Seedlings were grown for two weeks in a greenhouse with temperatures ranging between 22-35o C following a 12-hour lighting regiment emitted by high-pressure sodium light bulbs. Fertilizer, Nu-Crop , was incorporated in the irrigation water. After two weeks, seedlings were moved outside for three days to harden them to field conditions before transplanting. Seedlings were transplanted into the field at a depth of 5cm, covering the cotyledons. We grew three trials, designated as “environments,” planted in July 2020, May 2021, and July 2021 at the UC Davis Plant Sciences Farm in Davis, CA. The 2020 trial was designed as a randomized complete block design with two replications. The trial contained 132 entries, with three entries acting as controls, two of which were day neutral and early flowering and one daylength sensitive and late flowering . Each entry was represented by five plants in a plot. Some entries were included as additional replications within blocks. Plots were transplanted on July 6, 2020, at the UC Davis Plant Sciences Farm in Davis CA. Seedlings were transplanted into the middle of raised soil bed, with beds spaced 1.5m on center. Each plot was 4.5 m long within a bed with five plants per plot spaced 90 cm apart, cannabis drying rack with no additional end to-end space between plots. The trial plots were bordered by two rows of hemp plants on either side of the plot area and maize was planted as border surrounding the entire perimeter of the field.
The May 2021 trial was designed as a randomized complete block design with four replicates with 64 plots per replication. The trial contained 88 entries because some entries were present in fewer than four replications. Entries included 14 cloned genotypes which were propagated from plants selected from the 2020 field, 33 families previously grown in the 2020 trial, 38 new families, and the same three check cultivars. Seedlings were transplanted on May 12, 2021 at the UC Davis Plant Sciences farm. The trial was planted similarly to the June 2020 trial. Two rows of hemp plants were transplanted to act as borders and sorghum surrounded the perimeter of the field. The July 2021 trial was designed as a randomized complete block design with four replications each with 63 plots. This trial contained 69 entries, with the same three control entries, 34 previously grown in the May 2021 field, and 32 new families. Seedlings were transplanted on July 7, 2021 and each plot contained six plants grown in two rows 90 cm apart on 1.5m beds. Plants were spaced 120cm apart within rows and offset 60cm between rows. No additional space was added between plots on a bed. The 2020 trial site was cover cropped with wheat in the fall. Prior to transplanting, field crews added a total of 168 kg/ha using a mixture of Starter 8-24-6, UAN-32, and CAN-17. Irrigation was applied for 6 hours daily Monday through Friday providing approximately 4 cm of water a week to the field.
Coragen was read ministered a month into the trial due to continued presence of corn ear worm . Before planting in May 2021, 28lbs N/acre was applied to the field. An additional 10lbs of N/acre using the same starter fertilizer, Nu-Crop , was applied 2 weeks after transplanting. UAN-32 fertilizer was applied at 28 kg/ha along with Coragen insecticide through the irrigation on June 30 and July 30, 2021. For the July 2021 field, fertilizer was applied at transplant at a rate of 200 lbs N/acre using a mixture of Starter 8-24-6, UAN-32, and CAN-17. Two weeks after transplanting, 10 lbs of N/acre using the same starter fertilizer was applied 2 weeks after transplanting; the irrigation scheduling was the same as the previous year. For all fields, seedlings were watered using a surface-drip system for a month before being switched to a sub-surface drip system to minimize weeds.Leaf samples were collected from 5000 individuals from all trials one month after the transplant date of that field. We removed a single leaflet from the top three inches of each plant and placed it in a 96 well plate, which was subsequently sealed and frozen in a -20o C freezer for at least one week before DNA was extracted. The extraction protocol follows that of Clark et al. . Following DNA extraction, quality and concentration were checked and assessed using a nanodrop spectrophotometer.
