关 键 词：

马铃薯 红薯 收获 设计 仿真 proe 三维 12 CAD

资源描述：

马铃薯-红薯收获机的设计与仿真含proe三维及12张CAD图,马铃薯,红薯,收获,设计,仿真,proe,三维,12,CAD

内容简介：

RESEARCH ARTICLECharacterization of the genetic diversity of Ugandas sweetpotato (Ipomoea batatas) germplasm using microsatellitesmarkersBarbara M. ZaweddeMarc GhislainEric MagembeGeovani B. AmaroRebecca GrumetJim HancockReceived: 7 April 2014/Accepted: 1 September 2014/Published online: 17 September 2014? Springer Science+Business Media Dordrecht 2014AbstractKnowledge about the genetic diversity andstructure of crop cultivars can help make better conser-vation decisions, and guide crop improvement efforts.Diversity analysis using microsatellite markers wasperformed to assess the level of genetic diversity insweet potato in Uganda, and evaluate the geneticrelationship between the Ugandas germplasm andsome genotypes obtained from Kenya, Tanzania,Ghana, Brazil and Peru. A total of 260 sweet potatocultivarswerecharacterizedusing93microsatelliteloci.The Ugandan collection showed a large number ofdistinct landraces, and very low (3 %) levels of geneticdiversitybetweengenotypesobtainedfromthedifferentagro-ecological zones. There was low (6 %) levels ofgenetic diversity observed between the East Africangenotypes; however unique alleles were present incollections from the various sources. Pairwise compar-isons of genetic differentiation indicated that Ugandasgermplasmwassignificantlydifferent(P0.001)fromcultivars from Tanzania, Ghana, Brazil and Peru. Thepresence of unique alleles in populations from variousUgandas agro-ecological zones and other globalregions, as well as the regional diversity patterns,suggestthateffortsshouldbemadetofurthercollectandcharacterize the germplasm in more depth.KeywordsCharacterization ? Crop breeding ?Ipomoea batatas ? Molecular markers ? SSRIntroductionA germplasm collection of crop cultivars with varyingenvironmental adaptive capacity can be both a sourceof genes for future crop improvement, as well as acritical resource for farmers. The highest levels ofgenetic diversity for the majority of the importantglobalfoodcropsisintheSouth,wherecropcentersoforigins are commonly found, and centers of diversityemerged due to prolonged periods of farmer selection(FAO 2008).B. M. Zawedde ? R. Grumet ? J. Hancock (&)Graduate Program in Plant Breeding, Genetics andBiotechnology, Michigan State University, Plant and SoilScience Building, 1066 Bogue Street, East Lansing,MI 48824, USAe-mail: hancockB. M. Zaweddee-mail: zaweddePresent Address:B. M. ZaweddeUganda Biosciences Information Center (UBIC), NationalCrop Resources Research Institute, 27 km Kampala -Zirobwe Road, Namulonge, Kampala, UgandaM. Ghislain ? E. MagembeCIP Sub-Saharan Africa, International Potato Center,P.O. Box 25171, Nairobi, KenyaG. B. AmaroEmbrapa Vegetable Crops, P.O. Box 218, Bras lia,DF CEP 70359-970, Brazil123Genet Resour Crop Evol (2015) 62:501513DOI 10.1007/s10722-014-0175-5Sweet potato, Ipomoea batatas (L.) Lam., is thefifth most important food crop in terms of weightharvested in Eastern Africa (FAO 2012). Sweet potatowas introduced to the East African borders from SouthAmerica by Portuguese explorers during the 16thcentury (Zhang et al. 2004). The oldest remains ofsweet potato have been found in the caves of theChilca Canyon in Peru and dated as 8,000 years old(Lebot 2010). However, based on morphologicalrelationships among related species, the center oforigin appears to be between the Yucatan Peninsula inMexico and the Orinoco River in Venezuela (Austin1977). It is also in that region that the wild species ofthe section Batatas, considered as putative ancestorsand wild relatives of the cultivated sweet potato, arefound (Andersson and de Vicente 2010). Evaluationsof genetic diversity patterns among germplasm fromdifferent parts of the world have resulted in thesuggestion that China, Southeast Asia, New Guineaand East Africa are secondary centers of diversity(Yen 1982; Austin 1983).Uganda has the highest production per capita inSub-Saharan Africa and numerous, diverse sweetpotato landraces are grown. In 2005, the nationalsweet potato program collected over 1,300 landraces,which were characterized using morphological meth-odologies to determine the level of genetic diversityand 946 of these were found to be morphologicallydistinct genotypes (Yada et al. 2010a). This high levelof diversity can be attributed primarily to theallogamous and hexaploidy nature of sweet potato(Lebot 2010), as well as variations in farmerspreferences (Veasey et al. 2008). The method ofpropagation by vine cuttings contributes also indi-rectly by maintaining cultivar diversity.Knowledge about the genetic diversity and struc-ture of existing crop cultivars can aid in making betterconservation decisions, and help direct breedingprograms. Characterization of crop diversity can beachieved through morphological and molecular tools.Morphological characterization is an important firststep in assessment of diversity; however there aremajor limitations in relying only on morphologicalcharacterization including low levels of polymor-phism, low