In order to get an overview on the genetic relatedness of sorghum (Sorghum bicolor) landraces and cultivars grown in low-input conditions of small-scale farming systems, 46 sorghum accessions derived from Southern Africa were evaluated on the basis of amplified fragment length polymorphism (AFLPs), random amplified polymorphic DNAs (RAPDs) and simple sequence repeats (SSRs). By this approach all sorghum accessions were uniquely fingerprinted by all marker systems. Mean genetic similarity was estimated at 0.88 based on RAPDs, 0.85 using AFLPs and 0.31 based on SSRs. In addition to this, genetic distance based on SSR data was estimated at 57 according to a stepwise mutation model (Deltamu-SSR). All UPGMA-clusters showed a good fit to the similarity estimates (AFLPs: r = 0.92; RAPDs: r = 0.88; SSRs: r = 0.87; Deltamu-SSRs: r = 0.85). By UPGMA-clustering two main clusters were built on all marker systems comprising landraces on the one hand and newly developed varieties on the other hand. Further sub-groupings were not unequivocal. Genetic diversity (H, DI) was estimated on a similar level within landraces and breeding varieties. Comparing the three approaches to each other, RAPD and AFLP similarity indices were highly correlated (r = 0.81), while the Spearman's rank correlation coefficient between SSRs and AFLPs was r = 0.57 and r = 0.51 between RAPDs and SSRs. Applying a stepwise mutation model on the SSR data resulted in an intermediate correlation coefficient between Deltamu-SSRs and AFLPs (r = 0.66) and RAPDs ( r = 0.67), respectively, while SSRs and Deltamu-SSRs showed a lower correlation coefficient (r = 0.52). The highest bootstrap probabilities were found using AFLPs (56% on average) while SSR, Deltamu-SSR and RAPD-based similarity estimates had low mean bootstrap probabilities (24%, 27%, 30%, respectively). The coefficient of variation (CV) of the estimated genetic similarity decreased with an increasing number of bands and was lowest using AFLPs.

In order to define the variation of the genomic proportion of the recurrent parent [G(RP)] and its relation to yield, G(RP) of individual BC 1 plants of two sorghum populations composed of a high-yielding cultivar as recurrent parent (RP) and a donor with superior drought resistance or grain quality, respectively, was estimated using AFLPs and SSRs. G(RP) in BC 1 ranged from 0.53 to 0.95 and averaged to 0.76 in the population (NP4453 · ÔSV-2Õ) · ÔSV-2Õ. G(RP) varied between 0.60 and 0.86 and averaged to 0.74 in the BC 1 of (ICV-219 · ÔSV-2Õ) · ÔSV-2Õ. Results show that plants with a G(RP) equivalent to BC 2 (0.875) or BC 3 (0.938), respectively, can be selected from BC 1 . Yield performance of BC 1 S 1 families was tested in field trials carried out in South Africa. The correlation between yield and G(RP) in BC 1 was low. Selection according to G(RP) did not result in an effective preselection for yield.

Retraction type (multiple responses allowed):  Unreliable findings  Lab error  Inconsistent data  Analytical error  Biased interpretation X Other: One of the coauthors informed us that none of the coauthors was aware about the submission to our journal. The article was therefore retracted and the files were removed from our system  Irreproducible results  Failure to disclose a major competing interest likely to influence interpretations or recommendations  Unethical research  Fraud  Data fabrication  Fake publication  Other:  Plagiarism  Self plagiarism  Overlap  Redundant publication *  Copyright infringement  Other legal concern:  Editorial reasons  Handling error  Unreliable review(s)  Decision error  Other:  Other: Results of publication (only one response allowed): X are still valid.  were found to be overall invalid.

No abstract