Annatto is grown for its unique capability to produce bixin, a natural pigment found predominantly in the
outer coatings of its seeds and widely used in the food industry (Bittencourt et al., 2005). The demand for natural colorants of annatto has increased trade worldwide. Besides its natural nutrients, it helps lower glucose (Russell et al., 2005) and cholesterol (McGonigle and McCracken, 2002) in the blood. It has also displayed hypolipidemic (Lima et al., 2001) and anti-microbial activities (Silveira et al., 2003). All these properties make annatto a very attractive and safe substitute pigment for many synthetic colorants, including as an alternative tracking dye for gel electrophoresis (Siva et al., 2008).
However, annatto dye and grain yields vary from sample to sample and from plant to plant (Michelangeli et al., 2002; Valdez-Ojeda et al., 2008). With this in mind, several methodologies have been developed to accelerate true-breeding programs; that is, high yielding lines have been selected and vegetatively propagated for commercial plantations (Parimalan et al., 2007). In vitro regenera-tion of B. orellana has been used for the production of high-yield-ing genotypes (Madhuri and D’Souza, 2000). Portela de Carvalho
et al. (2005) induced polyploidy in vitro to duplicate the chromo-some complements in interspecific hybrid crosses. Morphologic and genetic assessments have been used for parental selection (Rivera-Madrid et al., 2006; Valdez-Ojeda et al., 2008).
Critical decisions in breeding programs include the choice of mating, the intensity of selection applied and proper manage-ment of inbreeding (Sleper and Poehlman, 2006). All of these require the consideration of the mating system to allow the breeder to select the most effective breeding scheme.
Before this study we generated information concerning the mating system in B. orellana through artificial pollination (Rivera-Madrid et al., 2006). We found 57% outcrossing and 31.4% selfing (Rivera-Madrid et al., 2006). To support the existence of both reproduction systems in B. orellana, and because SRAP mark-ers (Li and Quiros, 2001) have been successfully used in Cyperus difformis (Merotto et al., 2009); they have been used here to obtain quantitative estimates of mating system parameters and to make appropriate estimates of outcrossing rates.
The SRAP markers combine simplicity, reliability, and a moderate throughput ratio and target coding sequences in the genome, without previous knowledge of the genome to which they apply (Li and Quiros, 2001). The SRAP is adequate for outcrossing estimation in annatto because it has been found to detect adequate levels of polymorphism in this species (Valdez-Ojeda et al., 2008). The dominant character of SRAP may be disadvantageous, but simulation studies have demonstrated that by multilocus estimation of outcrossing, the variances decrease when more loci are included in the estimation with intermediate gene frequencies, making this limitation readily overcome (Ritland and Jain, 1981). The multilocus estimation of outcrossing with dominant markers procedure (MLDT)
ABSTRACTSeveral methodologies have been proposed to accelerate breeding programs in annatto (Bixa orellana L.). However, before imple-menting them it is important to understand the mating system in annatto, a tropical species that is highly commercial for its unique capability to produce bixin. This report focuses on estimating outcrossing rates in an open-pollinated population of B. orellana by sequence-related amplified polymorphism (SRAP). Eleven open-pollinated progeny arrays of 10 individuals each were used for this determination. Sequence-related amplified polymorphism data indicated a multilocus outcrossing rate (tm) of 0.748 and an average single-locus outcrossing rate (ts) of 0.651. Outcrossing was predominant in the B. orelllana open-pollinated population. The mul-tilocus estimates differed from the single-locus estimates (0.097), suggesting the existence of mating among relatives (biparental inbreeding). A Wright’s fixation index (F) of 0.332, indicated high homozygosity in the progeny evaluated with random mating, suggesting low variability. Furthermore, we found that based on experimental analysis, a minimum of 40 markers was required to obtain tm estimates that stabilized at approximately 0.7, and did not change significantly with an increasing number of markers. The molecular markers SRAP detect adequate levels of polymorphism in B. orellana for evaluating the mating system of this species.
R. Valdez-Ojeda, M. de Lourdes Aguilar-Espinosa, and R. Rivera-Madrid, Centro de Investigación Científica de Yucatán, Calle 43 No. 130, Col. Chuburná de Hidalgo, 97200 Mérida, Yucatán, México; C.F. Quiros, Department of Vegetable Crops, University of California, Davis, CA 95616. Received 15 Dec. 2009. *Corresponding author ([email protected]).
