Crop Science 42:291-296 (2002)
© 2002 Crop Science Society of America
PLANT GENETIC RESOURCES
Comparative Assessment of Variation among Sorghum Germplasm Accessions Using Seed Morphology and RAPD Measurements
J. A. Dahlberg*,a,
X. Zhangb,
G. E. Hartb and
J. E. Mulletb
a NGSP, P.O. Box 5309, Lubbock, TX 79408
b Crop Improvement Center, Southern Crop Improvement Facility, Texas A&M Univ., College Station, TX 77843-2123
* Corresponding author (jeff{at}sorghumgrowers.com)
 |
ABSTRACT
|
|---|
The sorghum germplasm collection currently contains over 42 000 accessions, a number that is too large to manage efficiently. The specific objective of this research was to compare clusters developed from agronomic descriptors with phylogenetic groupings based on random amplified polymorphic DNA (RAPD) fingerprinting of selected sorghum [Sorghum bicolor (L.) Moench] races. Our intent was to identify one approach using agronomic descriptors that would most closely approximate the groupings produced by RAPD markers. Ninety-four accessions of sorghum were grouped into four of the five major races. Differences among accessions determined by various clustering procedures based on agronomic characteristics were compared with clusters developed by means of RAPD markers. Each race varied in the degree of similarity between the four clustering approaches taken and the information provided by RAPD fingerprinting. Test 2, standardization of data by Z-scores and cluster analysis using the complete set of data, provided the highest similarity score for the race bicolor, while Test 3, standardization of data by Z-scores and cluster analysis based on a reduced set of variables selected from principle component analysis, provided the highest similarity scores for the races guinea. Test 1, random selection, was highest for the races caudatum and durra. When averaged over all the races, Test 2 provided the highest similarity score. The results of this study indicate that no one approach to develop clusters by means of agronomic descriptors closely approximate the groupings produced by RAPD markers. These results underscore the need for further research in the evaluation of techniques used to develop core collections and their validity.
|
INTRODUCTION
|
|---|
PLANT BREEDERS and other scientists interested in maintaining and enhancing genetic diversity have recognized for many years the importance of obtaining plant introductions through the collection of plants in native habitats. Collections have been made of many diverse genotypes of cultivated, wild, and weedy species that may have potential as breeding material in the future (Fehr, 1987). Despite the importance of sorghum as a food and feed crop in the world, existing literature on germplasm curation for it is limited. More effective methods are needed for identifying and using exotic sorghum germplasm. Growers and U.S. sorghum breeders have shown a strong interest in preserving exotic sorghum germplasm. This has led to a collection that currently inventories over 42 000 accessions, a number that is too large to manage efficiently.
Brown (1988) argued "that a better collection is one that is rationalized, refined and structured, around a small, well-defined and representative core." Frankel (1984) suggested that a core collection represent, "with a minimum of repetitiveness, the genetic diversity of a crop species and its relatives." Brown et al. (1987) developed the first core collection from the Australian collection of perennial Glycine species. Core collections have since been developed for various species such as annual Medics (Bauchan et al., 1992), perennial Medicago (Basigalup, 1991; Basigalup et al., 1995), annual Medicago (Diwan et al., 1994), and Arachis hypogaea L. (Holbrook et al., 1993).
The overall goal of the sorghum curator is to facilitate the development of a sorghum core collection from the 42 000 accessions of sorghum currently under USDA maintenance and supervision. Key information needed is a measure of the genetic variation within and among accessions. This could best be achieved by genetic profiling techniques which rely on simple sequence repeats (SSR), restriction fragment length polymorphism (RFLP), RAPD, and amplified fragment length polymorphic DNA (AFLP) analyses. With this information, the core collection could be formulated to ensure an adequate sampling of genetic variation. RAPDs have been used to determine genetic relationships in Zea mays L., Gliricidia sepium (Jacq.) Walp, G. maculata (Kunth) Walp., Mangifera indica L., and Phaseolus vulgaris L. (Welsh et al., 1991; Chalmers et al., 1992; Schnell et al., 1995; Skroch and Nienhuis, 1995). Each paper provides information that supported the use of RAPDs in the identification of unique cultivars or populations based on genetic diversity.
Though DNA profiling is the current method of choice in measuring genetic variation within germplasm collections, the use of such technologies in large collections is costly, and both labor and time consuming. Alternative strategies, therefore, are required. One such strategy is the use of imaging technology to gather quantitative data on seed morphology. Dahlberg and Wasylikowa (1996) used these techniques to classify 8000-yr-old sorghum seed as wild species and showed that these measurements can be used to distinguish between the various races and wild species.
