Printer Friendly

Genetic variation among tomato accessions from primary and secondary centers of diversity.

Vavilov (1926) divided regions of modern cultivation into primary centers, regions where domestication occurred, and secondary centers, regions away from the primary center where post-domestication events occurred. Secondary centers are expected to possess less variation than primary centers if the genetic diversity associated with adaptation to a particular ecogeographic region is proportional to the length of time a species inhabits that region (Willis, 1922).

The primary center of diversity for cultivated tomato, Lycopersicon esculentum Mill., has been located in a narrow belt along the South American west coast, limited by the equator and 30 [degrees] S and the Andes and the Galapagos Islands (Zeven and de Wet, 1982). This Andean region and Mexico hold the greatest morphological variability in tomato (Rick, 1958; Jenkins, 1948). The most complete taxonomic treatments of Lycopersicon (Muller, 1940; Luckwill, 1943) credit L. esculentum var. cerasiforme as the putative ancestor of cultivated L. esculentum. A commonly accepted hypothesis for the domestication of cultivated tomato is that the sub-species, cerasiforme, originated in the Andean region, spread to Mexico as a weed, became domesticated in Mexico, and then was disseminated to the Old World (Jenkins, 1948; Rick, 1958). The concordance of allozymes between the two taxa (Rick and Fobes, 1975) and the greater genetic variability in cerasiforme from the Andean region than found in accessions collected outside of this region (Rick and Fobes, 1975; Rick and Holle, 1990) support this hypothesis. However, Rick and Fobes (1975) and Rick and Holle (1990) recognize a second hypothesis that independent centers of domestication may exist in Mexico and the Andean region. Presently, cerasiforme grows in most tropical regions of the world (Rick, 1973).

In the sixteenth century, the tomato was taken to Europe and later disseminated to many areas of the world (McCue, 1952). Variation in L. esculentum has been reduced through domestication because of self-pollination, founder effects, and both artificial and natural selection (Rick, 1958; Rick and Fobes, 1975). Morphological and molecular variation is greater among genotypes from South America, Central America, and Mexico than among U.S. and European cultivars (Jenkins, 1948; Rick, 1958; Rick and Fobes, 1975; Miller and Tanksley, 1990; Williams and St. Clair, 1993). Cerasiforme germplasm has been found to be both more and less variable than L. esculentum for morphology (Rick, 1958), isozymes (Rick and Fobes, 1975), and molecular markers (Miller and Tanksley, 1990; Williams and St. Clair, 1993) depending on the germplasm studied.

The Asian Vegetable Research and Development Center (AVRDC), Shanhua, Tainan, Taiwan, maintains one of the world's largest collections of Lycopersicon germplasm. This collection presently contains 6074 accessions (Tay, 1989), with 722 accessions from Ecuador, Peru, and Chile and 225 accessions from Mexico, the putative centers of diversity and domestication of L. esculentum (Zeven and de Wet, 1982; Jenkins, 1948; Rick and Fobes, 1975). In addition, 5127 accessions are from outside of these regions including Africa, Asia, Canada, the Caribbean, Europe, and the USA along with other South and Central American countries. Approximately 60% of the collection is from secondary centers, and it is unclear whether these accessions represent unique sources of genetic variation compared with accessions received from the primary center or geographic regions contiguous with the primary center. At least 31 cerasiforme accessions have been identified through routine characterization and seed increase. However, there has been no specific attempt at AVRDC to identify cerasiforme accessions throughout the entire collection. Pedigree information is usually not available for genebank accessions, rendering the relationships among accessions unclear. Relationship information is necessary to allocate resources to different parts of the collection and to guide future users in their sampling strategy. Molecular marker variation can be used to quantify the genetic structure and variation within and among populations. Williams and St. Clair (1993) have demonstrated the usefulness of RAPDs (Welsh and McClelland, 1990; Williams et al., 1990) for germplasm organization in Lycopersicon species. Although Rick and Fobes (1975), Miller and Tanksley (1990), and Williams and St. Clair (1993) have studied genetic variation patterns in tomato, these studies primarily compared New World landraces with U.S. and European cultivars. The comprehensive L. esculentum collection held at AVRDC provides an opportunity to sample a wider range of germplasm adapted to many environments. The first objective of this study was to compare subpopulations of L. esculentum germplasm from the AVRDC collection on the basis of RAPD-marker determined genetic relationships and diversity levels. Sub-population comparisons included (i) accessions from the primary compared with secondary centers, (ii) accessions from regions containing and contiguous with the primary center compared with secondary centers, and (iii) L. esculentum compared with cerasiforme accessions. The second objective of this study was to compare RAPD polymorphism patterns among AVRDC accessions with patterns among California processing tomatoes generated in an independent study (Villand, 1995).