repeatability, late expression for certaintraits; phenotypic plasticity and parallel evolution(Karuri et al. 2010; Yada et al. 2010b). A number ofmolecular markers including random amplified poly-morphic DNAs (RAPDs), restriction fragment lengthpolymorphism (RFLPs), amplified fragment lengthpolymorphism (AFLP), microsatellites or simplesequence repeats (SSRs), single nucleotide polymor-phisms (SNPs) have been developed and used tocomplement morphological characterization. Selec-tion of any particular DNA marker in a crop dependslargely on the objective of the research, availableresources and technical skills (Otoo et al. 2009).In sweet potato research, a number of molecularmarkers have been used to study the crops geneticdiversity including RAPDs (Connolly et al. 1994;Gichuki et al. 2003; He et al. 2006), DNA amplifica-tion fingerprinting (DAF) (He et al. 1995), AFLPs(Zhang et al. 2004, Elameen et al. 2008), inter simplesequence repeats (ISSRs) (Hu et al. 2003), selectiveamplificationofmicrosatellitepolymorphicloci(SAMPL) (Tseng et al. 2002), and SSRs markers(Gichuru et al. 2006; Veasey et al. 2008) andmicrosatellites or SSRs (Jarret and Bowen 1994;Buteler et al. 1999; Hu et al. 2004; Yada et al. 2010b;Tumwegamire et al. 2011).Sweet potato is a hexaploid (2n = 6x = 90) cropbelieved to have originated from natural hybridiza-tion between several wild species (Lebot 2010).There are three hypotheses for generation of hexa-ploid sweet potato: autopolyploidy (Kobayashi 1983;Shiotani and Kawase 1987, 1989); allo-autopolyp-loidy (Schafleitner et al. 2010) and allopolyploidy(Austin 1977; Nishiyama 1971; Srisuwan et al. 2006;Gao et al. 2011). The hexaploid nature of sweetpotato and the complexity of its genomic makeupcreate a challenge when using molecular markers forgermplasm characterization (Lebot 2010). However,SSR markers have proved to be useful in unveilinggenetic relationships that are closely similar tomorphological characterization in hexaploid Guineayam Dioscorea rotundata Poir. (Mignouna et al.2003) and detection of polymorphism in sweet potato(Veasey et al. 2008). Small sets of SSR markers, aslow as 10 primer pairs, have been found to efficientlytrace the movement of specific alleles betweenpopulations (Lebot 2010), and distinguish distinctindividuals within a population (Yada et al. 2010b).In Uganda, two studies have previously used SSRmarkers to analyze the genetic diversity of the sweetpotato germplasm in Uganda. One study focused onassessed the genetic relationship among 192 selectedsuperior landraces (high yielding, sweet potato virusdisease or Alternaria blight disease resistance, or high502Genet Resour Crop Evol (2015) 62:501513123dry matter content) and concluded that there were 190distinct landraces (Yada et al. 2010b). The secondstudy used 75 cultivars from the same collection toassess the genetic relationship between East Africanorange-fleshed and the white-fleshed landraces, andhow they related to cultivars from China, USA, PapuaNew Guinea and Peru (Tumwegamire et al. 2011).They found that the orange-fleshed types which werefirst developed in Kenya were distinct from those ofnon-African origin, andthat orange-fleshedand white-fleshed landraces from East Africa are closely related.Since 1995, the national breeding programme hasreleased twenty cultivars (Mwanga et al. 2011). Mostof these cultivars (Sowola and NASPOT 1 to 11) wereselected from bulked seed from a polycross between18 and 24 parents.ThepurposeofthisstudywastouseSSRmarkerstodetermine the structure of genetic diversity in Ugan-dans sweet potato, to assess genetic relationshipbetween landraces and the released cultivars inUganda, as well as to relate it to germplasm fromother parts of the world. This information can be usedto make recommendations that will improve andpossibly enhance efficient low-cost conservation ofsweet potato diversity.Materials and methodsPlant materialUsing the collection database maintained by the Sweetpotato Program in the National Crop ResourcesResearch Institute at Namulonge, a total of 168Ugandan genotypes were selected based on threecategories: response to weevil infestation (Muyinzaet al. 2012); agro-ecological zone (Yada et al. 2010a);and varietal classification (Mwanga et al. 2011).Landraces were selected from the five major sweetpotato growing agro-ecological zones of Uganda:northernregion;easternregion;centralregion;westernregion;andsouthernregion.Todeterminetheirgeneticrelationship with genotypes from other parts of theworld, the Ugandan genotypes were compared withgenotypes from other sweet potato growing Africancountries, Brazil and Peru. Improved cultivars wereselected from bulked seed of three different polycross-es, each containing between 18 and 24 parents(Mwanga et al. 2003, 2009, 2011). Planting materialof selected cultivars was grown in a screen house for2 months and young leaves were put immediately onice and stored at -80 ?C until DNA extraction.DNA extraction and simple sequence repeat (SSR)amplificationGenomic DNA was isolated from 200 mg of frozen leaftissueusingamodifiedCTABmethod(DoyleandDoyle,1990), according to (Yada et al. 2010b). A total of 308DNA samples were quantified, diluted to 20 ng/ll andprepared for polymerase chain reaction (PCR) amplifi-cation. 