The specific objective of this research was to compare phylogenetic groupings of four of the major sorghum races based on seed morphology clustering with phylogenetic groupings based on RAPD profiles subjected to UPGMA clustering. Our intent was to identify one approach using agronomic descriptors that would most closely approximate the groupings produced using RAPD markers.
|
MATERIALS AND METHODS
|
|---|
Ninety-four sorghum accessions were chosen from the National Collection. Accessions were planted at the USDA-ARS Tropical Agriculture Research Station farm located at Isabela, Puerto Rico. The soil is an Oxisol (Tropeptic Haplorthox). Accessions were evaluated for race characterization. The accessions were not planted in a replicated trial, but handled as typical regenerations for sorghum within the National Plant Germplasm System. Because sorghum is photoperiod sensitive, most of the accessions within the world collection must be regenerated in the tropics. Given the size of the collection, replicated increases for descriptive characterization is not a viable option.
These accessions were then grouped into four of the five major races of sorghum on the basis of work done by Harlan and de Wet (1972). Sorghum bicolor subsp. bicolor contains all of the cultivated sorghums. Dahlberg (2000) provides a review of the classification and characterization of sorghum. On the basis of this review, sorghum can be classified into five major races and 10 intermediate races (for descriptions of each, see Dahlberg 2000).
The accessions used in the studies included 37 caudatums, 21 bicolors, 16 guineas, and 20 durras (Table 1). Five sorghum kernels from each accession were analyzed by means of imaging techniques described by Dahlberg and Wasylikowa (1996). Measurements on Area, Breadth, Circularity, Major Axis Length, Perimeter Length, and Rectangularity (BioScan, Inc., 1992) were replicated five times. Both Circularity and Rectangularity are mathematical expressions that give a measure of how closely a boundary approximates either a circle or a rectangle (BioScan, Inc., 1992). Data were analyzed by SAS (1985).
View this table:
[in this window]
[in a new window]
|
Table 1. ID, plant introduction (PI) number, international sorghum (IS) number, race, and country of origin of the 94 sorghum accessions used in this study.
|
|
DNA was extracted by a modified cetyltrimethylammonium bromide (CTAB) protocol. Leaf tissue was collected and frozen at -80°C from bulks of 10 to 15, 3- to 4-wk-old sorghum plants grown in a greenhouse. The RAPD methods described by Pammi et al. (1994) were used. Approximately 500 random 10-mer primers (Operon Technologies, Inc., Alameda, CA) were used to screen BTx623 and IS 3620C, the parents of a sorghum recombinant inbred population that is being used to integrate several sorghum genetic maps. Ninety-five percent of the primers amplified genomic DNA of the two lines. From 16 to 42 medium and bold fragments were amplified by each primer. Primers were selected for use in RAPD fingerprinting of sorghum accessions on the basis of their ability to amplify several major fragments in BTx623 and IS 3620C.
PCR reaction conditions were as follows: 7.5 µL reaction volume, 0.0375 µL Ampli-Taq Stoffle, 0.75 µL of reaction buffer, 0.1 mM MgCl2, 1 mM each of dATP, dGTP, and dTTP; 0.125 µL of 1 mM dCTP and 0.08 µL of dCTP32 to label PCR products; 4.8 ng of DNA and 75 ng operon random primer and ddH2O to bring the reaction volume to 7.5 µL. Amplifications were performed in a Perkin Elmer (Norwalk, CT) Cetus Thermal Cycler programmed for 30 cycles of 30 s at 92°C, 30 s at 48°C, and 45 s at 72°C. PCR products were resolved on 5% (w/v) polyacrylamide sequencing gels and detected by autoradiography. The RAPDs were scored as present (1) or absent (0). Using the SIMQUAL program and DICE coefficients on 53 RAPD bands, a matrix was computed and phenograms developed by a UPGMA clustering of the matrix using NTSYS-pc (Rohlf, 1994). Trees were printed by the tree function within the graphics programs window. The races bicolor and guinea were grouped on the basis of similarity levels of 7.4, while race durra was grouped on the basis of a similarity level of 6.4 and the race caudatum was grouped based upon a similarity score of 6.0.
Four clustering procedures were used to develop phenograms within each race. The first approach grouped accessions within each race on the basis of simple random selection. The second approach used seed trait measurements that were standardized by Z-scores. Ward's minimum variance cluster analysis (SAS, 1985) was then used on this data set to cluster accessions. The third approach used standardized seed measurements to develop a reduced set of variables by principle component analysis. This data set was then clustered by means of Ward's minimum variance cluster analysis. The fourth approach included standardization of the seed measurements by Z-scores and the development of clusters based on retained principal components was performed by Ward's minimum variance cluster analysis (see Basigalup et al., 1995). On the basis of these analyses, four groups clustered within each race.