Reduced representation libraries were created for genotyping-bysequencing following the protocol of Elshire et al. , with the exception that a HindIII restriction enzyme was used. 1500 individuals from families which were segregating for day neutrality were used in our GBS. Libraries were size selected and sequenced on an ABI Prism® 3730 Genetic Analyzer at the UC Davis Sequencing Facility.A genome wide association study was first done using a general linear model in TASSEL 5.0 for both AGDD and Day Length at Flower. For both traits, the QQ-plot , showed a large inflation of p-values, suggesting that many of our SNP markers could be considered as false-positive associations with the traits. Given that our GWAS population contained well defined families of known relatedness, we expected that including kinship in the model would be necessary. Therefore, the GWAS was re-run using a mixed linear model , which incorporated the population structure using a Q matrix and integrated the relatedness of individuals with a kinship matrix. This deflated the magnitude of p-values and reduced the presence of false-positive results . Fifty-four combined SNPs between AGDD and Daylength at Flower were marked as hits by having a p-value above the significance threshold set by a Bonferroni correction, 6.007: AGDD and Daylength at flower, both of which had the same top 5 SNP hits, were located on Chromosome 2 in the hemp genome . The GWAS shows SNPs associated with both traits in several regions on Chromosome 2, possibly because of the manner in which our families were constructed and the large linkage disequilibrium that they contained.We expected families to differ for flowering time because they included both day neutral and daylength sensitive phenotypes. We also observed a significant family × environment interaction, suggesting that flowering differed across trials. This is likely to be related to the earlier planted May 2021 trial compared to the two July trials, given the differences in photoperiod. The two traits measured in our experiment had a strong and significant inverse correlation as expected. Due to hemp plants typically being short day plants which need to reach a critical photoperiod to initiate flowering, plants which need very short days would accumulate more AGDD; due to our planting dates in May to June, days only begin to shorten past the summer solstice, around June 21st, and only then, and sometimes even a few weeks after that, do hemp plants initiate flowering.After categorizing individuals in families within two groups, early flowering and late flowering , we saw no significant differences in the proportion of segregants in each family for all except for one, KNR2C. Regarding KNR2C, the family had a larger proportion of earlier flowering individuals in May 2021 than July 2021, where the proportion of early to late flowering individuals was reversed. There was significant G×E interactions when observing the dataset, and in some capacity, there may be a way in which the individuals within KNR2C are responding to changes in the environment; specifically, because of differences which arise in temperature and daylength between fields planted weeks apart. The May 2021 trial had an uneven distribution of nitrogen, with plots at the southern end of our field, showing N deficiency and possibly inducing a larger proportion of KNR2C individuals to flower earlier compared to the same family planted in the July trial. The KNR2C family had few individuals and was not included in all replications, vertical grow system and hence the difference between fields could have been due to sampling. In addition, with any field trial, the individuals transplanted into the field could have been mislabeled or placed incorrectly. With both traits, across all fields, broad sense heritability ranged between 0.60-0.82 .
This falls just out of line with similar conclusions from Petit et al., 2020 which found broad sense heritability of flowering time in hemp to be 0.94 – 0.95 for beginning and full flowering, respectively. Broad sense heritability of a specific trait “is the proportion of phenotypic variation attributable to genetics” . The heritability of both traits was markedly lower in the July 2020 trial in which we had two replications vs. four for the 2021 trials. The moderate heritability of both traits in 2020 indicates that while variation for flowering time has a large genetic component, it can still be influenced by the environment, particularly if the experimental design is not robust. Across all three environments, heritability of AGDD was 0.55 and of Daylength at Flower, 0.44. Our parental germplasm was quite diverse, including different sources of high cannabinoid material as well as grain types with desirable architecture. Therefore, we controlled for both population structure and kinship in our association study. We identified numerous SNP located on Chromosome 2 for both traits. The QQ-plot for our model, while better than a general linear model that excluded population structure and kinship, still suggested that some of these associations could be false positives, as seen for SNPs trending above the line between the expected and actual statistical significance . Using mixed linear models to find associations between traits and SNPs have been used extensively, not just in plants in general, but when looking at flowering traits specifically. Papers identifying SNPs for flowering timing have been found for hemp , alfalfa , and cowpea . When a GWAS was run with all trials, we observed that the five most significant SNPs were shared between the two phenotypic traits . We looked at the C. sativa L. variety ‘Purple Kush’ on the GenBank website and tried to identify if there were any candidate genes which might have been within, or just downstream of our significant hits. We investigated a 31 Mb region of Chromosome 2 between our two most and distal SNPs in order to see if we could identify genes which were linked to day-neutrality much like CONSTANS . What we have found was one gene, flowering locus K homology domain , which was located on position 28,634,491 of chromosome 2, 7Mb away from two significant hits shared between both traits. In Arabidopsis, the flowering locus K is a regulator of other flowering time genes. Arabidopsis plants with mutations to FLK were shown to effect relative expression levels of CONSTANS in the individual plants . With this, it might be possible that FLK works in the same way protein DAY NEUTRAL FLOWERING acts in Arabidopsis. DNF works in the same pathway as CO and works to maintain low levels of CO in short day settings . DNF when mutated, shows to initiate earlier flowering in individuals over expressing the protein, and also in instances in mutant dnf individuals.Lignocellulosic biomass can be a renewable carbon resource for the production of fuels and chemicals to replace non-renewable fossil carbon sources such as natural gas, oil, and coal. Lignocellulosic biomass can be converted into a wide range of chemicals and energy carriers through a closed-loop bio-refinery. Despite decades of scientific and technological advances in this field, many technical challenges associated with the recalcitrance of lignocellulose to deconstruction into fermentable carbon still hamper the establishment of economically feasible bio-refineries. These include the lack of efficient and inexpensive processes to depolymerize all components of biomass and convert them to fuels and chemicals at high yields by microbial fermentation. To improve the accessibility of the biomass carbohydrate polymers to hydrolytic enzymes, a pretreatment step is frequently required. Numerous pretreatment technologies have been developed, including physical methods, chemical methods, combined physico-chemical methods, or biological methods. The discovery and implementation of biocompatible ionic liquids such as cholinium lysinate provide a compelling alternative to the aforementioned approaches because of their high pretreatment efficiency and lower inhibitory effect to enzymes and microbes. These features have allowed the development of a one-pot reaction system where all the process steps, such as pretreatment, enzymatic hydrolysis and fermentation can be consolidated into a single vessel without any separation, potentially reducing operating costs.