31-labeled SSR primer pairs used in previoussweet potato studies (Karuri et al. 2010; Yada et al.2010b;Tumwegamireetal.2011)forDNAamplificationwere tested andonly 19 SSR markers (Table 1) wereselected and used in this study. Other SSR markers wererejected due to non-specific amplification or for beingmonomorphic.Atotalreactionvolumeof10 llcontain-ing 1 ll of the 10X PCR buffer (10 mM TrisHCl pH9.0, 30 mM KCl, 1.5 mM MgCl2), 0.1 ll of lyophilized5 U/ll of Taq DNA polymerase, 0.2 ll of 10 mMdNTPs, 0.5 ll of 10 lM of the primers, 2.5 ll of 20 ng/ll DNA template and 5.7 ll of double distilled water,wasusedforPCR.ThePCRprogramwassetasfollows:94 ?C for 3 min for initial denaturation, followed by 35cycles each consisting of denaturation at 94 ?C for 30 s,annealing at the indicated temperature for each primerpair (Table 1) and polymerization at 72 ?C for 30 s, andthenfinalextensionat72 ?Cfor20 min.AmplifiedPCRproducts were prepared for each DNA sample and usedfor fragment analysis using an ABI 3730 capillarysequencer (Applied Biosystems).Allele scoring and data analysisGenemapper v3.7 software (Applied Biosystems) wasused for peak detection and fragment size estimation.AlleloBin software (Prasanth et al. 1997) was used tocorrectanyerrorsinthescoredallelesduetoslippageofDNA polymerase during PCR (Schlotterer and Tautz1992). Analyses were performed using two methods ofdata coding that were previously employed to describediversity in polyploids (Esselink et al. 2004; Jrgensenet al. 2008; Kloda et al. 2008; Garc a-Verdugo et al.2009;SampsonandByrne2012).Firstly,themultilocusdata were transformed into binary arrays of thepresence/absence of an allele for each individual, usingALS-Binary software (Prasanth and Chandra 1997).Genet Resour Crop Evol (2015) 62:501513503123Secondly, the MAC-PR (microsatellite DNA allelecounting-peakratios)method(Esselinketal.2004)wasemployedtoobtainallele-dosageof16markersforeachof the polyploidy individuals. The MAC-PR methodmakes use of the quantitative values for peak areasprovided by the GeneMapper software. For each locus,all alleles were analysed in pairwise combinations inorder to determine their copy numbers in the individualsamples. This was accomplished by calculating ratiosbetween peak areas for two alleles in all samples wherethese two alleles occurred together. Access to individ-uals with six different alleles or three different allelesprovided a baseline for calculating the single- anddouble copy dosage, respectively.The binary data was used to determine thepolymorphic information content (PIC), which is themeasure of the usefulness of each marker in distin-guishing between individuals that was formulated byWeir (1996) as:PICi 1 ?Xnj1Pij2where PICiisthe PIC of amarker i; Pijisthe frequencyof the jth pattern for marker i, and the summationextends over n patterns. The standard measures ofgenetic diversity for each population included numberof polymorphic loci (P), percentage of polymorphicloci (%P) and Neis (1973) gene diversity (D),estimated from binary data using GenAlex 6.4 (Peak-all and Smouse 2006).Analysisofmolecularvariance(AMOVA)wasalsoperformed using GenAlEx version 6.4 to estimate thetotal variance and distribution of diversity within andbetween populations. Wrights F-Statistic (FST, fixa-tion index) was also computed, using GenAlExsoftware, to estimate the amount of genetic variancethat can be explained by population structure (Hol-singer and Bruce 2009).Fixation index; FSTHT? HIHTwhere HIis the mean observed heterozygosity perindividual within subpopulations and HTis theexpected heterozygosity in a random mating totalpopulation.FSTcanrange from 0.0 (nodifferentiation)to 1.0 (complete differentiation, that is, subpopula-tions fixed for different alleles).The phylogenetic relationship among populationswas assessed using DARwin version 5 (Perrier andJacquemoud-Collet 2006). Similarity matrices wereconstructed from the binary data with JaccardsTable 1 Characteristics ofSSR markers used toevaluate East African sweetpotato cultivars: name,labeled dye, motifs,annealing temperature andreferenceaTa- annealingtemperature;bunpublisheddata developed from 2002to 2003 at CIP,cunpublished datadeveloped from 2005 to2006 at CIPNameDyeMotifTaaReferenceIB-S076-FAM(tgtc)760Benavides (unp.)bIB-R03PET(gcg)558Benavides (unp.)IB-R16VIC(gata)459Benavides (unp.)IB-R13NED(ttc)658Benavides (unp.)J10APET(aag)6TD (5762)Solis et al. (unp.)cJ175VIC(aatc)4TD (5762)Solis et al. (unp.)IBCIP-96-FAM(cca)2ac(acc)6TD (5060)Yan ez (2002)IBCIP-13NED(acc)3 ? (cgg)2 ? (tgc)3 ? (gtc)2TD (5762)Yan ez (2002)IB-R19PET(cag)5bTD (5762)Benavides (unp.)BU6919846-FAM(tgg)6TD (5762)Hu et al. (2004)JB1809VIC(cct)6(ccg)6TD (5060)Solis et al. (unp.)IBSSR09NED(gaa)5(gag)3TD (5762)Hu et al. (2004)IB-R08PET(t3a)4TD (5060)Benavides (unp.)BU690524VIC(cag)5TD (5762)Hu et al., 2004BU6907086-FAM(ccg)5TD (5762)Hu et al. (2004)1B-S186-FAM(tagc)4TD (5762)Benavides (unp.)IBCIP-16-FAM(acc)7aTD (5762)Yan ez (2002)IBSSR07PET(ct)7a(tc)4TD (5762)Hu et al. (2004)IB-R12NED(cag)5aTD (5762)Benavides (unp.)504Genet Resour Crop Evol (2015) 62:501513123coefficients(Jaccard1908).Jaccardscoeffi-cient = Nab/(Na? Nb), where Nabis the number ofalleles shared by two individuals a and b, Nais totalnumber ofalleles in samplea,andNbistotal number ofalleles in sample b. Genetic distances between popula-tions were obtained by computing the usual Euclidiandistance matrix based on haplotype frequencies. Fromthis matrix, a dendrogram was constructed using theneighbor joining method (NJ) from Saitou and Nei(1987).The significance ofeachnodewas evaluated bybootstrappingdataoveralocusfor5,000replicationsofthe original matrix. We examined hierarchical geneticvariation between individuals using the un-weightedpair group method analysis (UPGMA), as suggested bySneath and Sokal (1973).Clustering patterns of individuals and populationswere examined using STRUCTURE version 2.3.3(Pritchard et al. 2000), which is reported to have thecapability to generate population structuring (Prit-chard et al. 2009). Using the allele dosage (MAC-PR)data for each individual, individuals were assignedprobabilistically to genetic clusters (K). The STRUC-TURE program was run using no prior assumptions ofpopulationstructurewithanadmixtureancestrymodeland the recommended methods for recessive alleles,and allele frequencies correlated. The analysis wasused to determine whether biologically relevant clus-ters could be determined among the plants sampled,and establish the proportion of an individuals genome(Q) that originated from each cluster. For all analyses,the Markov chain Monte Carlo (MCMC) parameterswere set to a burn-in period of 50,000 with 50,000iterations. The optimum K, indicating the number oftrue clusters in the data, was determined from 20replicate runs for each value of K (K set to 10) usingthe method described by Evanno et al. (2005) and theadhocQuantityDeltaK,basedontherateofchangeinthe log probability of the data between successiveK values. Parameters of the method of Evanno et al.(2005) were calculated using the program StructureHarvester version 0.6.92 (Earl and vonHoldt 2012).Similarity among different runs was calculated by themethod of Jakobsson and Rosenberg (2007) as used intheir computer program CLUMPP 1.1.2. This methodcalculates a similarity coefficient h0, which allows theassessment of the similarity of individual runs of theprogram STRUCTURE. The optimal alignment of 20replicates of K values was determined using thecomputer program CLUMPP 1.1.2 (Jakobsson andRosenberg 2007) and clusters were visualized usingthe program DISTRUCT 1.1 (Rosenberg 2004).ResultsSSR markers amplificationA total of 107 alleles were scored for the 19 SSRmarkers (Table 2). The number of alleles per locusranged from 3 to 9. Three markers had very low PIC;IB-S07(0.22),JB1809(0.19)andIBSSR09(0.31)thuswere excluded from further analyses.Determining relatedness between cultivarsin UgandaA total of 10 newly improved cultivars released by thenational program were compared with 158 Ugandanlandraces. The unweighted neighbor joining (NJ)algorithm cluster analysis generated numerous clus-ters (Fig. 1). Improved cultivars were scattered intoTable 2 Observed base pair (bp) range, number of alleles andpolymorphic information content (PIC) for the SSR markersused to characterize sweet potato genotypes from Uganda,Kenya and TanzaniaNamebp rangeNo. of allelesPICIB-S0717317750.22IB-R0324425850.60IB-R1619421250.66IB-R1322629690.77J10A17522980.82J17511214460.81IBCIP-917619440.80IBCIP-1319637470.52IB-R1919120840.63BU69198425226730.67JB180919626440.19IBSSR0919620850.31IB-R0820521640.64BU69052427229380.80BU69070823525970.811B-S1821123540.76IBCIP-113518980.89IBSSR0715817840.70IB-R1230334170.79Genet Resour Crop Evol (2015) 62:501513505123many of these clusters together with landraces.Noteworthy (6/10) improved cultivars were groupedtogetherwithaKenyancultivarKakamega,which waspurposely included in this analysis because it is aknown maternal parent for many of these improvedcultivars.Genetic relationship between genotypesfromUgandasagro-ecologicalzonesandcultivarscollected from other East African countriesAnalysis of Molecular Variance (AMOVA) indicatedthatonly6 %ofthegeneticvariationwasexplainedbydifferences among the sources (Table 3). Analysis ofthe sixteen microsatellites yielded a total of 93presumptive loci in the 228-sweet potato genotypesfrom the eight predefined populations (Table 4). Anaverage of 70 polymorphic loci was observed in eachpopulation. The level of genetic diversity variedamong the different populations. Most regions inUgandahadpopulationswith fewuniquealleles (14),except the south-western region, which had none.Tanzanian cultivars also had few unique alleles (4),but had the highest level of heterozygosity (D).Overall the level of heterozygosity (D) for thecollected samples was low. The significant differencebetween the cultivars from Tanzania and the popula-tions from Uganda and Kenya is clearly shown in thegenetic distance matrix (Fig. 2).Genetic relationship between Ugandan genotypesand cultivars collected from elsewhereAnalysis of Molecular Variance (AMOVA) indicatedthat only 24 % of the genetic variation was explainedby differences among the countries (Table 5). Pair-wise comparisons of genetic differentiation amongcountries indicated that Ugandas germplasm wasWestern UgSouthern UgCentral UgEastern UgNorthern UgImproved UgKenyaTanzaniaFig. 