Ninety-six of the 2000 morphologically characterized L. esculentum accessions in the AVRDC collection were chosen randomly by stratified proportional sampling within geographic regions: South America, Central America, Mexico, the Caribbean, Asia, Africa, Europe, and the USA. Two AVRDC breeding lines, accessions TL1496 and TL1328, were included by interest of the authors to understand their relationship to other accessions in the genebank. Geographic regions were sub-divided into (i) the primary center, (ii) a region containing and contiguous with the primary center including countries of South America, Central America, and Mexico, and (iii) secondary centers, including all countries except those of South America, Central America, and Mexico. For this study, countries in the primary center were Ecuador, Peru, and Chile. Passport data indicated 21 accessions were collected or received from Ecuador, Peru, and Chile, and an additional 37 accessions were collected or received from other countries of South America, Central America, and Mexico; 38 accessions were collected or received from secondary centers (Table 1). Cerasiforme accessions were discriminated from L. esculentum accessions by a key for the Lycopersicon species (Rick et al., 1990), which classifies cerasiforme as having two loculed fruit 1.5 to 2.5 cm in diameter. Characterization data collected by AVRDC for fruit diameter and locule number (not shown) resulted in classification of 10 accessions as cerasiforme: L00158, L00167, L00171, L00172, L00173, L00174, L00191, L00493, L01156, and L04360. For accessions where locule and fruit diameter data were not available, sub-species status was assigned as L. esculentum only if the fruit weight would preclude fruits from being 1.5 to 2.5 cm. Accessions L04313 and L04637 could not be classified on available data.

Prior to distribution by AVRDC, 30 plants per accession were grown for seed increase in cages to eliminate outcrossing and the seed was bulked. Therefore, individuals within accessions are assumed to be homozygous; however, variation among individuals within an accession is observed.

DNA Extraction and RAPD Reactions

Approximately 10 seeds were germinated for each accession and immature leaves from approximately six plants were harvested for DNA extraction. The DNA isolation procedure followed Johns et al. (1997). The RAPD reaction mixtures followed Skroch and Nienhuis (1995a) while RAPD cycling conditions followed Johns et al. (1997). The 10mer primers used in this study were A07, A15, A18, AB09, AC12, AD02, AFl4, AW08, AX20, B15, C02, C18, E18, F06, G02, G06, G16, I01, I12, I20, J20, K07, K16, M02, M05, N08, O10, P17, Q01, Q05, S01, S07, T14, T16, U03, U12, W06, W13, X11, Y02, and Y09 (Operon Technologies, Inc., Alameda, CA). Data was scored as the presence (1) or absence (0) of a fragment for each polymorphic marker.

RAPD Analysis

Genetic distances were calculated among all pairwise combinations of the 96 accessions. Genetic distances were calculated by the complement to the simple matching coefficient (Gower, 1972), Genetic [Distance.sub.i,j] = [n.sub.i[not equal to] j] / [n.sub.i=j] where n is the number of discordant or concordant marker comparisons between accessions i and j. Thus in this study, genetic distance is the ratio of discordant comparisons ([n.sub.i-j]) to the total number of comparisons ([n.sub.i[not equal to] j] + [n.sub.i=j]). A genetic distance equaling 0 or 1 indicates maximum similarity or difference between the pair of accessions, respectively. To visualize the relationships among the accessions, the distance matrix was converted to two dimensional coordinates by the monotonic multi-dimensional scaling procedure (MDS) with the Kruskal scaling option in Systat ver. 5.2 (Wilkinson et al., 1992).

Results from an independent study (Villand, 1995) were combined with the results from the present study to compare processing tomato cultivars to AVRDC accessions for the number of polymorphic RAPD bands amplified per RAPD primer and the correlations among RAPD markers (described below). Villand (1995) used RAPD markers to describe relationships among 20 California processing tomato genotypes that trace the development of the ideotype, VF65, to modern California processing tomato varieties. Twenty-seven RAPD primers were common between the present study and Villand (1995).

