RAPD Based Assessment of Genetic Diversity in Cultivated Races of Sorghum Abebe Menkir, Peter Goldsbrough, and Gebisa Ejeta* ABSTRACT Analysis of the extent and distribution of genetic diversity in crop plants is essential for optimizing sampling and breeding strategies. We used random amplified polymorphic DNA (RAPD) markers to assess genetic diversity and taxonomic relationships in 190 accessions sampled to represent the cultivated races of sorghum [Sorghum bicolor (L.) Moench], A high level of variation was detected among cultivated genotypes. Partitioning the genetic variation in cultivated sorghum with Shannon's diversity index revealed that 86% of the total genetic variation occurred among accessions and 14% among races. We also examined the degree of association of accessions with their geographic areas of origin by using Shannon's diversity index. The results indicated that only 13% of the total genetic variation was attributable to diver- gence among regions. Further tests using principal component analysis also failed to show separation of accessions into discrete racial or geographic groups. Despite such limited differentiation among races or regions, RAPD markers successfully identified races and regions with maximum genetic diversity. For example, accessions within races bicolor and guinea had greater genetic diversity than accessions from race kafir. Accessions from southern Africa had a lower level of genetic diversity than accessions from Far and Middle East, Central and Eastern Africa, and Western Africa. Thus, use of RAPD markers may optimize sampling of genetically divergent accessions for intro- gression into breeding pools. S ORGHUM is fifth in worldwide economic importance among cereal crops (Doggett, 1988). The range of sorghum-growing environments including longitude, lat- itude, precipitation, elevation, temperature, and length of the growing period is quite broad. The crop also is cultivated for a wide variety of purposes by peoples of different cultural heritage. Owing to the selection pres- sure imposed by such diverse environmental factors and cultivators, the sorghum crop has evolved into a vast array of forms (Morden et al., 1989). Analyses of the extent and distribution of genetic varia- tion in a crop are essential in understanding the evolution- ary relationships between accessions and to sample ge- netic resources in a more systematic fashion for breeding and conservation purposes. Traditionally, taxonomists classified genetic resources in sorghum based mainly on morphological markers (Harlan and de Wet, 1972; Murty etal., 1967; Snowden, 1936). Harlan and de Wet (1972) Abebe Menkir and Gebisa Ejeta, Dep. of Agronomy, Purdue Univ., W. Lafayette, IN 47907-1150; and Peter Goldsbrough, Dep. of Horticulture, Purdue Univ., W. Lafayette, IN 47907-1165. Contribution of the Indiana Agric. Exp. Stn. Purdue J. no. 14962. Received 3 March 1996. "Corre- sponding author (E-mail: gejeta@dept.agry.purdue.edu). Published in Crop Sci. 37:564-569 (1997). used a small number of easily recognizable traits, includ- ing grain shape, glumes, and panicle shape, to partition variability in cultivated sorghums into five races and 10 intermediate (hybrid combination of the major races) forms. The morphological traits used by Harlan and de Wet (1972) are conditioned by a relatively small number of genes (Doggett, 1988). However, several complex quantitative traits which are related to habitat adaptation and particular end use of the crop exhibit enormous variability among sorghum accessions within each race (de Wet et al., 1976). Thus, classifying germplasm acces- sions based solely on a few discrete morphological char- acteristics may not provide an accurate indication of the genetic divergence among the cultivated genotypes of sorghum. Since both natural and human selection have contributed to genetic differentiation in sorghum (Murty et al., 1967), landraces of the same race grown in different habitats may have greater genetic dissimilarity than those of different races from the same habitat. Biochemical and molecular markers are now widely used as tools to assess the validity of taxonomic classifi- cation in crop plants. Allozyme markers have been used extensively to evaluate the extent and patterns of genetic variation in sorghum (Aldrich et al., 1992; Morden et al., 1989, 1990). Allozymes did not clearly separate the various races of cultivated and wild sorghum. Instead, these markers showed some degree of differentiation related to geographic area of origin. Because allozymes only measure variation at a very limited number of sites, these results may not reflect overall patterns of genetic variation throughout the genome (Aldrich et al., 1992). Restriction fragment length polymorphism (RFLP) and RAPD markers can overcome the limitation of allozymes because they have the potential to identify a large number of polymorphisms with good coverage of the entire ge- nome (Melchinger, 1993). These techniques have been used to characterize genetic diversity and phylogenetic relationships in sorghum (Aldrich and Doebley, 1992; Cui et al., 1995; Deu et al., 1994; Tao et al., 1993). However, these studies provided conflicting results con- cerning the degree of differentiation among cultivated races of sorghum. These studies also assessed genetic diversity either in a relatively small number of accessions (Aldrich and Doebley, 1992; Cui et al., 1995; Tao et al., 1993) or in non-random accessions selected on the basis of a priori information from other studies (Deu et al. 1994). Thus, extensive random sampling from the Abbreviations: RAPD, random amplified polymorphic DNA; RFLP, restriction fragment length polymorphism. Published March, 1997