PubAg

Main content area

Assessment of Stalk Properties to Predict Juice Yield in Sorghum

Author:
Carvalho, Geraldo, Jr, Rooney, William L.
Source:
BioEnergy research 2017 v.10 no.3 pp. 657-670
ISSN:
1939-1234
Subject:
Sorghum (Poaceae), agronomic traits, biomass, breeding programs, economic sustainability, equations, ethanol, ethanol production, feedstocks, ideotypes, juices, models, phenotype, photosensitivity, prediction, sweet sorghum
Abstract:
Sweet sorghum is an outstanding feedstock choice for bioethanol production, but the gap between theoretical and commercial ethanol yields must be reduced to improve economic viability. Extractable juice yield is a primary limiting factor for higher ethanol yield, but current phenotyping techniques to measure juice yield in sorghum can be laborious. Therefore, alternative approaches to measuring juice yield during selection are needed. The objectives of this study were to investigate the relationship between stalk-related traits and juice yield and to assess the ability to predict juice yield using agronomic traits and stalk properties across and within a diverse set of sorghum ideotypes (photoinsensitive, photosensitive, biomass, grain, and sweet types). Stalk weight, stalk volume, stalk diameter, and plant height had significantly strong associations with juice yield, which were consistent across different sorghum ideotypes. The direct and indirect effects of multiple predictive traits on juice yield varied greatly with the distinct sorghum subsets. However, equation modeling demonstrated that juice yield is satisfactorily predicted by jointly assessing stalk weight and stalk moisture. Moreover, alternative prediction models involving distinct combinations of agronomic and stalk-related traits had similarly good prediction accuracy. Altogether, this suggests that several prediction models can be used to accelerate phenotyping for juice yield, which will improve the selection process. Overall, the results indicate that increasing sorghum juice yield via indirect selection is possible, but the choice of prediction model depends on the ideotypes and resources available in a breeding program.
Agid:
5763417