To use the SPQV, all QTL with a sufficiently high logarithm of the odds score should be assessed. The function provides the higher and lower confidence limits, as well as the combined CI for each mathematically related set of QTL. If any QTL within a set exceeds the combined CI for that set, then the whole set can be deemed successful. QTL with observed gene counts that exceed the upper confidence limits should of course be considered to have attained significance. If at least one QTL in an experiment is determined to be significant by the SPQV, the mapping experiment on a whole can be deemed a success. The QTL that contain lower numbers of genes might be identifying previously unknown genes, and therefore their non-significance does not detract from the significance of other QTL. Because of the potential for the identification of new phenotype associated genes in QTL mapping, QTL containing significantly low numbers of identified genes might also be of interest, as they may have been placed on ’empty’ regions due to previously unidentified genes that have a large impact on the phenotype. The QTL containing significantly low numbers of genes should only be considered interesting when at least one other QTL in the same mapping experiment has been proven to be significant so as to avoid the suggestion that all QTL in a faulty mapping study were significant. The intake, transportation, and storage of elements within a plant is vital to its appropriate growth and development. Plants must maintain balance between excessive uptake, which can cause toxicity and necrosis through the production of free radicals or the exclusion of other nutrients, and sufficient uptake of essential inorganic nutrients . Tight control of ion homeostasis allows plants to respond appropriately to environmental conditions such as temperature, soil pH, and water availability . This regulatory control must be responsive to both the concentration of ions within the soil and to those concentrations within the plant. Understanding the genetic control of elemental uptake and transportation will contribute to efforts to improve crops,vertical farming tower for sale and to the ultimate sustenance of the growing global population.
Examining the concentration of elements within a plant sample allows for exploration of the genetic and physiological processes involved in adaptation to a particular environment. To this end, we and others have developed a pipeline that is designed to cheaply and efficiently measure the concentrations of 20 different elements via inductively coupled mass spectrometry . This process is known as ionomics, which is defined as the quantitative study of the mineral nutrient and trace elemental content of an organism; that is to say, its ionome . In cereal crops such as Setaria italica, the flag leaf emerges just before the panicle, and therefore marks the specific developmental time point at which the plant has taken up the majority of its total mass. This tissue is therefore ideal for ionomic inquiry. Additionally, nutrient loading of the grain is accomplished through the remobilization of organic and inorganic materials from the leaves . The flag leaf specifically is instrumental in the loading of photo assimilates and other micro-nutrients , and is therefore commonly used as representative tissue to assess the composition of a grass . The species Setaria viridis, or green foxtail millet, is a member of the Panicoideae that utilizes C4 photosynthesis . It is therefore a good model system for several related, economically important crops, including sorghum and maize. The compact stature, short life span, and sequenced genome of S. viridis have also contributed to its status as an emerging model organism . In addition, S. viridis is the wild ancestor of the crop species Setaria italica ; these two species have a semi-permeable boundary between them, as their primary difference is phenotypic and they are still readily crossed . Foxtail millet is a member of the small millet species, a group of ancient grains that are relatively nutritionally dense when compared to rice and wheat and which are often cultivated as subsistence crops . Because of the combination of the high nutritional value of the S. italica grain and its resistance to abiotic and biotic stressors, breeding elite cultivars of S. italica for increased nutritional content is an attractive prospect. Understanding the contribution that different regions of the S. italica genome make to total nutrient content is an important first step for breeding purposes.
Structured populations are useful tools for dissecting the relationship between elemental accumulation and the genetic content of a species. Recombinant inbred lines have various advantages when it comes to quantitative trait loci analyses. Repetitive selfing allows for the break up of large linkage blocks, which in turn allows for finer mapping . Additionally, once established, RILs may be continuously maintained in a fixed homozygous state. This makes it possible to assay the same combinations of alleles in multiple different environments . The resulting phenotypic and genotypic data can then be compared through various statistical means in order to identify QTL. Here, we use elemental profiling of a RIL population resulting from a wide cross between S. italica and S. viridis grown in multiple environmental conditions to identify QTL associated with the ionic content of leaf material. Overall, we identified 251 QTL, 171 of which were associated with a single element and 80 of which were associated with a principal components analysis of the ionome. The use of traits defined by the differences between treatments in an experiment allowed for the quantification of the influence of the environment on Setaria’s ionome. Experiments were conducted in the summers of 2013 and 2014. Experiments assaying the effect of density of planting on ionic content were conducted in 2013 and 2014. A single drought experiment was conducted in 2014. A total of 189 F7 RILs resulting from a wide cross between the B100 cultivar of S. italica and the A10 line of S. viridis, together with their parent lines, were used as the study material . In every experiment, lines were planted in triplicate in a block design in the field in Creve Coeur, MO. Treatments in the density experiment consisted of either five centimeter spacing between neighboring plants, or twenty centimeter spacing between neighbors. Plants in the drought experiment were either well watered until the time of sample collection, or were subjected to drought stress from eight weeks post planting. The data used for this work included measurements for 20 different elements in flag leaf tissue collected from a recombinant inbred line population resulting from the cross of the B100 cultivar of Setaria italica and the A10 line of Setaria viridis. A drought experiment was conducted in Creve Coeur, Missouri in 2013 ; both a drought and a density experiment were conducted in the same location 2014 . 179 of the RILs were planted in at least two of the three experiments, while 116 were grown in all three.