1 Non-metricmultidimensional scalingrepresentation of theprincipal coordinatesanalysis (PCoA) comparinggenotypes from differentUgandas agro-ecologicalzones andotherEast Africancountries. The percentage ofvariation explained by thethree axes is 89.1 %506Genet Resour Crop Evol (2015) 62:501513123significantly different (P0.001) from genotypesfrom Brazil, Peru and Ghana (Table 6). The Jaccardssimilarity coefficients ranged from 0.0 to 0.95 with amean of 0.56. More than 70 % of the pair-wisesimilarity coefficients were between 0.50 and 0.63.The dendrogram generated by DARwin softwarerevealed three clusters (Fig. 2). To efficiently visual-ize the results, the dendrogram was pruned from thecomplete tree to show clustering only between geno-types that had bootstrap values greater than 60. Thispruned tree showed similar broad clustering patternsas the complete tree (data not shown). East Africangermplasm and cultivars from USA were found inCluster A, the majority of cultivars from Brazil andPeru were in cluster B while most Ghanaian cultivarswere found in Cluster C.The Bayesian model of STRUCTURE (Pritchardet al. 2000) assigned the individuals to two majorgenetic clusters, as the highest Delta K was observedat K = 2. All individuals appeared to have a compo-nent of both clusters in their genome; however, theUgandan and Kenyan cultivars had a very highproportion of their genome originating from clusterK1, Brazil, Ghana and Peru had a very high proportionof their genome originating from cluster K2, whileTanzaniancultivarswerecomposedofamixtureofthetwo clusters (Fig. 3).DiscussionThe number of alleles per primer pair observed in thiswork is close to that obtained by Yada et al. (2010b)using the same SSR markers. However, our number ofalleles varies somewhat from those reported by Tumw-egamireetal.(2011)onsimilarsweetpotatogermplasmusing the same markers. Higher number of alleles wasobservedforsomemarkersinthisworkandthisislikelydue to a larger genotype sample size and a higherresolution of DNA fragment. Yada et al. (2010b)assessed 192 samples using the ABI system similar towhat was used in this work, while Tumwegamire et al.(2011) screened 75 samples with the LiCOR system.A total of 92 out of 106 markers were highlypolymorphic, which confirms the high discriminatingpower of the SSR markers (Hwang et al. 2002;Gichuru et al. 2006, Veasey et al. 2008, Yada et al.2010a, b; Tumwegamire et al. 2011). Hwang et al.Table 3 AMOVA for genetic diversity within and among the populations of sweet potato from different Ugandas agro-ecologicalzones and other East African countries (Kenya and Tanzania) as well as Ugandas improved cultivarsSourcedfSSMSEst. Var.%FSTP valueAmong regions7275.2434.410.8160.0560.010Within regions2203,206.7013.6513.6594Total2273,481.9314.45100Table 4 Number of individuals, number of unique alleles, and genetic diversity parametersafor sweet potato populations fromvarious Ugandas agro-ecological zones and other East African countries (Kenya, Tanzania), as well as Ugandas improved cultivarsRegionsIndividualsRAP%PDNorthern Uganda2326975.30.433 0.032Eastern Uganda6617884.40.433 0.029Central Uganda3417379.20.442 0.031Western Uganda4347985.70.449 0.027South-western Uganda1406267.50.431 0.032Improved cultivars1105762.30.347 0.028Kenya1526267.50.457 0.028Tanzania2247884.40.514 0.025Mean7075.80.438 0.010Total2281393aWhere P is total number of polymorphic loci per region, %P is percentage of polymorphic loci, and D is Neis (1973) gene diversityestimated with computer program GenAlex 6.4 (Peakall and Smouse 2006)Genet Resour Crop Evol (2015) 62:501513507123BAC508Genet Resour Crop Evol (2015) 62:501513123(2002) suggested that the high level of polymorphismof molecular markers in sweet potato was due to itslarge genome size and high levels of heterozygosity.Such high levels of heterozygosity would be expectedif the sweet potato is an allopolyploid as many suggest(Austin 1977; Nishiyama 1971; Srisuwan et al. 2006;Gao et al. 2011) and high levels of heterozygositywould be maintained by the common practice ofvegetative propagation of sweet potato. In addition,the mating system of sweet potato, which is outcross-ing due to frequent self-incompatibility, would sup-port high levels of heterozygosity.There was a low mean genetic similarity coefficientof 0.56 between the Ugandan cultivars and cultivarsfrom other African countries, similar to that observedby Tumwegamire et al. (2011). This suggests thatthere is a large amount of sweet potato geneticdiversity in Uganda that needs to be preserved. Veryfew Ugandan landraces had high genetic similaritycoefficients (0.05) with other cultivars, suggestingthat there is high level of genetic diversity beingmaintained by the farmers. This observation wassupported by the AMOVA results, which showed overthat 94 % of the variation was found within popula-tions. Similar results were observed in previousstudies assessing East African germplasm (Gichuruet al. 2006; Tumwegamire et al. 2011). The high levelof genetic diversity within a given gene pool isinfluenced by a number of factors such gene flowthrough intra-specific introgression, and movement ofplant materials between localities during farmer plantmaterial exchange and levels of farmer selection(Gichuru et al. 2006; Veasey et al. 2008). Landraceexchangecan explain ourobservationthat anumberofbFig. 