The Relationship of Physiological, Morphological and Metabolic Traits to Biomass Accumulation and Drought Response in Sorghum.

Sorghum bicolor is a multipurpose crop used in food, forage and biofuels markets. Sorghum is genetically diverse and displays phenotypic variation for biomass yield. This suggests possibility to breed lines with enhanced biomass, while also improving traits for other crop uses. While sorghum is considered relatively drought-tolerant, genetic variation for this trait may be used to improve performance in water-limited environments. Thus, there exists a need to better understand interactions among morphological, physiological and metabolic mechanisms which influence biomass yield. This study examined trait variation to uncover potential targets for breeding to improve sorghum for multiple uses in both benign and stressful environments. We evaluated 11 lines of varying biomass yield and resistance to drought for morphological and physiological traits. Non-targeted metabolite profiling by gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS) was performed and cellulosic components of partitioned plant tissues were evaluated via Near Infrared Spectroscopy (NIRS). Significant variation was observed for all measured morphological, physiological and global metabolic profiles across lines. In general, small biomass plants shared other morphological traits and had higher physiological rates, while large plants had lower physiological rates. Interestingly, although smaller lines were types bred for grain and larger types were those bred for biomass and/or forage yields, high biomass plants did not produce less seed weight indicating a possibility to breed for improved dual purpose crops. Cell-wall compositional constituents had few correlations with other phenotypic traits on a percentage basis. However, conversion of percent lignin and xylan to total weight per plant showed strong positive correlations with plant size (biomass, stem harvest index and height). A model to predict final biomass from metabolite data was developed using elastic net regression. These results provide putative phenotypic and molecular markers to facilitate breeding for enhanced biomass through the modification of primary and secondary metabolism.