Written by Dr. Maany Ramanan, University of California Davis
As we come out of another successful barley season, I am excited to share findings from my doctoral research on malting barley, conducted under the guidance of Professors Glen Fox and Christine Diepenbrock at UC Davis. Our study investigated how different barley varieties, growing locations in California, and crop years (2017-2022) affect productivity, grain quality, and the composition of starch and protein. Supported by the American Malting Barley Association, this research focused on the combined effects of genotype (genetic makeup) and environment (growing conditions), aiming to innovatively use proteomics and machine learning to improve grain/malt quality predictions and operational efficiencies for farmers, maltsters, and brewers.
An overall theme of our findings was the substantial impact of the environment on barley quality in comparison to genotype. Our study found that environmental factors, such as soil type, geography, weather patterns, and farming practices, accounted for the largest variance (a measure of how much numerical values of a trait like total protein content differ from their mean across the entire study) in the majority of traits. Specifically, environment explained the largest variance in yield (44%), total protein content (72%), grain size (30%), starch gelatinization traits (34-74%), total starch (31%), and proportion of small to large starch granules in the endosperm (50%). On the other hand, genotype explained the largest variance in thousand kernel weight (33%).
Starch is a particularly vital component in barley that influences its malting and brewing quality, and our research revealed that environmental factors largely determine starch content and its structural properties. For example, the onset and end of starch gelatinization—critical temperatures for malting—are mostly influenced by the crop year. This variability means that even if two barley varieties have the same starch content, the quality of starch can differ significantly, impacting the malting process and the final brewing efficiency.
While these variables are interesting enough on their own, understanding the interactions between genotype and environment is particularly important for unlocking certain quality improvement outcomes. Interactions between genotype and environment can be assessed in addition to the main effects from genotype and environment using linear mixed-effect models. For instance, the interaction between barley variety and environment explained the largest variance in endosperm texture traits (45-87%). This highlighted the importance of maltsters selecting the right variety and keeping detailed records of yearly conditions for achieving optimum grain hardness to facilitate easy milling and starch conversion. For instance, if a variety is consistently meeting specifications in a certain location but is not meeting specifications in a different location, maltsters can determine which variety needs to be selected to be grown in a target location by keeping yearly records.
Our research also employed advanced proteomics, allowing us to compare differences in barley protein composition among over 3,000 unique proteins. We found that location explained the largest variance in the overall proteome, with 16 proteins showing significant differences in abundance compared to the overall mean. These proteins had storage, DNA/RNA binding, and enzymatic functionalities. Notably, storage protein, comprising 50-75% of the total protein content, was found to be more variable due to environmental factors than total protein content. This insight suggests that focusing on protein composition, rather than just total protein content, can provide a more accurate measure of barley’s malting potential when environmental variability is high.
Combining proteomics with machine learning was another approach we employed to predict barley quality traits. We used four different machine learning models which were trained on the entire proteomics data, and subsequently used to predict grain and malt quality traits. These models showed promising results in predicting traits such as total protein content, alcohol-soluble protein content, malt protein content, and malt fine extract. These results support the use of proteomics and machine learning to forecast barley and malt quality traits before harvest, and warrant further validation in other barley growing regions.
In summary, the intersection of environment, genotype, and technology offer an opportunity to track/forecast malt and brewing quality traits more effectively. While this research at UC Davis focused on California-grown barley, the varieties and conditions assessed are applicable to all barley regions in the United States. It would be worthwhile to validate these methodologies and findings in other barley growing regions in the future. Craft maltsters can enhance their practices by tracking quality on a yearly basis, looking beyond total protein content at storage protein content instead, and partnering with research institutions to implement the use of technologies like proteomics and machine learning to improve their processes.
Resources:
- Ramanan, M., Gielens, D. R. S., de Schepper, C. F., Courtin, C. M., Diepenbrock, C., & Fox, G. P. (2024). Environment found to explain the largest variance in physical and compositional traits in malting barley grain. Journal of the Science of Food and Agriculture. https://doi.org/10.1002/jsfa.13704
- Ramanan, M., Nelsen, T., Lundy, M., Fox, G., & Diepenbrock, C. (2023). Effects of genotype and environment on productivity and quality in Californian malting barley. Agronomy Journal, 115(5), 2544–2557. https://doi.org/10.1002/agj2.21433
You must be logged in to post a comment.