UC Berkeley

Most cultures obtain the majority of their caloric and mineral nutrition from cereals either directly or through livestock; therefore, understanding the genetic architecture underlying the acquisition of ions from the soil in these crops is important for agricultural purposes. The ionome of an organism is intricately interconnected; many transporters and enzymes involved in ion uptake and storage have been implicated in the homeostasis of multiple elements. Thus, an elemental signature of iron deficiency can be detected in an organism through analysis of the entire ionome. The focus of this thesis is on the ionome of two closely related species of grass known as Setaria viridis and S. italica.

This thesis uses standing diversity and constructed inbred lines to illustrate the ionomic variation present in the Setaria species complex, and to determine which regions of the genome are associated with this variation. Chapter 1 explores the effect of iron concentration upon the ionome of a diverse group of Setaria landraces and cultivars. Each accession was grown in the soil as well as a hydroponic medium containing three varying concentrations of iron. This experiment aimed to assess the variation in iron accumulation present in the species. Plants were grown until senescence and then harvested and dried. Dry root and shoot weight was assessed, as was yield by weight. Additionally, the uppermost leaf was taken and the concentrations of 20 different elements in the plant were assessed. The unsupervised machine learning algorithm DBSCAN allowed for the identification of two separate ionomically defined groups and three morphologically defined groups. These groupings did not show strong intercorrelation, with the exception of one ionomic group, which did correspond to the African morphology group. The second ionomic group appeared to show a constitutive phosphate deficiency response.

Chapter 2 addresses the need to statistically validate the quantitative trait loci (QTL) identified in association studies such as those completed in Chapter 3 and offers a mathematical solution to this issue along with an open source R module. The Scanning Probabilistic QTL Validator calculates the significance of a QTL of a particular length overlaying any N genes that were previously identified for the trait of interest. The strategy put forth in this chapter takes into account the number of QTL that are identified for a particular model, their lengths, the number of markers used for the initial mapping, and the overall gene distribution in the organism of interest. The SPQV’s usage is demonstrated by validating the results of a QTL mapping experiment in the TeoNAM RIL population that was aimed at identifying genes associated with branching.

Chapter 3 details the use of the quantitative trait locus (QTL) mapping to identify regions of the Setaria genome associated with ionic homeostasis in the species complex. A RIL population resulting from the interspecific cross between S. viridis and S. italica was grown in treatments designed to assay the effects of planting density and water availability. As in Chapter 1, the flag leaves of these plants were harvested and subjected to ICP-MS to assay their ionomic content. These phenotypic data were then used to map the regions of the genome that are associated with alterations in the ionome of the species complex. Further mapping was performed using the rotated loadings of principal components analyses which were performed on the phenotypic data resulting from the initial grow-outs. A total of 251 QTL were identified. Multiple concentrated regions of QTL were identified that overlapped with regions previously identified as important for the trait of water use efficiency.