Cultivars Sonora, Foisy, and Chiddam Blanc de Mars were selected. The primary criterion was the total root biomass with Sonora ranked among the highest, Foisy was intermediate, and CBdeM had a low total root biomass. Additionally, the three also have other contrasting phenotypes for traits such as drought tolerance, plant height, days to heading, awn type, and seminal root angle. These parents are what could be considered as “traditional cultivars” in that they were all selected from land races. Sonora was selected from a land race in Durango, Mexico and is known for good drought tolerance but its height makes it susceptible to lodging. Cv. Foisy was selected by Mr. Foisy in Oregon in 1865 and typically yields more than CBdeM and Sonora. CBdeM originates from Ville de Paris, France, and was selected from an English land race. None of these cultivars have a place in commercial agriculture today but still are grown by traditional or artisan farmers as so called heirloom varieties of wheat for bread making. More information about the parents can be found on the UC Davis small grains web page in the 2011 California cultivar descriptions publication . Crosses were made in a triangular manner to form a set of “nested” mapping populations with any given two populations having a single parent in common so that we get the populations Sonora x CBdeM , Sonora x Foisy , and CBdeM x Foisy.
This design provides a built in system for verification of QTLs across populations and genetic backgrounds. For each of the three populations ca. 200 lines were planted on July 11th, 2013 in an air-conditioned greenhouse on the UC campus in Riverside, California,growing lettuce hydroponically in one gallon pots with two plants per pot. These were used for seed increase, leaf tissue for DNA extraction, and for phenotyping of simple traits. Doubled haploid plants in each pair were compared and expected to be identical, however, not all were and any lines with clear differences between the two plants were discarded. A second seed increase was planted on April 4th, 2014 in a similar manner and was also used to collect phenotype data. During this increase plants were grown under 18 hour days with supplemental lighting. In 2015 a two-location field trial was established. Experiments were planted in October 2015 and harvested by May 2015. The two locations were the University of California, Riverside Agricultural Experiment Station in Riverside, California, and at the Coachella Valley Agricultural Research Station . Experiments were set up in randomized augmented designs with three check varieties replicated in each block. The check varieties were Blanca Grande 515, Summit 515, and Cal Rojo.
There were 32 blocks per treatment per location with 16 plots per block. Additional “blank” plots of Summit 515 were planted to make blocks square but were not included in the analysis. Each plot consisted of six rows spaced 20cm apart and 122cm long planted at a density of 560 seeds per plot. Each location had two treatments, one well-irrigated and the other which received limited irrigation after 60% of the genotypes were booting. The well-irrigated treatment received water as needed based upon soil moisture and plant indications. For the limited irrigation treatment water was withheld until plots showed moderate to severe wilting at which point they were irrigated to prevent death.The R statistical package “lmerTest” was used to obtain the predicted mean values for all traits evaluated which included; days to heading, plant height, yield/m2 , and 1000 grain weight . SNP calls from Genome Studio were converted into “A” and “B” genotypes by comparison against parental scores for each population. Markers that were polymorphic between parents for each population were imported to JoinMap 4.1 and used to construct linkage maps. Chromosome and marker index number were used to name markers previously mapped by Wang et al. , for example 5A_6716, and markers that were not previously mapped were named using an underscore and the marker index number with no chromosome indication. Identical individuals were excluded from the genotypes used to construct linkage maps and likely arose as artifacts from the DH procedure or were a result of labeling errors.
Also, individuals with greater than 10% missing data for marker calls were excluded. Initially linkage groups were generated based upon markers mapped by Wang et al. using eight doubled-haploid mapping populations. This was done on a chromosome by chromosome basis including only markers mapped to a given chromosome. Identical markers for the given chromosome were removed prior to mapping. Groupings were made using the default calculation settings for independence LOD and linkage groups were mapped using the default settings for the maximum likelihood algorithm. For linkage groups that failed to generate maps or lacked a sufficient number of markers, additional markers were added from the unmapped pool of SNPs. To do this, all unmapped markers were selected along with the mapped markers for the given chromosome and then the steps listed above were repeated to give new linkage groups with a more suitable number of markers. These newly added markers were then BLASTed against the wheat arm survey sequence to verify their correct linkage group assignment. Phenotypic data for awn type, days to heading, and plant height collected during 2013, 2014 greenhouse evaluations and 2015 field evaluations were used to map QTLs by the software package ICImapping . For greenhouse data, the linkage maps and mean value for two plants of each doubled haploid line were used to map QTLs and for field data the predicted mean values for each genotype were used. The composite interval mapping method with a step of 1 cM was used and the threshold for QTLs detection was determined using 1000 permutations at α = 0.05. Originally, populations SC, SF, and CF consisted of 257, 244, and 214 lines, respectively. During the 2013 seed increase populations were assessed for vernalization requirement, hybrid necrosis,ebb and flow table and uniformity. Of the 200 lines planted for SC, SF, and CF population about 1, 2, and 7%, respectively, showed what appeared to be segregation but could have arisen from multiple unknown reasons. Winter growth habit appeared in 18.5, 7.5, and 1.5% of the SF, CF and SC populations, respectively. This was despite the fact that all three parents are spring wheats and require no vernalization. The appearance of winter growth habit may reflect some combination of recessive alleles of vernalization genes . Hybrid necrosis was rated on a scale of 1 to 10 with 1 being minor and 10 being lethal. Hybrid necrosis was fairly prevalent in SF with 31% of the genotypes showing some level of the phenotype; the CF population had 25% of the genotypes showing some level of hybrid necrosis and the SC population only 6.5%. Any genotype with a score greater than 3 was excluded from genotyping. Because of possible contamination , sterility and with some lines showing winter growth habit and/or hybrid necrosis, populations were reduced in size. Additionally, population sizes had to be limited to genotyping of 150 lines for several reasons, from the cost/practicality issue to future experiment manageability. However, all non-genotyped lines are preserved and can be accessed, if needed. During the 2014 greenhouse evaluations the remaining 150 lines for each population were characterized for vernalization requirement, days to heading, plant height, and awn type . For field trials populations were reduced to 133, 121, and 115 lines for SC, SF, and CF respectively due to winter habit or late flowering of some lines and lack of seed for others.