The number of polymorphic RAPD bands amplified per RAPD primer in this study was compared with results from Villand (1995) by sub-sampling the data to standardize genotype number (n = 20) for the 27 common primers. Random samples of 20 accessions were sampled with replacement 2000 times. For each sample, the mean number of polymorphic RAPD bands per primer was calculated and the mean over 2000 sub-samples was reported. Sub-sampling 2000 times was used in this and subsequent analyses since this sample number was sufficiently greater than that number where sampling variation becomes relatively unimportant.

To test the association among RAPD markers, simple linear correlations (Steel and Torrie, 1980) were computed for all pairwise combinations of polymorphic RAPD marker scores. RAPD marker correlations were also computed for a standardized number of genotypes (n = 20) to compare correlations in this study to results from Villand (1995). Random samples of 20 accessions were sampled with replacement 2000 times. For each sub-sample, the fraction of correlations among pairwise comparisons of RAPD marker scores was computed for each of 20 equally spaced intervals between 0.0 and 1.0. For each interval in the distribution, the mean fraction over 2000 sub-samples was reported.

Experimental error was calculated by sampling seed from accessions L00173 and L01175 twice and treating each separate DNA extraction as a separate entry. Experimental error was defined as the failure of accessions to be scored consistently over replications because of seed heterogeneity and random error in the generation and scoring of data. The number of scored inconsistencies between replicates was divided by the total number of markers scored for replicated accessions. For these two accessions, the replicate with the least missing data was used in the analysis.

Frequency for each RAPD marker was calculated as the frequency of the presence of RAPD amplification among accessions in a sub-population. Comparisons included (i) accessions collected in the primary center compared with secondary centers, (ii) accessions collected in South America, Central America, and Mexico compared with accessions from secondary centers, and (iii) L. esculentum compared with cerasiforme accessions. Spearman rank correlations between RAPD marker frequencies for sub-populations were calculated (SAS Institute, 1994). The magnitude of the difference in the frequency of RAPD fragment amplification for each RAPD marker was also used to compare sub-populations. For each pairwise comparison of sub-populations, the significance of the mean absolute difference in the frequency of RAPD fragment amplification was determined by randomization tests (Edgington, 1980) based on 10 000 permutations appropriate to each sub-population comparison. In addition, to evaluate the effect of large sample size differences between L esculentum and cerasiforme, the mean absolute difference in RAPD marker frequency and the corresponding P-value was computed for a comparison of 10 L. esculentum accessions randomly sampled with replacement and the 10 cerasiforme accessions, for each of 2000 sub-samples. Differences in RAPD marker frequency for individual RAPD loci evaluated were analyzed by two-way cross tabulation tables (FREO procedure; SAS Institute, 1990). The low frequency of RAPD alleles (described in Results and Discussion) required the computation of significance tests based on exact probabilities rather than approximations and thus, P-values were computed by Fisher's exact test (FREQ procedure; SAS Institute, 1990; Steel and Torrie, 1980).

Genetic diversity was estimated as RAPD marker diversity for each sub-population by Nei's gene diversity at a locus, h = (1 - [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII])n / (n - 1) = 2 pqn / (n - 1), where [x.sub.i] is the frequency of the ith allele, p is the frequency of the presence and q is the frequency of the absence of RAPD amplification among n accessions in a sub-population (Nei, 1987). Comparisons of sub-populations were done by t-tests for paired observations (Nei, 1987; Steel and Torrie, 1980). In addition, to evaluate the effect of large sample size differences between L. esculentum and cerasiforme, differences in mean RAPD marker diversity values and corresponding P-values were computed for a comparison of 10 L. esculentum accessions randomly sampled with replacement and the 10 cerasiforme accessions, for each of 2000 samples.


