers assumetheavailabilityofmarkersformul- potential savings compared with con- tipletraitsandtakeintoconsiderationthe ventional selection when it is necessary advantages of MAS under optimum situa- to screen for traits whose expression tions(Dreheret al.,2002;Dudley,1993). depends on seasonal parameters. Using • Gene stacking for a single trait. MAS molecular markers, at any time of the offers potential savings compared with year, breeders can screen for the pres- conventional selection when it allows ence of an allele (or alleles) associated breederstoidentifythepresenceofmul- with traits that are expressed only dur- tiplegenes/allelesrelatedtoasingletrait, ing certain growing seasons. For exam- and the alleles do not exert individually ple, CIMMYT’s wheat breeding station detectable effects on the expression of in northern Mexico is usually used for the trait. For example, when one gene screeningsegregatinggermplasmforleaf confers resistance to a specific disease rust resistance. However, expression of orpest,breederswouldbeunabletouse leaf rust is not uniform in all growing traditional phenotypic screening to add seasons. The same concept is true for anothergenetothesamecultivarinorder field screening for drought tolerance. to increase the durability of resistance. Whenthereareseasonswithlowexpres- In such cases, MAS would be the only sion of leaf rust or less intense drought feasible option, provided markers are due to unexpected rainfall, markers, if availableforsuchgenes. available,canbeavaluablealternativeas • Early detection.MASofferspotentialsav- atoolforscreening. ings compared with conventional selec- • Geographical considerations.MASoffers tion when it allows alleles for desirable potential savings when it is necessary
398 Marker-assisted selection – Current status and future perspectives in crops, livestock, forestry and fish to screen for traits whose expression investment.Tocomplicatemattersfurther, dependsongeographicalconsiderations. theprivatesectordominatesmanyfieldsof Using molecular markers, breeders in biotechnology research and therefore has onelocationcanscreenforthepresence proprietaryrightstomanytechnologiesand of an allele (or alleles) associated with productsthathaveimmediateapplications traitsexpressedonlyinotherlocations. in developing countries (e.g. transgenic • Multiple genes, multiple traits. MAS technology). This is quite different from offers potential savings when there is a conventional plant breeding technologies, needtoselectformultipletraitssimulta- mostofwhichweredevelopedbypublicly- neously.Withconventionalmethods,itis fundedresearchprogrammesandthushave oftennecessarytoconductseparatetrials remainedmoreaccessible. toscreenforindividualtraits. There is no single answer to meeting • Biological security considerations. MAS these challenges, especially as developing offerspotentialadvantagesoverselection countries are not uniform in their public based on the use of potentially harm- agricultural research capacities. Broadly ful biological agents (e.g. artificial viral speaking,developingcountriesfallintothe infections or artificial infestations with followingcategories: insectpests),whichmayrequirespecific • countries (a few) with strong public securitymeasures. sector research infrastructure enabling Inviewoftheabove-mentionedfactors, biotechnology applications, as well as itisdesirabletoconsiderMASapproaches upstream research capability to develop onacase-by-casebasis,takingintoaccount toolsfortheirownspecificneeds; factorssuchastheimportanceofatraitin • countries with intermediate capacity in theoverallbreedingscheme,theamountof applied plant breeding, as well as in availableresourcesintermsofbothstaffand using biotechnology tools that are pub- operationalexpenditures,andthenatureof liclyavailableorcanbeacquiredthrough the breeding materials. There are no “one bilateral partnerships with the private size fits all” recommendations that can be sector; made for MAS approaches. Usually, no • countries (a considerable number) with breedingschemefocusesonimprovingjust moderate plant breeding capacity and onetrait.Atcurrentlevelsofcapacity,MAS practicallyno,orverylittle,capacityfor islikelytobeusedtoachievegeneticgains biotechnologyapplications. for single traits such as host plant resist- More advanced developing countries ance to pests and/or diseases. Therefore, withmajorcommercialfarmingsectorsare MASactivitiesshouldbeintegratedintoan more likely to succeed in adopting agri- overallbreedingprogramme. cultural biotechnology. In addition, the presence of commercial opportunities will Challenges for developing countries attractinvestmentbyprivateindustryand The rapid expansion of agricultural bio- thus allow the country to benefit from technology is generating a wide array of future advances in biotechnology. This is methodologies with potential applica- notalwaysapositiveoutcomeforthepublic tions, and therefore national programmes sector because, as competition increases, in developing countries face the difficult it may be more difficult to justify large challenge of identifying priority areas for public investments in biotechnology. This
Chapter 19 – Technical, economic and policy considerations on marker-assisted selection in crops 399 hasoccurredtosomedegreeinmaizebio- MASforcropimprovement.Biotechnology technology,evenintheUnitedStates. toolscanbeusedtoenhancegeneticgains Developing countries, in which agri- for a few traits in a few crops, but their culture is still dominated by subsistence ultimateimpactdependsonhowwellthey farming and where there is limited or no are adopted and integrated into existing capacity for biotechnology research, are plantbreedingactivities.Thisisasobering at an added disadvantage. Resource-poor thought,becauseinmanydevelopingcoun- farmersinsuchcountriesrarelyofferade- triespublicsectorresearchcapacityisbeing quate market incentives for the private erodedandpublicsectorextensionservices industry that dominates biotechnology arebeingseverelycurtailed. research. For example, the involvement of Otherfactorsessentialforthesuccessful the private sector in research and devel- application of biotechnology tools are opment activities for root crops or grain training and capacity building. Many bio- legumes is doubtful as these crops are technology applications require learning grown mainly by small-scale farmers in new skills, some research infrastructure poorer regions of the world and there and effective operational capacity. It is would be potentially low returns on especially important to train and nur- investment.Therefore,itisimportantthat ture national scientists capable of using international development agencies ensure emerging technologies. In general, it may that neither the “orphan commodities” not be possible for older plant scientists yielding broad socio-economic benefits, to acquire the capacity for biotechnology northelessadvantagedandleastdeveloped applications. Therefore, policy-makers in countries, are left out from the prospect developingcountrieshavetoconsiderlong- of harnessing potential benefits associ- terminvestmentsintrainingandnurturing atedwithbiotechnology.Indoingso,they anewgenerationofscientifictalent.They must evaluate what biotechnology tools also need to consider how to utilize this canbeofimmediatebenefittosuchcrops talent effectively by providing adequate and countries and then develop strategies resources and optimum work environ- leadingtosuccessfuladoptionbythetarget ments. Specialized technical training must groups. This can only be accomplished inturnbeunderpinnedbycomplementary if the efforts made are serious, long-term governmentinvestmentsinbasiceducation, and sustainable. Many examples can be e.g. by including biotechnology-related citedwhereinternationalaidagencieshave subjectsinnationaluniversitycurricula. investedinpurchasingequipmentdesigned Although it is widely assumed that for biotechnology research in developing enormousinvestmentsareneededtoestab- countries but, when the aid programmes lishacapacitytocarryoutMAS,thisisnot terminate their short-term involvement, alwaystrue.Certainly,aminimumlevelof thecapitalinvestmentseitherhavenotbeen investment is needed for laboratory facili- optimallyutilizedorhaveremainedidle. ties,equipmentandtrainedstaff.However, Policy-makers in different national consideringthatmostMASworkindevel- programmes must also bear in mind that oping countries is likely to be geared sustained capacity in public agricultural towardstheuseofexistingmarkersrather research is a pre-requisite for successful thanthedevelopmentofnewones,invest- applicationofbiotechnologytoolsincluding ments in facilities and capital need not be