2 A dendrogram of the unweighted pair group methodanalysis (UPGMA) cluster analysis on the basis of Jaccardssimple sequence repeat (SSR) showing genetic similaritiesamong 102 cultivars. BR, GH, KE, PE, RW, TZ, UG stands forBrazil, Ghana, Kenya, Peru, Rwanda, Tanzania, and Uganda,respectively. Other cultivars were obtained from USA. Thedendrogram was pruned from the complete tree to showclustering between genotypes that had bootstrap values greaterthan 60. This pruned tree shows similar broad clustering asobserved for the complete tree (data not shown)Table 5 AMOVA for genetic diversity within and among sweet potato populations from Uganda, other African countries (Ghana,Kenya, Tanzania, Othersa), and American countries (Brazil, Peru, USA).aOthers stand for other African countries including Rwandaand MozambiqueSourcedfSSMSEst. Var.%FSTP valueAmong countries7328.0946.871.95240.2440.001Within countries2521,513.796.036.0376Total2591,841.887.98100Table 6 Pair-wise FSTafor populations from Uganda, other African countries (Ghana, Kenya, Tanzania, Othersb), and Americancountries (Brazil, Peru, USA)cRegionsOthersBrazilGhanaKenyaPeruTanzaniaUgandaUSAOthers0.000Brazil0.1870.000Ghana0.2580.3470.000Kenya0.0060.2940.3140.000Peru0.1810.1270.2210.3080.000Tanzania0.0100.1970.2460.1900.1660.000Uganda0.0000.2850.3870.0680.2990.1000.000USA0.0290.2250.2570.2100.1160.0330.1330.000aFSTis the mean reduction in observed heterozygosity of a subpopulation (relative to the total population) due to genetic drift amongsubpopulations. 0.0 means no differentiation while 1.0 means complete differentiationbOthers stand for cultivars obtained from other African countries including Rwanda and MozambiquecItalicised values indicate significant difference between the populations (P0.001)Genet Resour Crop Evol (2015) 62:501513509123Ugandan landraces had high genetic similarity withcultivars from Kenya, Rwanda and Tanzania.While the majority of genetic diversity was spreadevenly across Uganda and East Africa, there weredistinct regional clusters of genotypes. Pairwise com-parisons of genetic differentiation among regionsindicated that cultivars from Tanzania were signifi-cantly different (P0.001) from all the UgandanpopulationsandtheKenyanone;Ugandasgermplasmwas significantly different (P0.001) from cultivarsfrom Brazil, Peru and Ghana. This suggests that thesame genes were being assorted in different arrays inthe various regions.Even though sweet potato was likely introducedinto Africa from Brazil, we observed substantialdifferentiation between the Brazilian cultivars andall of the African populations. This could be explainedby the low representation of the Brazilian germplasm,which in this study consisted of only 22 genotypes.Another explanation for the higher degree of differ-entiation among populations is that the sweet potatogene pool evolved to local conditions and humanpreferences as it was spread from America throughoutAfrica. This would explain why the majority of theAmerican cultivars were found in one cluster, andmost of the East African germplasm were found inanother cluster. However, some of the cultivars fromTanzania clustered more closely with the Americangroup. This may be because the Tanzanian collectioncontained ancient genotypes that have been cultivatedin isolation on some of Tanzanias islands. TheGhanaian germplasm was more closely related to theAmerican cultivars than with the East African geno-types, although they still formed a unique cluster. Thecultivars from the USA had a higher genetic similaritywith the East African germplasm compared with otherAmerican genotypes, probably because these geno-types were introduced in Uganda about 15 years agoand have been used as paternal parents in thepolycrosses (Mwanga et al. 2003). It is likely thathybridization and introgression has occurred betweenthese genotypes and Ugandan germplasm.The majority of the improved cultivars in Ugandawere found in one cluster probably because of by thesource of parents used in breeding the improvedcultivars. These improved cultivars were generatedfrom a polycross of a single cultivar crossed with1824parentsincludingintroductions fromotherpartsof the world (Mwanga et al. 2003, 2009, 2011). Forexample NASPOT 7-11 were generated from apolycross of 24 parents with Kakamega (a Kenyanlandrace) as the maternal parent. The cross, includedZapallo(from Peru),Beauregardand