2 The research challenge: Genetic improvement of sorghum and millet crop residues

Introduction

While crop residues, particularly cereal straws, provide the bulk of livestock feed across South Asia, their nutritive value is often so low that farmers must supplement them with feed grains and other concentrates. Improving the nutritional value of straws and the efficiency of their use in mixed diets is an important option for increasing livestock production in the region. A number of technologies have been developed to address this constraint. These include biological, physical and chemical treatments of these ligno-cellulosic materials. However, in general, adoption of these technologies has been poor (Singh et al 1997; Devendra et al 1998). Therefore, there is a need to develop alternative technologies for crop residue improvement that are more sustainable and responsive to smallholder farmers’ needs.

The goal of the research is to develop new germplasm, i.e. millet and sorghum varieties with residues of higher nutritive value, without sacrificing the gains in grain yields already achieved. This will increase the availability of nutrients to ruminants and thus increase the production of milk, meat and draft power.

Research constraints and opportunities

Genetic improvement of millet and sorghum crop residues is an area of research that is relatively new since there have been few, if any, previous attempts to remove digestibility constraints in cereal crops relevant to the semi-arid tropical region.

It is possible that genetic variation in the quality of sorghum and millet stover could be exploited to develop improved crop germplasm with stover of higher nutritive value. Small increases in roughage digestibility have been reported to result in considerable increases in milk and meat (ILRI 1995). In vitro dry-matter disappearance heritabilities for forages were found to range from 50% to 70%, and heritability estimates for fibre components and tannins were high for neutral-detergent fibre and tannin content of sorghum forage (Saini et al 1977). Such data suggest that digestibility of these species is under genetic control, indicating that improvements should be possible through breeding. Successful improvements in digestibility through selectively breeding forage grasses have encouraging implications for the probability of similar research success with millet and sorghum.

Stover versus grain yield

Parallel increases in grain and stover yield may be difficult to achieve in high production systems, but this is not the prevailing situation in the low-input agriculture of developing countries. Research on sorghum and millet conducted at ICRISAT and elsewhere (Badve et al 1994; Lynch et al 1995; Rattunde 1998) seems to indicate that, at the present levels of grain production and with the genetic diversity available, both stover and grain biomass can be genetically increased simultaneously. Increased biomass, however, must be digestible to contribute to livestock productivity increases. Hence the need for a collaborative approach involving both crop and animal scientists.

Stover quantity versus stover quality

Few quantitative data are available on yield and relative digestibility of components of forage grass varieties. An increased digestibility of 5% in pearl millet brown mid-rib (bmr) mutants was associated with a 23% reduction in dry-matter yield (Degenhart et al 1995), indicating a possible trade-off between gains in quality and reduction of stover biomass. However, reported increases in dry-matter digestibility of some forage grass species were weakly correlated with forage grass yield (Burton et al 1968; Lamb et al 1984; Ray et al 1996). This implies that improvement of quality could be possible without sacrificing biomass yield.  

Advances in maturity and accompanying morphological development increase the proportion of less digestible plant components and tissues. Different genotypes react differently to these changes (Fribourg et al 1976). Stay-green character in sorghum is one example where leaves and stem senesce more slowly and nutrient content differs from senescent genotypes (Thomas and Smart 1993).

Genotype versus genotype ´ environmental effects

Environmental factors could have significant interactions with genetic effects on productivity and quality traits as indicated by Arora et al (1975) for sorghum genotypes.   Large genotype ´ environment interactions would require more extensive testing and limit average gains. However, Lodhi (1993) observed a considerable amount of genetic variability and heterosis for forage quality and yield characters, indicating that substantial gains from genetic improvement are possible.

Research plan

Important questions to be answered in any ex ante impact assessment are: ‘How long will the research be likely to take?’, and ‘What exactly is involved?’ With this particular research challenge, there are two possible approaches: a conventional trait-based selection approach, and a ‘marker-assisted selection’ (MAS) route for the backcross transfer of previously identified stover quality traits to elite genetic backgrounds—forming near-isogenic pairs for use in confirming the utility of these traits and the map locations of quantitative trait loci (QTLs) controlling them.

In pearl millet and sorghum three seed-to-seed generations per year can be achieved at most using either conventional or MAS methods. It would be preferable to be conservative and plan to achieve only two. Once a trait is identified—by either conventional or MAS methods—backcrossing will be used to transfer it to elite genetic backgrounds. Segregating backcross progenies will be evaluated for the presence of the trait(s). Using conventional selection methods for the backcross transfer, single plants from F₂ generations of the first, third and fifth backcrosses would need to be evaluated at the very minimum. Using MAS, there will be a need to genotype individual plants in five segregating generations—first, second, third and fourth backcross F₁ and fourth backcross F₂ progenies—to identify those carrying the target markers and associated stover quality genes. By increasing the number of progenies and selecting simultaneously for target markers and elite background genotype markers in other regions of the genome, finished products could be identified somewhat earlier, but it is unlikely that this could be achieved earlier than the third backcross. The process to reach final products (elite genotypes in which genes for improved stover quality have been introduced and fixed in homozygous form) by the two routes is summarised in Table 2.1. The conventional method requires 10 generations (or 5 years), while MAS requires 7 generations (3½ years), to complete backcross transfer of the stover quality trait from its original source to one or more agronomically elite genetic backgrounds.

Thus MAS could shorten the research period required to generate genetically elite sorghum and pearl millet materials with improved stover quality. However, the identification of a specific marker cannot be guaranteed in a given period of time. Therefore, both routes will be pursued. This will also permit a direct comparison of the cost-effectiveness and practicality of conventional and MAS methods for stover quality traits in these two most important dual-purpose (grain and stover) crops of dryland agriculture in the SAT.