Comparison of clusters developed by these techniques with those developed from RAPD information were evaluated for similarity between clusters by means of "Objective criteria for evaluation of clustering methods" described by Rand (1971). Rand surmised that "... a basic unit of comparison between two clusterings is how pairs of points are clustered. If the elements of an individual point-pair are placed together in a cluster in each of the two clusterings, or if they are assigned to different clusters in both clusterings, this represents a similarity between the clusterings, as opposed to the case in which the elements of the point-pair are in the same cluster in one clustering and in different clusters in the other. From this, a measure of similarity between two clusterings of the same data, Y and Y', can be defined ..." From this, a criterion was developed to evaluate similarity between clusterings developed from RAPD data and those developed from seed morphological data were then evaluated. From n objects (n = number of accessions within each race), the total number of pairs (TP) can be calculated as [n(n - 1)]/2. Let x = number of pairs of accessions that are in the same cluster in both clusterings; y = number of pairs of accessions that are in different clusters in both clusterings; and z = number of pairs of accessions that are in the same cluster in one clustering and different clusters in the other clustering, then x + y + z = TP. The number of pairs that are in agreement is x + y and the proportion of correctly clustered pairs = (x + y)/(x + y + z).
|
RESULTS AND DISCUSSION
|
|---|
Analysis of variance of image analyses of sorghum kernels indicated significant differences existed between each race for each of the measurements taken. Means, CV, and LSD values are presented in Table 2. Measurements for the races guinea and durra were larger than for the races bicolor and caudatum except for ArCircularity and ArRectangularity. Larger measurements generally correspond to larger, though not necessarily heavier, seed size and both durra and guinea seed tend to be large seeded sorghums. Ranges for each measurement are presented in Table 2 and it was this variation within the races that were used to develop clusters for comparison purposes. The ANOVA indicated that significant differences existed within each race for the various agronomic measurements taken.
View this table:
[in this window]
[in a new window]
|
Table 2. Means and ranges () for measurements taken on seed samples of germplasm accessions of four major sorghum races.
|
|
Seventeen primers were used to detect RAPDs among the 94 accessions. Table 3 summarizes the profiles generated by the 17 primers. More than 70% of the medium and bold fragments that were amplified were shared by all accessions. The trees produced from the RAPD data and cluster groups are presented in Fig. 1 through 4
. Results from these comparisons are presented in Table 4.
View this table:
[in this window]
[in a new window]
|
Table 4. Similarity scores within sorghum germplasm races for comparison of clusterings based on RAPD profiles with four clustering techniques using seed morphological measurements.
|
|
Each race varied in the degree of similarity between the four clustering approaches taken and the information provided by RAPD fingerprinting. Test 2, standardization of data by Z-scores and cluster analysis based on the complete set of data, provide the highest similarity score for the race bicolor, while Test 3, standardization of data by Z-scores and cluster analysis based on a reduced set of variables selected from principle component analysis, provided the highest similarity scores for the races guinea. Test 1, random selection, was highest for the races caudatum and durra. When averaged over all the races, Test 2 provided the highest similarity score. Though the methods of comparison were different, these results differ from those put forth by Basigalup et al. (1995), who suggested that either directed selection of a core based on evaluation traits or development of a core based on random selection of a cluster using retained principal component analysis provided the highest overall score of similarity.
Brown and Schoen (1994) proposed that core collections should be based on "structured or stratified random sampling." The target collection (or area, or species) is first divided into a number of groups—the groups being genetically, ecologically, or geographically distinct. A sample is then drawn from each group." The results of this study indicate that no one approach to develop clusters by means of agronomic descriptors closely approximate the groupings produced from RAPD markers. In the case of two sorghum races, caudatum and durra, random selection was just as effective as other techniques in comparing agronomic clusters and RAPD clusters. These results underscore the need for further research in the evaluation of techniques used to develop core collections and their validity.
In many respects, these finding were not unexpected. High correlations between agronomic descriptors and RAPD or any other current DNA techniques may be difficult. Seed morphology may be an important trait for developing a core collection; however, it is quite likely that seed morphology, along with most phenotypic traits, are controlled by several genes which may be highly influenced by environment. Most of these traits also sample a very small region of the genome. In many cases, DNA profiling, given the current technology, is a refined random evaluation of the genome. In the case of RAPDs, they are samples of DNA markers randomly distributed throughout the genome. Finding a high correlation between these different sampling techniques would be very difficult. It may require that the full genome of a crop species be identified and genes sequenced before a true measure of genetic variation can be measured and used for core collection development. Given the current limitations of the knowledge base of the sorghum genome, clustering on the basis of agronomic characteristics will continue to be the most economical method by which to measure diversity within the collection, though further work is needed to evaluate the best method for this sampling.