The 41 RAPD primers produced 98 scored polymorphic RAPD markers. Between one and six polymorphic bands per primer were scored with a mean ([+ or -] SD) of 2.27 ([+ or -] 1.25) bands per primer. This level of polymorphism was higher than among the population of 20 processing tomato cultivars in which Villand (1995) observed 1.66 ([+ or -] 0.79) bands per primer. Primers in both studies were chosen for the ability to generate a polymorphism across U. S. tomato cultivars. However, sample sizes and some RAPD primers differed in these two studies. By means of a common set of RAPD primers and identical genotype sample sizes, this study yielded a lower mean of 1.53 ( [+ or -] 1.76) bands per primer whereas the 20 processing tomato cultivars yielded 2.07 ( [+ or -] 1.28) bands per primer. Thus, the opposite result was obtained when the genotype sample sizes were standardized for both sets of germplasm. The original sample size of 96 AVRDC accessions resulted in the identification of many low frequency RAPD markers that were less likely to be detected in the much smaller sample of 20 processing tomato genotypes.

The mean frequency of the presence of a RAPD marker was 0.401 ([+ or -] 0.399). However, the frequency distribution (Fig. 1) illustrates that most RAPD markers are amplified with a very low or very high frequency. In addition, low and high frequency markers are conserved across sub-populations of accessions. The marker frequency distribution illustrates the problem that plant breeders and genebank curators face in the need to sample large numbers of accessions within the tomato germplasm to ensure inclusion of low-frequency alleles.

Genetic distances among the 4560 pairwise combinations of accessions ranged from 0.011 to 0.500 with a mean genetic distance of 0.164 ([+ or -] 0.084). The mean experimental error between replicates of accessions L00173 and L01175 was 2.5%. This level of experimental error indicates that differences among accessions of 0.025 or less could arise because of experimental error alone and may not reflect true genetic differences. In the present study, 23 pairwise comparisons of accessions had a difference [is less than or equal to] 0.025. Five RAPD markers produced inconsistent amplification across replications and these markers were removed from the data. All analyses were performed on the remaining 93 RAPD markers. Scoring errors described by Skroch and Nienhuis (1995b) were non-randomly distributed with a large portion of scoring errors distributed within a relatively small subset of RAPD markers. Therefore, the removal of RAPD markers with inconsistent amplification across replications is warranted.

Gene introgression from other Lycopersicon species has allowed plant breeders to broaden the germplasm base of L. esculentum cultivars (Rick, 1973). Williams and St. Clair (1993) reported an increase in molecular marker polymorphisms among modern L. esculentum cultivars compared with cultivars released in or before the 1960s, suggesting the presence of introgressed chromosomal regions from other Lycopersicon species. Rick and Fobes (1974) detected the same allozyme in six nematode resistant cultivars while this allele was not present in any other L. esculentum cultivar they studied. Lycopersicon peruvianum (L.) Miller was the source of the nematode resistance in the tested cultivars.

Villand (1995) found an increase in RAPD polymorphism in groups of tomato cultivars with genes introgressed from other Lycopersicon species compared with other processing tomatoes, suggesting that multiple polymorphisms were genetically linked to introgressed genes. In that study, consistent with the hypothesis of genetic clustering of RAPD polymorphisms near introgressed genes, 6.5% of all pairwise comparisons of RAPD markers (432 of 6689 comparisons) generated across processing tomato cultivars were perfectly correlated (identical RAPD marker patterns). In the present study, less than 0.2% of all pairwise comparisons of RAPD markers (8 of 4214 comparisons) were perfectly correlated, suggesting relatively less genetic clustering of RAPD polymorphisms in the AVRDC germplasm. However, the sample size of AVRDC accessions was larger than the collection of processing cultivars and thus high correlations among RAPD markers were less likely to occur by chance. When the genotype sample size for AVRDC accessions was limited to 20, the distributions for the AVRDC and processing tomato genotypes were similar (Fig. 2). In the evaluation of 2000 subsamples of 20 AVRDC accessions, some samples were observed having fractions of RAPD marker pairs with identical distributions among AVRDC genotypes (perfect correlations) greater than that observed for the processing genotypes ([is greater than] 6.5%). However, these samples occurred at a low frequency (P [is less than] 0.05) indicating that the fraction of pairwise marker comparisons having perfect correlations was significantly greater for the processing genotypes than for the AVRDC accessions. This result provides statistical support for the hypothesis that genetic clustering of RAPD polymorphisms is greater among processing tomato genotypes.