Abbreviations: F, Wright’s fixation index; MLDT, multilocus estimation of outcrossing with dominant markers procedure; OP, open pollinated; PCR, polymerase chain reaction; SRAP, sequence-related amplified polymorphism; tm, multilocus outcrossing rates; ts, average single-locus outcrossing rate.
requires grouping the population samples in family arrays and using neutral dominant loci that segregate independently. The main aim of this study was to estimate the outcrossing rates in a population of B. orellana in Yucatán, México by SRAP.
MATERIALS AND METHODSPlant Materials
Bixa orellana populations bloom from September to Novem-ber in the fields of Yucatán, México. The plant material was collected from the agronomic region of Yucatán located at lati-tude 20º13´ to 20º29´ and longitude 88º55́ to 89º54́ named
Yaxcaba. The selected trees were sufficiently distant from each other to allow the assumption of nonrelatedness among the individuals. Eleven trees subjected to open-pollination (OP) were randomly selected for progeny testing in 2005 in a total population of 100 individuals. Among the sampled trees, to compare the mating system of different floral phenotypes, we selected four plants that exhibited white flowers and seven with pink flowers (Fig. 1). From each tree, seeds were randomly collected from different fruits, disinfected, and grown under greenhouse conditions with soil and agrolite (2:1) at the Center of Scientific Investigation of Yucatán (CICY). The analysis was performed on 10 seeds from each of the 11 OP trees. A total of 110 individuals were analyzed.
DNA Isolation and Sequence-Related Amplified Polymorphism Procedure
From each 1-mo-old seedling, two to four leaves were ground in liquid N and total genomic DNA was extracted using a DNeasy Tissue Kit (Qiagen, Hilden, Germany). Following the method of Li and Quiros (2001), the SRAP protocol was implemented using primer combinations selected based on their polymorphism in B. orellana (Valdez-Ojeda et al., 2008), as well as others selected for this study. Nineteen primer combina-tions were used to detect 50 polymorphic bands, which were used for this study (Table 1). The SRAP analysis was performed with forward primers labeled with IRDye 800 or IRDye 700 (LI-COR, Lincoln, NE). A standard PCR cocktail mix was prepared according to Valdez-Ojeda et al. (2008) and subjected to the following conditions: 1 cycle of 94ºC for 2 min; 5 cycles of 94ºC for 30 s, 35ºC for 30 s, and 72ºC for 1 min; 35 cycles of 94ºC for 30s, 50ºC for 30 s, and 72ºC for 1 min; followed by a final extension of 72ºC for 2 min. Three microliters of blue stop solution were added to the PCR products before denaturing at 94ºC for 5 min. After denaturing, 0.6 μL of the reaction mixture was loaded into KB 6.5% gel matrix (Li-Cor, Biosciences) and
Fig. 1. Maternal trees showing different traits in flowers and fruits. White flowers are indicated by numbers 16, 19, 58, and 87. Pink flowers are indicated by numbers 54, 55, 56, 67, 68, 77, and 79.
Table 1. Primer combinations and sequences of the sequence-related amplified polymorphism (SRAP) primers used in B. orellana in this study.
continuous 1x TBE buffer. The samples were run in triplicate to ensure reproducibility. The PCR products were run in the sequencer IR2 4200, LI-Cor, Biosciences, (Lincoln, NE).
Data Analysis
The bands were scored considering two possible alleles: band presence or absence. The mating system was analyzed using the software: multilocus estimation of outcrossing with dominant markers procedure (MLDT) using the mathematical formu-lae provided by Ritland and Jain, (1981), Ritland (1990), and Ritland (2002). The MLDT specifies that both selfing and out-crossing occur in a population (Shaw and Allard, 1982). From progeny array data, the MLDT simultaneously estimated the tm by the Newton-Raphson method (which includes biparental inbreeding) and the average ts (which excludes most biparental inbreeding). Thus, the difference between these estimates was considered to be the rate of biparental inbreeding. The average single-locus inbreeding coefficient (also known as Wright’s fixa-tion index) of maternal parents (F) via the Newton-Raphson method and the average gene frequencies via the expectation-maximization method were also estimated (Ritland and Jain, 1981; Ritland, 2002). One thousand bootstraps of among the families sampled were used to estimate the standard error for all outcrossing rates (including tm – ts) and allele frequencies. A χ2 statistic was calculated for each locus to determine the goodness of fit of the data to the model. It tests the differences between the observed and expected number of offspring of each genotype for each maternal genotype under the mixed mating model, estimating the proportion of the progeny result-ing from self-fertilization and that from outcrossing.