Received for publication February 3, 1999.
|
REFERENCES
|
|---|
- Basigalup, D.H. 1991. Development of a core collection for alfalfa (Medicago sativa L.). Ph.D. Thesis (Diss. Abstr. 91-25791), Univ. of Minnesota, Minneapolis, MN.
- Basigalup, D.H., D.K. Barnes, and R.E. Stucker. 1995. Development of a core collection for perennial Medicago plant introductions. Crop Sci. 35:1163–1168.[Abstract/Free Full Text]
- Bauchan, G., N. Diwan, and M. McIntosh. 1992. Development of an annual Medic core collection. p. 5. Proc. 33rd N. Am. Alfalfa Improv. Conference, Atlanta, GA. June 14–18, 1992. NAAIC, Beltsville, MD.
- BioScan, Inc. 1992. BioScan optimas technical reference, Fifth Edition. BioScan, Incorporated, Edmonds, WA.
- Brown, A.H.D. 1988. Core collections: A practical approach to genetic resources management. CSIRO., Div. Plant Industry, Canberra, Australia.
- Brown, A.H.D., and D.J. Schoen. 1994. Optimal sampling strategies for core collections of plant genetic resources. p. 357–370. In V. Loeschcke et al. (ed.) Conservation genetics. Birkhuser Verlag, Basal, Switzerland.
- Brown, A.H.D., J.P. Grace, and S.S. Spear. 1987. Designation of a "core" collection of perennial Glycine. Soybean Genet. Newsl. 14:59–70.
- Chalmers, K.J., R. Waugh, J.I. Sprent, A.J. Simons, and W. Powell. 1992. Detection of genetic variation between and within populations of Gliricidia sepium and G. maculata using RAPD markers. Heredity 69:465–472.
- Dahlberg, J.A. 2000. Classification and characterization of sorghum. p. 99–130. In C.W. Smith and R.A. Frederiksen (ed.) Sorghum: Origin, history, technology, and production. John Wiley & Sons, Inc., New York.
- Dahlberg, J.A., and K. Wasylikowa. 1996. Image and statistical analyses of early sorghum remains (8000 B. P.) from the Nabta Playa archaeological site in the Western Desert, southern Egypt. Veg. Hist. Archaeobot. 5:293–299.
- Diwan, N., G.R. Bauchan, and M.S. McIntosh. 1994. A core collection for the United States annual Medicago germplasm collection. Crop Sci. 34:279–285.[Abstract/Free Full Text]
- Fehr, W.R. 1987. Principles of cultivar development: Vol. 1 Theory and technique. Macmillan Publ. Co., New York.
- Frankel, O.H. 1984. Genetic perspectives of germplasm conservation. p. 161–170. In W. Arber et al. (ed.) Genetic manipulation: Impact on man and society. Cambridge Univ. Press, Cambridge, England.
- Harlan, J.R., and J.M.J. de Wet. 1972. A simplified classification of cultivated sorghum. Crop Sci. 12:172–176.[Abstract/Free Full Text]
- Holbrook, C.C., W.F. Anderson, and R.H. Pittman. 1993. Selection of a core collection from the U.S. germplasm collection of peanut. Crop Sci. 33:859–861.[Abstract/Free Full Text]
- Pammi, S., K. Schertz, G. Xu, G. Hart, and J.E. Mullet. 1994. Random-amplified-polymorphic DNA markers in sorghum. Theor. Appl. Genet. 89:80–88.
- Rand, W.M. 1971. Objective criteria for the evaluation of clustering methods. J. Am. Stat. Assoc. 66:846–850.[ISI]
- Rohlf, F.J. 1994. NTSYS-pc: Numerical taxonomy and multivariate analysis system, Ver. 1.80. Exeter Software, Setauket, NY.
- SAS Institute Inc. 1985. SAS user's guide: Statistics, 5th edition. Cary, NC.
- Schnell, R.J., C.M. Ronning, and R.J. Knight, Jr. 1995. Identification of cultivars and validation of genetic relationships in Mangifera indica L. using RAPD markers. Theor. Appl. Genet. 90:269–274.
- Skroch, P.W., and J. Nienhuis. 1995. Qualitative and quantitative characterization of RAPD variation among snap bean (Phaseolus vulgaris) genotypes. Theor. Appl. Genet. 91:1078–1085.[ISI]
- Welsh, J., R.J. Honeycutt, M. McClelland, B.W.S. Sobral. 1991. Parentage determination in maize hybrids using the arbitrarily primed polymerase chain reaction (AP-PCR). Theor. Appl. Genet. 82:473–476.
Related articles in Crop Science:
- This issue in Crop science
Crop Science 2002 42: 1-2.
[Full Text]
TOP
ABSTRACT
INTRODUCTION