During the 2015 field evaluations populations were characterized for days to heading and plant height, awn type and 1000 grain weight . However, not all field data could be analyzed and/or were unreliable, thus distributions of trait values in the field for days to heading, plant height at Coachella Valley are shown in Figure 1.3 for SC, SF, and CF.Using phenotypic data collected during the greenhouse evaluations and the field evaluations QTLs were mapped to verify the quality of the genetic maps and provide some basic genetic information about the three populations. Table 1.1 summarizes regions that were mapped in the SC, SF, and CF populations for awn type, days to heading , plant height , and hybrid necrosis . Only the regions consistent through multiple years are discussed. Table 1.2 demonstrates how QTLs can be verified within and between populations by using awn type QTLs as an example. For simplicity of the exercise lines were classified into two groups: awned and awnless. Two genome regions were consistently identified that explained 23.76 – 92.67% of the phenotypic variation within the populations. The first locus was consistently identified in the SF and CF populations on chromosome arm 5AL. The QTL in the SF population covered a 2.2 cM region between the markers 5A_9620 and 5A_6716 with the peak around 266 cM explaining 36-39% of the phenotypic variation across all years and environments. In the CF population, the QTL covers a 3.2 cM region between 5A_9620 and 5A_6716 with the peak around 169 cM. The QTL explains 31-93% of the phenotypic variation observed in the population across all years and environments. This QTL shares the same two markers in common with the QTL identified in the SF population. Additionally, Mackay et al. mapped the same QTL using a wheat MAGIC population and verified it using an association mapping population. They identified the marker BobWhite_c8266_227 as being the closest linked to the QTL which in these populations mapped to the same genetic location as 5A_6716 identified here. The second QTL was consistently identified in the SC and SF populations as being on the chromosome arm 6BL. In the SC population the QTL covers a 4.2 cM region between markers 6B_606 and 6B_1614 with the peak around 103 cM. It explains 62-73% of the phenotypic variation for this population across all three years and two environments. In the SF population the QTL covered a 0.71 cM region in 2013 and 2014, and a 2.1cM region in 2015 with the peak being around 79 and 80 cM respectively. This QTL explains 23-29% of the phenotypic variation in the population. In 2013 and 2014 the QTL was between 6B_45514 and 6B_606, however, in 2015 it shifted by a couplemarkers to 6B_68633 and 6B_84 covered a larger region. However, its peak was still near the same point and the two markers associated with the phenotype in 2013 and 2014 were present in the 2015 region. The QTL shares the 6B_606 marker in common with that identified in the SC population . These results indicate that Sonora carries the dominant allele for B2 on 6BL and that Foisy has the dominant allele for B1 on chromosome arm 5AL. Since CBdeM is fully awned it must have the hd b1 b2 genotype. The trait „heading date‟ or „days to heading” in wheat is determined by several factors, including vernalization requirement controlled by the Vrn genes , the photoperiod genes play a role in determining the sensitivity to photoperiodism and the Earliness per segenes are responsible for controlling flowering time regardless of photoperiod. In the three populations studied here, five major QTLs were found responsible for the heading date character, located on chromosomes 2D, 5A, 5B, and 5D. Two consistent QTLs on chromosome 2D were identified in the SC and CF populations in 2015. In the SC population the QTL covers a 0.67 cM region with its peak around 112 cM between markers 2Dx_32130 and 2Dx_79444. This QTL explains 18.43% of the phenotypic variation seen in the population and has an average additive effect of 7.69 days. In the CF population the QTL covered a 5.8 cM region with a peak around 47 cM between markers 2Dx_7001 and 2Dx_13208. This QTL explains 70.06% of the phenotypic variation in this population and has an average additive effect of -14.24 days. These QTLs are most likely the Ppd-D1 gene described by Beales et al. . Sonora contributed the day length sensitivity allele in the SC population and Foisy contributed the allele in the CF population. This explains why no segregation for the locus was seen in the SF population.