We also empirically estimated the minimum number of domi-nant markers necessary to obtain a consistent outcrossing rate estimate. In accordance with Gaiotto et al. (1997), we initially estimated tm with four randomly selected markers. Subsequently tm was reestimated by increasing the number of markers used by four until the total available number of markers was analyzed. Each tm estimated and its respective standard error were calcu-lated based on 100 bootstrap data sets by MLDT software using the mathematical formulae provided by Ritland and Jain (1981), and Ritland (2002). The estimated tm obtained from each four markers and its respective standard error was plotted.
RESULTS AND DISCUSSIONThe SRAP markers were registered by band presence or
absence (Fig. 2). The average number of amplicons per primer pair was three, ranging from one to five per primer set, con-firming that SRAP marker polymorphism in B. orellana was adequate for evaluating the mating system of this species.
The multilocus outcrossing rates based on all 50 loci were high (0.748 ± 0.092) and significantly different from 1.0, as determined by the standard error. This indicates that B. orel-lana is predominantly an outcrossing species (Table 2). The estimates of multilocus outcrossing rates (tm) and the average single-locus outcrossing rates (ts) obtained from the MLDT of SRAP data clearly show that some selfing also occurs (26%).
Bixa orellana seeds in Yucatán are the product of a mixed mat-ing system. In this context, previous artificial pollination studies developed in experimental fields in Yucatán have suggested that annatto can tolerate both pollination types, as indicated by the
fruit set of 31.4% resulting from self-pollination and 57% from allogamy (Rivera-Madrid et al., 2006). From artificial pollina-tion studies, flower maturation in annatto has been shown to be asynchronous on the same panicle and between different panicles. Annatto has also been proved to be protandrous, that is, the anthers shed their pollen before the stigma of the same flower is receptive (Rivera-Madrid et al., 2006), therefore pro-moting outcrossing. This reproductive mechanism (documented in 50% of all angiosperms) is a prezygotic mechanism that favors outcrossing, reducing the frequency of selfing and biparental inbreeding (Charlesworth and Charlesworth, 1987). The char-acteristics found in annatto are also exhibited in the medicinal plant Myracrodruon urundeuva (Anacardiaceae). Estimations of the multilocus outcrossing rates of this species have revealed that it possesses a mating system with a high proportion of outcross-ing due to two events: flowering asynchrony and the behavior of different visiting pollinators (Freitas et al., 2004). On the other hand, the prezygotic mechanism in tarbush, Flourensia cernua, apparently favors a high rate of outcrossing in this species and a variable degree of self-compatibility that in some cases can also be very high (Ferrer et al., 2004).
Bixa orellana is normally considered by its variability of contrasting morphological traits to be a cross-pollinated crop (Portela de Carvalho et al., 2005). However, in crops that are normally classified as cross pollinated, some self pollination usually occurs as a result of numerous factors such as inflores-cence morphology (Sleper and Poehlman, 2006). The repro-ductive biology of annatto has not been explained by floral morphology, although floral morphology in annatto (Rivera and Flores, 1988), indicated by the position of female and male sexual organs, appears to promote selfing.
The difference between the multilocus rate and the average single-locus estimates (0.097 ± 0.060) were not significantly different from zero, suggesting the existence of mating among relatives (biparental inbreeding) (Table 2). This biparental inbreeding, or mating between relatives, apparently causes selfing or increased homozygosity (Ritland, 2002), which is indicated by the single locus rate in this study.
Artificial pollination studies have indicated seed set per fruit to be higher in cross pollination of the white flower variant with the pink and/or purple flower variants, than in crosses involving only white flower variants (Rivera-Madrid et al., 2006). This suggests inbreeding depression with respect to white flower variants because they are probably related by ascendance. In our study, we used four maternal plants with white flowers and seven with pink flowers; therefore it was possible to test whether different outcrossing rates applied to these morphological types. To test this hypothesis we analyzed white and pink variants separately. We found high outcrossing rates tm (0.856 ± 0.172) for the white and pink flower variants tm (0.825 ± 0.088), with no significant differences between them. However, of particular interest was the Wrights’ fixation index F: for the white flower variants we observed high values of F (0.395), demonstrating an excess of homozygotes; however, with the pink flower variants this fixation index was –0.006, indicating an excess of heterozy-gotes. Therefore we can effectively assume that the white flower variants were intermating, contrary to the pink flower variants. On the other hand, the difference in outcrossing rates tended to
Agronomy Journa l • Volume102, Issue4 • 2010 1343
Table 2. Allele frequencies (Freq), standard deviation (σ), and χ2 statistic according to the mixed crossing model for domi-nant markers (MLDT, Ritland and Jain, 1981; Ritland, 2002) in Bixa orellana.