As noted above, the polymorphism level for processing tomato cultivars was greater than that of 20 AVRDC accessions. This result suggests that the diversity level of these processing tomato cultivars has actually been increased even though the processes of domestication and selection in breeding programs have acted to reduce diversity in cultivated tomato (Williams and St. Clair, 1993). However, examination of pedigrees and evaluation of the relationships among RAPD markers through correlation analysis suggests that the increase in polymorphism level observed for the processing tomato genotypes can be partly explained by genetically related markers associated with introgressed genes and not a genome-wide increase in marker polymorphism.

Divergence of Primary and Secondary Centers

The MDS analysis of 96 accessions (Fig. 3) indicated little difference in clustering among sub-populations. However, an alternative analysis to identify differences among sub-populations is to define genetically each subpopulation by allele frequencies, or analogously by RAPD marker frequencies. Any observed differences in these frequencies may reflect genetic differences among sub-populations (Nei, 1987).


The Spearman rank correlation for RAPD marker frequencies between primary and secondary center subpopulations was 0.86 (P [is less than] 0.01). The rank correlation for RAPD marker frequencies between South America, Central America, and Mexico and secondary center subpopulations was 0.90 (P [is less than] 0.01). Correlations illustrate the similarity of these populations in terms of RAPD markers. However, significant differences for individual RAPD marker frequencies between primary and secondary centers occurred for 12 of the 93 markers (Table 2). Individual RAPD marker frequency differences between South America, Central America, and Mexico compared with secondary centers occurred for seven of the 93 markers (Table 2) illustrating the small proportion of differences between the sub-populations on an individual RAPD locus basis. Overall, significant differences in RAPD marker frequency were observed between accessions from the primary center and those from the secondary centers, and accessions from all countries in South and Central America and Mexico, compared with secondary centers (P [is less than] 0.01) (Table 3). The significance of RAPD marker frequency differences is due not only to the individual RAPD markers listed in Table 2, but also the summation of many small differences at other RAPD loci since after removal of RAPD markers listed in Table 2, significant differences remain between sub-populations (data not shown).


High correlations among sub-populations are expected since all accessions are likely derived from one or few geographic regions. However, unique allele frequencies at specific RAPD loci and the overall difference in magnitudes of marker frequencies are consistent with the divergence of sub-populations after germplasm was moved from its original geographic range. Some accessions from all three sub-populations tested here should be included in collection, maintenance, or sampling of germplasm if the range of genetic variation available in the AVRDC collection is to be represented.

Genetic diversity in the primary center is expected to be greater than in the secondary centers because of the bottleneck effect (Nei, 1987). In this study, RAPD marker diversity, used to estimate genetic diversity, was greater in the primary center (mean = 0.219) (Table 3) than in either the secondary centers (0.137) or South America, Central America, and Mexico (0.175) suggesting a bottleneck or founder effect as germplasm was moved away from the primary center. However, reasons for the lack of a significant difference in RAPD marker diversity between accessions from South America, Central America, and Mexico compared with secondary centers include, but are not limited to, (i) the variation present in L. esculentum in many regions of the New World was adequately sampled as germplasm was moved to secondary centers, (ii) selection pressure reduced variation in cultivars, breeding lines, and landraces from both these regions in a similar manner, and (iii) other post-domestication events, such as adaptation, generated similar levels of diversity in germplasm in these regions. Alternatively, the samples used in this study, while indicating the expected difference, may not have been large enough for the difference to be statistically significant.

Rick and Fobes (1975) also reported differences between sub-populations of L. esculentum genotypes from Peru and Ecuador compared with genotypes from other South American countries, Central America, Mexico, the USA, and Europe on the basis of contingency tests of allozyme variation. In addition, Rick and Fobes (1975) reported greater allozyme variation among genotypes from Peru and Ecuador than from those outside of this region. The present study expands this comparison to include accessions from a wide range of secondary centers. Despite sampling germplasm from a wide geographic range, we did not identify germplasm as diverse as that found in the primary center. Thus, when the goal is to obtain a diverse subset of germplasm, variation can be obtained at a faster rate by sampling accessions from the primary center versus other geographic regions.

