Evaluation for drought and disease resistance in sorghum for use in molecular marker-assisted selection.
Reliable screening of traits is essential to the development of molecular markers for use in sorghum improvement programmes. Marker-assisted selection should be particularly useful for the traits which are the most difficult to evaluate. These traits, however, will be the most difficult to develop molecular markers for, because of germplasm classification problems. Drought resistance is a complex trait in sorghum. Water deficit can affect the plant at any stage of growth; it can vary in severity and duration and its effect is often compounded by other abiotic or biotic stresses. Sorghum genotypes respond differently to soil moisture stress depending upon the stage of growth at which the stress occurs. The research reviewed here focused on moisture stress in the GS2 (head differentiation to flowering) and GS3 (flowering to physiological maturity) stages. Some genotypes showed excellent tolerance to moisture stress during one stage, but were very susceptible at the other stage. Field evaluation in the dry environment of western Texas, USA proved effective in evaluating sorghum for pre- (GS2) and post-flowering (GS3) drought resistance. This evaluation is described and distinct visual traits indicating resistance or susceptibility at each stage are subjectively rated. Pre- and post-flowering drought-tolerant genotypes have been developed, and their reaction is repeatable across widely diverse locations and years. Anthracnose (Colletotrichum graminicola), grain mould (Fusarium spp.), Striga and downy mildew (Peronosclerospora sorghi) are important diseases for which marker-assisted selection should be useful and their evaluation is discussed. The development of the appropriate type of germplasm is critical to the evaluation process. Individual F2 plants can be utilised when evaluation is very precise. For complex traits that are difficult to evaluate, such as drought, advanced generation progeny such as F6 recombinant inbred lines will be most useful, since large quantities of genetically uniform seed can be generated for use over multiple locations and/or years, and will produce phenotypically uniform progeny. Near-isolines differing in the trait of importance can be useful in studies of specific genes.