Fig. 2. Sequence-related amplified polymorphism (SRAP) fragments generated with primer combinations SA–12 and EM–8. The families 1, 2, and 3 are indicated. The arrows indicate the markers detected. Size standard (in base pairs) is shown on the left.
1344 Agronomy Journa l • Volume102, Issue4 • 2010
be higher in the white flower variants (0.180 ± 0.154) than in the pink flower variants (0.074 ± 0.035).
In the progeny evaluated with random mating, F (0.332), indi-cated high homozygosity (Table 2). The likelihood of inbreeding was investigated by comparing the estimated and expected F values. Wright’s fixation index is an equilibrium statistic and it provides an estimate of what has happened historically in the population (e.g., F values close to zero suggest a history of outcrossing populations). The outcrossing rate t is an estimate of what has happened during the most recent generation and is less affected by factors such as selection and drift than F is. Therefore, F and t values may seem to suggest different levels of outcross-ing in a population, but they may not actually be contradictory (Cook and Soltis, 1999). Wright’s fixation value in the progeny screened with SRAP was lower than expected, based on the estimate of tm. With a tm value of 0.748, the expected fixation index was [F = (1 – t)/(1 + t)] = 0.144, while the estimated value was 0.332. A higher than expected index suggests an excess of homozygotes and greater inbreeding than expected in the prog-enies of B. orellana analyzed. Although the maternal plants were randomly collected, the excess of homozygotes indicated by the fixation index explain the 26% selfing observed in B. orellana.
The χ2 statistic to test the conformity of marker loci with the mixed mating model indicated that for 29 SRAP mark-ers the number of observed progeny individuals for each genotype class from each maternal genotype class departed from the expected numbers (Table 3). According to Ritland (1983) there are several factors that influence such deviations, including selection against homozygous genotypes, a genotype-dependent outcrossing rate, and unbalanced frequencies of pollen in the population. Results similar to ours were obtained by estimation of outcrossing rates in Moringa oleifera, a species adapted to a mixed mating system, which using AFLP markers showed a multilocus outcrossing rate (tm) of 0.74 (Muluvi et al., 2004). Also, as in annatto this species is self-compatible.
On the other hand, according to our empirical estimation, 40 markers were required as a minimum to estimate tm = 0.7, with minimum standard error (Fig. 3). Furthermore we observed that by increasing the dominant markers it was possible to obtain a robust estimate of tm (Ritland, 1990). Gaiotto et al. (1997) sug-gested, in their empirical estimation with the dominant markers RAPD and AFLP in Eucalyptus urophylla, that with 18 to 20 dominant markers the estimates of tm were consistent. Merotto et al. (2009) analyzed 79 loci to estimate the mating system of Cyperus difformis with SRAP markers and the minimum num-ber of loci required for empirical estimation of tm was 48.
In summary, we found that the mating system of annatto is pre-dominantly outcrossing. We also found differences in F between different flower color variants. High F values demonstrated an excess of homozygotes in the white flower types. This is probably explained by few pollinators visiting white flower phenotypes, as a result of a reduced preference for this flower color. This phenom-enon has already been reported in other plants such as Ipomoea purpurea (Epperson and Clegg, 1987; Clegg and Durbin, 2000).