Divergence between U esculentum and cerasiforme

Cerasiforme accessions were spread throughout the cluster of L. esculentum accessions in the MDS analysis (Fig. 3). The Spearman rank correlation for RAPD marker frequencies between L. esculentum and cerasiforme was 0.88 (P [is less than or equal to] 0.01). However, differences in individual RAPD marker frequencies between L. esculentum and cerasiforme occurred for eight of the 93 RAPD markers (Table 2). Overall, RAPD marker frequency differences did discriminate between sub-populations (P [is less than or equal to] 0.01) (Table 3). Consistent with this result, when equal numbers of accessions were compared on the basis of 1000 random sub-samples, the difference was always significant. However, upon removal of cerasiforme accessions L04360, L00158, and L00167, which appear to be the most unique (Fig. 3), the difference between L. esculentum and cerasiforme became nonsignificant (data not shown).

The high correlation observed for marker frequencies between sub-populations and the divergence of subpopulations indicated by the significance of the mean difference in marker frequency are consistent with the hypothesis that cerasiforme is the ancestor of L. esculentum. The sympatric nature of the two populations and their ability to hybridize (Rick, 1973) may have resulted in the similarity between L. esculentum and most of the cerasiforme accessions studied. However, uniqueness of some cerasiforme accessions indicates that this hypothesized mixing of L. esculentum and cerasiforme germplasm is incomplete. The similarity of a large number of RAPD loci between L. esculentum and cerasiforme was also found by Williams and St. Clair (1993), who reported that the majority of cerasiforme genotypes were interspersed among L. esculentum in a dendrogram based on allelic identity. The non-homogeneity between these sub-populations highlights the need for inclusion of cerasiforme accessions as a component of the genetic variation available at AVRDC but the limited number of accessions contributing to this variation renders case-by-case decisions necessary.

RAPD marker diversity was not significantly different between L. esculentum (0.155) and cerasiforme (0.224; P = 0.101) (Table 3). When equal numbers of accessions were compared (on the basis of 2000 samples), none of the comparisons were significant. Rick and Fobes (1975) and Rick and Holle (1990) reported considerable allozyme variation among cerasiforme genotypes collected in coastal Peru and Ecuador but little allozyme variation has been revealed for cerasiforme collected elsewhere. This study included only two cerasiforme accessions from Peru (L00158 and L00191) and one accession from Ecuador (L04360) (Table 1). Therefore, the result indicating a non-significant difference in marker diversity level between L. esculentum and cerasiforme in this study may be expected. The lack of an increase in RAPD marker diversity illustrates that inclusion of accessions identified as cerasiforme alone will not necessarily increase the diversity of the AVRDC collection. In addition, two of the most unique cerasiforme accessions in this study were collected in Peru (L00158) and Ecuador (L04360) further emphasizing the importance of germplasm collected in the primary center.

Germplasm Collection and Breeding Implications

With little pedigree information available to understand the relationships among accessions, characterization data is often collected. The cost (approximately $25; L.M. Engle, 1996, personal communication) to multiply and characterize one AVRDC accession precludes systematic characterization of all accessions. Approximately one-half of the AVRDC L. esculentum accessions are characterized by the International Board for Plant Genetic Resources tomato descriptor list (Tay, 1989). In addition, morphological data such as heat tolerance, bacterial wilt resistance, tomato mosaic virus resistance, nematode resistance, and agronomic characteristics have been collected and used by AVRDC breeders to sample germplasm for breeding programs in tropical regions (Opena et al., 1989). However, morphological variation may not clearly indicate the level of genetic variation since there is difficulty in distinguishing variation resulting from differences at one locus versus many loci (Brown, 1978). Molecular marker data can be used to complement passport and characterization data for understanding the distribution of genetic variation in crop species (Brown, 1989; Bernatzky and Tanksley, 1989; Beebe et al., 1995). This study demonstrates how RAPD markers can be used to understand relationships among AVRDC L. esculentum accessions.

Lacking knowledge of the distribution of variation in a species, Brown (1978) suggested that a diverse germplasm collection would be achieved by sampling from as many distinct geographic regions as possible; however, this strategy assumes that each region harbors the same level of variation. If it were possible to rank populations by diversity levels then sample sizes could be set in proportion to the relative diversity level of each population. Schoen and Brown (1993) describe how molecular markers might be used to set such sampling priorities. Similarly, in this study we have used RAPD markers to understand the uniqueness of different parts of the AVRDC tomato germplasm collection and to rank these sub-populations on the basis of marker diversity levels. The results indicate that unique diversity can be sampled from all sub-populations studied and therefore, germplasm from each region, including both L. esculentum and cerasiforme, must be sampled to obtain the range of genetic variation available in the AVRDC collection. However, comparison of marker diversity levels indicates that variation can be obtained at a faster rate by sampling accessions from the primary center of diversity versus other world regions.