The leaf samples from all experiments were treated in an identical fashion; samples were dried and stored in temperature and humidity controlled rooms before ionomic analysis. Each sample was profiled for the quantity of 20 elements using ICP–MS . The resultant measurements were normalized to the sample weight and technical sources of variation using a linear model. Experiment level analytical outliers were removed as in Davies and Gather 1993. Pursuant to this, the measured values for each element were transformed to normality using the Box Cox family of transformations, and Studentized deleted residuals were used to identify and eliminate further outliers within the measurements for each element. After outlier removal,hydroponic vertical farm phenotypes were derived by averaging the values for each line within an experiment and treatment. Both environment and genotype impacted the variation present in these data. Repeatability was generally lower than within experiment heritability , indicating that there was less variation in genotypic replicates within an individual experiment than across experiments. The broad sense heritability of 9 elements in the DN13 experiment, 14 elements in the DN14 experiment and 16 elements in the DR14 experiment exceeded 0.4. Certain elements including selenium, sulfur, and boron showed low repeatability; this is likely due to the fact that these elements tend towards analytical artifacts, as they accumulate to levels that are near the limits of detection of the methods described in this paper. The heritability of individual elements varied by up to 0.533 between different experiments. The function stepwiseqtl from the R package was used in order to identify a multiple QTL model for each of the elemental phenotypes. This function moves iteratively through the genome to test for significant allelic effects of each marker on the phenotype in question. When a significant locus has been identified, this is added to the model. A combination of forward and backward regression ultimately produces a genome wide QTL model for each trait. Each element was considered individually, as well as in combination with the others as a contributor to a principal components analysis that was run for each experiment . Five different metrics were used as the phenotype for each element in each experiment. These phenotypes include the ‘raw’ values for each treatment and the differential values for that trait .
The significance of a QTL was computed using the 95th percentile threshold resulting from 1000 iterations of the scanone function as a penalty for adding the QTL to the model. When all experiments are considered, a total of 251 QTL were identified . As expected from the heritability measurements, the majority of these were identified in the 2014 drought experiment . The 2013 and 2014 density experiments allowed for the identification of 75 and 71 QTL, respectively. Approximately a third of the QTL were identified within treatments; the remainder were identified using either the difference, relative difference, or ratio of the phenotypic values measured within the different treatments in a single experiment . Of the 251 QTL, 80 were identified for the mapping based on the principal components analysis; 39 of these resulted from the drought experiment, 21 from the 2013 density experiment, and the remaining 20 from the 2014 density experiment. Of the 251 QTL, 55 were located on chromosome 2 . The locations of the QTL were assessed for overlap with the locations of known ionic genes . Forty five of the QTL contained at least one gene within their 95% confidence intervals. Of the QTL that coincided with genes, 35 were identified using PC as the mapped trait. The QTL were assessed for significance using the Scanning Probabilistic QTL Validator. The QTL were divided into mathematically related sets based on experiment and the phenotypic metric that was used when they were mapped. The results of this assessment are reported in Table 3-3; each phenotypic metric that was used for mapping was associated with at least one set of QTL that identified a significant number of genes, indicating that the data curation was done effectively. There were several regions in which QTL were remarkably concentrated. Fifteen QTL were identified on chromosome 2 between 89.4 and 95.9 cM ; these QTL were discovered in both the 2013 density experiment and the 2014 drought experiment. The traits associated with these QTL included As, Al, Co, Cu, Mo, P, Rb, and Sr, as well as PC2 for the 2013 density experiment. In the context of the ionome, principal components analysis allows for the identification of regions of the genome that would not otherwise be found. While some of the PC QTL identified regions that overlapped with those identified by ion specific QTL, the majority of them, including many QTL identified for PCs which explained a large amount of the variance present in the data, did not. It is possible that for a single ion, the signal associated with the PC QTL regions is not sufficient for their identification, while the additive signal that is inherent in a principal component suffices. Moreover, many of the first few principal components overlay regions associated with water use efficiency , with concentrated regions of QTL identified on chromosomes 2, 5, 7, and 9 at positions 94, 111.9, 99.9, and 123.7, respectively.