The annatto variants analyzed in this study present agro-nomical characteristics desirable for the genetic improvement of this crop, favoring the attainment of high contents of bixin. In the case of the white variants, these are characterized by low bixin contents and indehiscent fruit, a characteristic that may
enhance crop yields by protecting the seeds and pigment quality; while the pink variant presents high bixin contents but dehis-cent fruits that expose the seeds to sunlight, leading to bixin photooxidation and decreased bixin content (Valdez-Ojeda et al., 2008). Although the white variant presented high F values, in previous artificial pollination studies, this same variant also presented greater receptability to the pollen of the pink flower variant and a higher proportion of mature fruits (Rivera-Madrid et al., 2006). This implies that an important advance has been made toward more effective genetic improvement of annatto, since crosses between the different variants have the potential to produce desired traits in hybrid offspring: indehiscent pods and high bixin contents. However, this investigation also strongly indicates a need for in-depth studies concerning the nature of pollination (floral structure and its relation to the mating system; and the contribution of pollinating agents and seed dispersal) since evidence suggests that the white variant can be both self and cross pollinated and that pollinators may play an important role in these processes. Knowledge of all of the above aspects, in conjunction with the results of this study, need to be taken into account in breeding program to improve the annatto crop.
ACKNOWLEDGMENTS
This work was supported by the International Foundation for Science (IFS) F/2932-3 and the Consejo Nacional de Ciencia y Tecnología (CONACYT) 46541 and UC Mexus. V-O R was sup-ported by CONACYT PhD grant no. 185874.
Fig. 3. Changes in mean ± standard error of estimates of the multilocus outcrossing rate (tm), using multilocus estimation of outcrossing with dominant markers procedure (MLDT) software. The first point on the graph represents tm calculated with four randomly selected markers. For each subsequent point tm was re-estimated by increasing the number of markers used by four until the total available number of markers was analyzed.
Table 3. Outcrossing rate (tm), average single-locus outcross-ing rate (ts), difference between these estimates (tm – ts), and Wright’s fixation index (F) with the respective standard error, calculated by the multilocus estimation of outcrossing with dominant markers procedure (MLDT) software with families of Bixa orellana.
Site Families Progeny tm ts tm – ts F
Yaxcaba 11 110 0.748 0.651 0.097 0.332
Standarderror (0.092) (0.060) (0.060) (0.000)
Agronomy Journa l • Volume102, Issue4 • 2010 1345
REFERENCESBittencourt, C., M.P. Felicissimo, J.-J. Pireaux, and L. Houssiau. 2005. Study of
annatto from Bixa orellana seeds: An application of time-of-flight secondary ion mass spectrometry. Spectros. Eur. 17:16–22.
Charlesworth, D., and B. Charlesworth. 1987. Inbreeding depression and its evolu-tionary consequences. Annu. Rev. Ecol. Syst. 18:237–268.
Clegg, M.T., and M.L. Durbin. 2000. Flower color variation: A model for the experimental study of evolution. Proc. Natl. Acad. Sci. USA 97:7016–7023.
Cook, L.M., and P.S. Soltis. 1999. Mating systems of diploid and allotetraploid populations of Tragopogon (Asteraceae). II. Artificial populations. Heredity 82:237–244.
Epperson, B.K., and M.T. Clegg. 1987. Frequency-dependent variation for out-crossing rate among flower-color morphs of Ipomoea purpurea. Evolution Int. J. Org. Evolution 41:1302–1311.
Ferrer, M.M., L.E. Eguiarte, and C. Montaña. 2004. Genetic structure and out-crossing rates in Flourensia cernua (Asteraceae) growing at different densities in the South-western Chihuahuan Desert. Ann. Bot. (London) 94:419–426.
Freitas, M.L.M., A.M. Sebbenn, M.L.T. Moraes, and E.G.M. Lemos. 2004. Mat-ing system of a population of Myracrodruon urundeuva F.F. & M.F. Allemão using the fAFLP molecular marker. Genet. Mol. Biol. 27:425–431.
Gaiotto, F.A., M. Bramucci, and D. Grattapaglia. 1997. Estimation of outcrossing rate in a breeding population of Eucalyptus urophylla with dominant RAPD and AFLP markers. Theor. Appl. Genet. 95:842–849.
Li, G., and C.F. Quiros. 2001. Sequence-related amplified polymorphism (SRAP), a new marker system based on a simple PCR reaction: Its application to map-ping and gene tagging in Brassica. Theor. Appl. Genet. 103:455–461.
Lima, L.R.P., T.T. De Oliveira, T.J. Nagem, A.D.S. Pinto, P.C. Stringheta, A.L.A. Tinoco, and J.F. Da Silva. 2001. Bixina, norbixina e quercetina e seus efeitos no metabolismo lipídico de coelhos. Braz. J. Vet. Res. Anim. Sci. 38:196–200. (In Portugeuse.)
Madhuri, S., and M.C. D’Souza. 2000. In vitro clonal propagation of annatto (Bixa orellana L.). Curr. Sci. 78:1532–1535.
McGonigle, K.E.M., and K.J. McCracken. 2002. Study on the effects of annatto (Bixa orellana) and Yucca Schidegera on cholesterol content and yolk colour. Br. Poult. Sci. 43:S41–S42.
Merotto, A., Jr., M. Jasieniuk, and A.J. Fischer. 2009. Estimating the outcrossing rate of Cyperus difformis using resistance to ALS-inhibiting herbicides and molecular markers. Weed Res. 49:29–36.
Michelangeli, C., A.M. Medina, P. Artioli, and J. Mata. 2002. Microsporogénesis y microgametogénesis de onoto (Bixa orellana L.). A. C. V. 53:171–175. (In Portuguese.)
Muluvi, G.M., J.I. Sprent, D. Odee, and W. Powell. 2004. Estimates of outcrossing rates in Moringa oleifera using Amplified Fragment Length Polimorphism (AFLP). Afr. J. Biotechnol. 3:146–151.
Parimalan, R., P. Giridhar, B.G. Harischandra, and G.A. Ravishankar. 2007. Organogenesis from cotyledon and hypocotyl-derived explants of japhara (Bixa orellana L.). A. B. C. 66:153–160.
Portela de Carvalho, J.F.R., C.R. De Carvalho, and O.W. Campos. 2005. In vitro induction of polyploidy in annatto (Bixa orellana). Plant Cell Tissue and Organ Culture 80:69–75.
Ritland, K. 1983. Estimation of mating systems. p. 289–302. In S.D. Tanksley and T.J. Orton (ed.) Isozymes in plant genetics and breeding, part A. Elsevier Sci-ence Publ., Amsterdam.
Ritland, K. 1990. A series of FORTRAN computer programs for estimating plant mating systems. Heredity 47:37–54.
Ritland, K. 2002. Extensions of models for the estimation of mating systems using n independent loci. Heredity 88:221–228.
Ritland, K., and S. Jain. 1981. A model for the estimation of outcrossing rate and gene frequencies using n independent loci. Heredity 47:35–52.
Rivera, D.I., and E.M. Flores. 1988. Morfología floral del achiote, Bixa orellana L (Bixaceae). Rev. Biol. Trop. 36:499–509.
Rivera-Madrid, R., R.M. Escobedo-GM, E. Balam-Galera, M. Vera-Ku, and H. Harries. 2006. Preliminary studies toward genetic improvement of annatto (Bixa orellana L.). Sci. Hortic. (Amsterdam) 109:165–172.
Russell, K.R.M., E.Y.S.A. Morrison, and D. Ragoobirsingh. 2005. The effect of annatto on insulin binding properties in the dog. Phytother. Res. 19:433–436.
Shaw, D.V., and R.W. Allard. 1982. Estimation of outcrossing rates in Douglas-fir using isozyme markers. Theor. Appl. Genet. 62:113–120.
Silveira, P.C.A.M., G.A. Da Silva, O.M. Carneiro, and J.K.Ch. Vieira. 2003. Ativi-dade antimicrobiana de Bixa orellana L. (Urucum). Revista Lecta. Bragança Paulista 21:47–54.(In Portuguese.)
Siva, R., G.J. Mathew, A. Venkat, and C. Dhawan. 2008. An alternative tracking dye for gel electrophoresis. Curr. Sci. 94:765–767.
Sleper, D.A., and J.M. Poehlman. 2006. Methods in plant breeding. p. 155–182. In D.A. Sleper and J.M. Poehlman (ed.) Breeding field crops. Blackwell Publ., Ames, IA.
Valdez-Ojeda, R., J.L. Hernández-Stefanoni, M. Aguilar-Espinosa, R. Rivera-Madrid, R. Ortiz, and C.F. Quiros. 2008. Assessing morphological and genetic variation in annatto (Bixa orellana L.) by sequence-related amplified polymorphism and cluster analysis. HortScience 43:2013–2017.
![[2010] «Adaptación al sistema móvil de los lienzos del retablo mayor de la iglesia de los santos Justo y Pastor de Granada». Revista ph, 76, Noviembre de 2010, pp. 68-77. ISSN:](https://static.documents.page/doc/80x56/6361a4a62fd302ae640953be/2010-adaptacion-al-sistema-movil-de-los-lienzos-del-retablo-mayor-de-la-iglesia.jpg)