RAPD data have indicated divergence of sub-populations of accessions from primary versus secondary centers of diversity. Consistent with this result, sampling from the full range of environments in the L. esculentum collection has been important to AVRDC in the development of heat tolerant lines that set fruit under high temperatures. After screening 4050 AVRDC accessions, 38 accessions were identified as heat tolerant, 30 of these originated from secondary centers (Villareal et al., 1978). Selection for heat tolerance as L. esculentum was disseminated to secondary centers may explain the unequal distribution of variation for this trait. More importantly, inclusion of accessions in the collection from a wide range of environments allowed identification of variation for this trait.


We thank Dr. Liwayway M. Engle for assistance and knowledge of the AVRDC genebank and Michell E. Sass for technical assistance.


Beebe, S.E., I. Ochoa, P. Skroch, J. Nienhuis, and J. Tivang. 1995. Genetic diversity among common bean breeding lines developed for Central America. Crop Sci. 35:1178-1183.

Bernatzky, R., and S.D. Tanksley. 1989. Restriction fragments as molecular markers for germplasm evaluation and utilization. In A.H.D. Brown et al. (ed.) The use of plant genetic resources. Cambridge Univ. Press, Cambridge, England.

Brown, A.H.D. 1978. Isozymes, plant population genetic structure and genetic conservation. Theor. Appl. Genet. 52:145-157.

Brown, A.H.D. 1989. Core collections: a practical approach to genetic resources management. Genome 31:818-824.

Edgington, E.S. 1980. Randomization tests. Marcel Dekker, Inc., New York.

Gower, J.C. 1972. Measures of taxonomic distances and their analysis. In J.S. Weiner and J. Huizinga (ed.) The assessment of population affinities in man. Clarendon Press, Oxford, England.

Jenkins, J.A. 1948. The origin of the cultivated tomato. Econ. Bot. 2:379-392.

Johns, M.A, P.W. Skroch, J. Nienhuis, P. Hinrichsen, G. Bascur, and C. Munoz-Schick. 1997. Gene pool classification of common bean landraces from Chile based on RAPD and morphological data. Crop Sci. 37:605-613.

Luckwill, L.C. 1943. The genus Lycopersicon, an historical, biological, and taxonomic survey of the wild and cultivated tomatoes. Aberdeen University Studies. No. 120, Aberdeen, Scotland.

McCue, G.A. 1952. The history of the use of the tomato: an annotated bibliography. Ann. Missouri Bot. Gard. 39:289-348.

Miller, J.C., and S.D. Tanksley. 1990. RFLP analysis of phylogenetic relationships and genetic variation in the genus Lycopersicon. Theor. Appl. Genet. 80:437-448.

Muller, C.H. 1940. A revision of the genus Lycopersicon. USDA Misc. Publ. 382. U,S. Gov. Print. Office, Washington, DC.

Nei, M. 1987. Molecular evolutionary genetics. Columbia Univ. Press, New York.

Opena, R.T., S.K. Green, N.S. Talekar, and J.T. Chen. 1989. Genetic improvement of tomato adaptability to the tropics: progress and future prospects, p. 70-85. in S.K. Green et al. (ed.) Tomato and pepper production in the tropics. Proc. Int. Symp. on Integrated Management Practices, AVRDC Publ. No. 89-317, Tainan, Taiwan.

Rick, C.M. 1958. The role of natural hybridization in the derivation of cultivated tomatoes of western South America. Econ. Bot. 12:346-367.

Rick, C.M. 1973. Potential genetic resources in tomato species: clues from observations in native habitats. In A.M. Srb (ed.) Genes, enzymes, and populations. Plenum Press, New York.

Rick, C.M., and J.F. Fobes. 1974. Association of an allozyme with nematode resistance. Tomato Genetics Coop. Rep. 24:25.

Rick, C.M., and J.F. Fobes. 1975. Allozyme variation in the cultivated tomato and closely related species. Bull. Torrey Bot. Club 102: 376-386.