Jewel (fromUSA), and three introductions from the InternationalInstitute of Tropical Agriculture (IITA) based inNigeria. Hence, it is quite remarkable to note thelevel of genetic diversity present in this relativelysmallandinterbredsampleofimprovedcultivars.ThisBRPEGHTZ KE AFUGUSA1.00.20.0Fig. 3 Probabilities of membership for each individual inclusters K1 and K2 as determined by the Bayesian clusteringmethod implemented in STRUCTURE version 2.3.3 (Pritchardet al. 2000). BR, PE,GH, TZ, KE,AF, and UG standsforBrazil,Peru,Ghana,Tanzania,Kenya,otherAfricancountries(Mozambique and Rwanda) and Uganda, respectively. Othercultivars were obtained from USA. Each individual is repre-sented as a vertical bar and the shading correspond to itsmembership probabilities in clusters K1 (light) and K2 (dark)510Genet Resour Crop Evol (2015) 62:501513123finding is of particular significance for genetic diver-sity conservation because farmers are likely to grad-ually shift from their poor performing cultivars toimproved ones. The three sub-clusters identified,which were composed of only landraces also suggeststhat there is still a large genetic diversity in Ugandathat has not been tapped by breeding programs.ConclusionsOverall, the sweet potato has high levels of geneticdiversity. However, the presence of unique alleles inpopulations from various Ugandas agro-ecologicalzones and other global regions, as well as the regionaldiversity patterns, indicates the value of collecting andcharacterizing the germplasm in more depth. The useofmicrosatellitemarkerdatacanbeparticularlyusefultomakebetterchoicesofwhatneedstobepreservedinorder to increase genetic diversity and representationoflandraces acrossAfrica. These genotypesneed tobeincorporated in the collections at the national genebank and managed to ensure their long-term conser-vation. Finally, the origin of sweet potato germplasmin East Africa doesnt appear to be strictly of a singleBrazilian origin but rather successive introductionfrom several sources.AcknowledgmentsWe are very grateful to the Norman E.Borlaug Leadership Enhancement in Agriculture Program(LEAP) for funding this research. Our sincere gratitude goesto Dr. Joseph Nduguru and Luambano Nessie at MikocheniAgricultural Research Institute, Tanzania for providing us withthe samples from Tanzania. We are also grateful to FrancisOsingada and Jimmy Akono at the Biosciences Facility of theNational Crop Resources Research Institute, Uganda, BramwelWanjala of Biosciences eastern and central Africa (BecA) HubandMaggieMwathiofCIP-OfficeattheInternationalLivestockResearchInstituteinNairobi,Kenya,forthetechnicalassistanceprovided to conduct the research.ReferencesAndersson MS, deVicenteMC (2010)Gene flowbetween cropsand their wild relatives. Johns Hopkins University Press,Baltimore, p 564Austin DF (1977) Hybrid polyploids in Ipomoea section ofBatatas. J Hered 68:259260Austin DF (1983) Variability in sweet potato in America. Proc.Amer. Soc. Hort. Sci. 27(pt. B):1526Buteler MI, Jarret RL, La Bonte DR (1999) Sequence charac-terization of microsatellites in diploid and polyploidIpomoea. Theor Appl Genet 99:123132. doi:10.1007/s001220051216Connolly AG, Godwin ID, Cooper M, Delacy IH (1994) Inter-pretation of randomamplified polymorphic DNA markerdata for fingerprinting sweet potato Ipomoea batatas (L.)Lam. genotypes. Theor Appl Genet 88:332336. doi:10.1007/BF00223641DoyleJJ,DoyleJL(1990)IsolationofplantDNAfromfreshtissue.Focus 12:1315. http:/ci.nii.ac.jp/naid/10003365693/Earl DA, von Holdt MB (2012) STRUCTURE HARVESTER: awebsite and program for visualizing STRUCTURE outputand implementing the Evanno method. Conserv GenetResour 4(2):359361. doi:10.1007/s12686-011-9548-7Elameen A, Fjellheim S, Larsen A, Rognli AO, Sundheim L,Msolla S, Masumba E, Mtunda K, Klemsdal SS (2008)Analysis of genetic diversity in a sweet potato (Ipomoeabatatas L.) germplasm collection from Tanzania asrevealed by AFLP. Genet Resour Crop Evol 55:397408.doi:10.1007/s10722-007-9247-0EsselinkGD,NybomH,VosmanB(2004)Assignmentofallelicconfiguration in polyploids using the MAC-PR (microsat-ellite DNA allele counting-peak ratios) method. TheorAppl Genet 109:402408. doi:10.1007/s00122-004-1645-5Evanno G, Regnaut S, Goudet J (2005) Detecting the number ofclusters of individuals using the software STRUCTURE: asimulation study. Mol Ecol 14:26112620. doi:10.1111/j.1365-294X.2005.02553.xFAO(2008)Thestateoffoodinsecurityintheworld2003.Foodand Agriculture Organization, Rome. /3/a-i0291e.pdf. Accessed 16 May 2012FAO (2012) FAO crop production statistics. Food and Agri-culture Organization. /. Accessed 03March 2013Gao M, Ashu GM, Lakeisha S, Akwe WA, Njiti V, Barnes S(2011) Wx intron variations support an allohexaploid ori-gin of the sweet potato Ipomoea batatas (L.) Lam. Eu-phytica 177:111133. doi:10.1007/s10681-010-0275-zGarc a-Verdugo C, Fay MF, Granado-Yela C, Rubio De CasasR, Balaguer L, Besnard G, Vargas P (2009) Geneticdiversity and differentiation processes in the ploidy seriesof Olea europaea L.: a multiscale approach from subspe-cies to insular populations. Mol Ecol 18:454467. doi:10.1111/j.1365-294X.2008.04027.xGichuki ST, Berenyi M, Zhang D, Herman M, Schmidt J, GlosslJ, Burgh K (2003) Genetic diversity in sweet potatoIpomoea batatas (L.) Lam. in relationship to geographicsources as assessed with RAPD markers. Genet ResourCrop Evol 50:429437. doi:10.1023/A:1023998522845Gichuru V, Aritua V, Lubega GW, Edema R, Adipala E, Ru-baihayo PR (2006) A preliminary analysis of diversityamong East African sweet potato landraces using mor-phological and simple sequence repeats (SSR) markers.Acta Hort (ISHS) 703:159164. /books/703/703_18.htmHe G, Prakash CS, Jarret RL (1995) Analysis of genetic diver-sity in a sweet potato (Ipomoea batatas) germplasm col-lection using DNA amplification fingerprinting. Genome38(5):938945. doi:10.1139/g95-123He X, Liu Q, Ishiki K, Zhai H, Wang Y (2006) Genetic diversityand genetic relationships among Chinese sweet potatoGenet Resour Crop Evol (2015) 62:501513511123landracesrevealedbyRAPDandAFLPmarkers.BreedSci56:201207. doi:10.1270/jsbbs.56.201Holsinger KE, Bruce SW (2009) Genetics in geographicallystructured populations: defining, estimating and interpret-ing FST. Nat Rev Genet 10(9):639650. doi:10.1038/nrg2611HuJ,NakataniM,LalusinAG,KuranouchT,FujimuraT(2003)Geneticanalysisofsweetpotatoanditswildrelativesusinginter-simplesequencerepeat(ISSRs).BreedSci53:297304. doi:10.1270/jsbbs.53.297Hu J, Nakatani M, Mizuno K, Fujimura T (2004) Developmentand characterization of microsatellite markers in sweetpotato. Breed Sci 54:177188. doi:10.1270/jsbbs.54.177Hwang SY, Tseng YT, Lo HF (2002) Application of simplesequence repeats in determining the genetic relationshipsof cultivars used in sweet potato polycross breeding inTaiwan.SciHort93:215224.doi:10.1016/S0304-4238(01)00343-0JaccardP(1908)Nouvellesrecherchessurladistributionflorale.Bull Soc Vaud Sci Nat 44:223270 (In French)Jakobsson M, Rosenberg NA (2007) CLUMPP: a clustermatching and permutation program for dealing with labelswitching and multimodality in analysis of populationstructure. Bioinformatics 23(14):18011806. doi:10.1093/bioinformatics/btm233Jarret RL, Bowen N (1994) Simple sequence repeats (SSRs) forsweet potato characterization. Plant Genet Resour Newsl100:911Jrgensen MH, Carlsen T, Skrede I, Elven R (2008) Microsat-ellites resolve the taxonomy of the polyploidy Cardaminedigitata aggregate (Brassicaceae). Taxon 57:882892./stable/27756715Karuri HW, Ateka EM, Amata R, Nyende AB, Muigai AWT,Mwasame E, Gichuki ST (2010) Evaluating diversityamong Kenyan sweet potato genotypes using morpholog-ical and SSR markers. Int J Agric Biol. 12:3338. http:/published_papers/31561_.pdfKloda JM, Dean PDG, Maddren C, MacDonald DW, Mayes S(2008) Using principal component analysis to comparegenetic diversity across polyploidy levels within plantcomplexes: an example from British Restharrows (Ononisspinosa and Ononis repens). J Hered 100:253260. doi:10.1038/sj.hdy.6801044Kobayashi M (1983) The Ipomoea trifida complex closelyrelated to sweet potato. In: Shideler SF, Rincon H (eds.)Proceedings of the 6th symposium of the InternationalSociety of Tropical Root Crops, Lima, Peru, 2126 Feb-ruary, 1983, International Potato Center, Lima, Peru,pp 561568Lebot V (2010) Sweet potato, chapter 3. In: Bradshaw JE (ed)Root and tuber crops. Handbook of Plant Breeding 7.Springer Science & Business Media, Berlin, pp 97125Mignouna HD, Abang MM, Fagbemi SA (2003) A comparativeassessment of molecular marker assays (AFLP, RAPD andSSR) for white yam (Dioscorea rotundata Poir.) germ-plasm characterization. Ann Appl Biol 142:269276.doi:10.1111/j.1744-7348.2003.tb00250.xMuyinza H, Talwana HL, Mwanga ROM, Stevenson PC (2012)Sweet potato weevil (Cylas spp.) resistance in Africansweet potato germplasm. Int J Pest Manag 58(1):7881.doi:10.1080/09670874.2012.655701Mwanga ROM, Odongo B, Ocitti pobwoya C, Gibson RW,Smith N, Carey E (2003) Release of six (NASPOT 1 toNASPOT 6) sweet potato cultivars in Uganda. Hort-Science 38:475476. /content/38/3/475.full.pdfMwanga ROM, Odongo B, Niringiye CN, Alajo A, Kigozi B,Makumbi R et al (2009) NASPOT 7, NASPOT 8,NASPOT 9 O, NASPOT 10 O, and Dimbuka-Buk-ulula sweet potato. HortScience 44:828832. http:/content/44/3/828.full.pdfMwanga ROM, Niringiye CN, Alajo A, Kigozi B, Namakula J,Mpembe I et al (2011) NAPOT 11, a sweet potato cul-tivar bred by a participatory plant-breeding approach inUganda.HortScience46:317321.http:/hortsci./content/46/2/317.fullNei M (1973) Analysis of gene diversity in subdivided popu-lations. Proc Natl Acad Sci USA 70:33213323Nishiyama I (1971) The origin of the sweet potato plant. In:Barrau J (ed) Plants and the migrations of Pacific peoples.Press Honolulu, Bishop Mus, pp 119128Otoo E, Akromah R, Kolesnikova-Allen M, Asiedu R (2009)Delineation of pona complex of yam in Ghana using SSRmarkers. Int J Genet Mol Biol 1:616 (Article number77EE4812732)Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis inExcel. Population genetic software for teaching andresearch. Mol Ecol Notes 6:288295. doi:10.1111/j.1471-8286.2005.01155.xPerrier X, Jacquemoud-Collet JP (2006) DARwin software.http:/darwin