Rick, C.M., and M. Holle. 1990. Andean Lycopersicon esculentum var. cerasiforme: Genetic variation and its evolutionary significance. Econ. Bot. 44(Supp.):69-78.

Rick, C.M., H. Laterrot, and J. Philouze. 1990. A revised key for the Lycopersicon species. Tomato Genetics Coop. Rep. 40:31.

SAS Institute, Inc. 1990. SAS user's guide: Statistics, ver. 6, 4th edition. SAS Inst. Inc., Cary, NC.

SAS Institute, Inc. 1994. JMP, ver. 3.1. SAS Inst. Inc., Cary, NC.

Schoen, D.J., and A.H.D. Brown. 1993. Conservation of allelic richness in wild crop relatives is aided by assessment of genetic markers. Proc. Natl. Acad. Sci. (USA) 90:10623-10627.

Skroch, P.W., and J. Nienhuis. 1995a. Qualitative and quantitative characterization of RAPD variation among snap bean (Phaseolus vulgaris) genotypes. Theor. Appl. Genet. 91:1078-1085.

Skroch, P.W., and J. Nienhuis. 1995b. Impact of scoring error and reproducibility of RAPD data on RAPD based estimates of genetic distance. Theor. Appl. Genet. 91:1086-1091.

Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics: a biometrical approach. McGraw-Hill Inc., New York.

Tay, D.C.S. 1989. Genetic resources of tomato and pepper at AVRDC. p. 10-21. In S.K. Green et al. (ed.) Tomato and pepper production in the tropics. Proc. Int. Symp. on Integrated Management Practices, AVRDC Publ. No. 89-317, Tainan, Taiwan.

Vavilov, N.I. 1926. Studies on the origin of cultivated plants. Bull. of Applied Botany and Plant Breeding 16:139-245.

Villand, J.M. 1995. Comparison of molecular marker and morphological data to determine genetic distance among tomato cultivars. M.S. Thesis. Univ. of Wisconsin., Madison.

Villareal, R.L., S.H. Lai, and S.H. Wong. 1978. Screening for heat tolerance in the genus Lycopersicon. HortScience 13:479-481.

Welsh, J., and M. McClelland. 1990. Fingerprinting genomes using PCR with arbitrary primers. Nuc. Acids Res. 18:7213-7218.

Wilkinson, L., M. Hill, S. Miceli, P. Howe, and E. Vang. 1992. SYSTAT for the Macintosh, vet. 5.2. SYSTAT Inc. Evanston, IL.

Williams, J.G.K., A.R. Kubelik, K.J. Livak, J.A. Rafalski, and S.V. Tingey. 1990. DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nuc. Acids Res. 18:6531-6535.

Williams, C.E., and D.A. St. Clair. 1993. Phenetic relationships and levels of variability detected by restriction fragment length polymorphism and random amplified polymorphic DNA analysis of cultivated and wild accessions of Lycopersicon esculentum. Genome 36:619-630.

Willis, J.C. 1922. Age and area. A study in geographical distribution and origin of species. Cambridge Univ. Press, London.

Zeven, A.C., and J.M.J. de Wet 1982. Dictionary of cultivated plants and their regions of diversity. Centre for Agricultural Publishing and Documentation, Wageningen, the Netherlands.

J. Villand,(*) P. W. Skroch, T. Lai, P. Hanson, C. G. Kuo, and J. Nienhuis

J. Villand, P.W. Skroch, and J. Nienhuis, Dep. of Horticulture, 1575 Linden Dr., Univ. of Wisconsin, Madison, WI 53706; and T. Lai, P. Hanson, and C.G. Kuo, The Asian Vegetable Research and Development Center, Shanhua, Tainan, Taiwan 741. Received 14 April 1997. (*) Corresponding author (

Abbreviations: AVRDC, Asian Vegetable Research and Development Center; cerasiforme, Lycopersicon esculentum var. cerasiforme; MDS, multi-dimensional scaling; RAPD, random amplified polymorphic DNA.
COPYRIGHT 1998 Crop Science Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Villand, J.; Skroch, P.W.; Lai, T.; Hanson, P.; Kuo, C.G.; Nienhuis, J.
Publication:Crop Science
Date:Sep 1, 1998
Previous Article:Applications of biotechnology in turfgrass genetic improvement.
Next Article:Identification of diverse soybean germplasm using RAPD markers.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters