THESIS | PRESENTATION | ABSTRACT
This portfolio highlights a chapter from my M.S. thesis focused on estimating the heritability of liver abscesses in beef-on-dairy cattle. Featured components include project objectives, analytical approaches, and key findings, showcasing my experience in genetic evaluation, mixed-model analyses, and the application of quantitative genetics to livestock production challenges.
The objective of this study was to estimate the heritability of liver abscess incidence in fed, beef on dairy crossbred progeny. Liver abscesses affect approximately 20% to 30% of livers at slaughter. They are commonly identified during viscera inspection at slaughter due to difficulty with antemortem diagnosis. All livers with abscesses are condemned, causing a profit loss for producers. This loss is increased when the abscesses are ruptured or adhered to the gut due to additional condemnation of the gut. The most common prevention techniques include antimicrobial feed additives and nutritional management; however, mitigation has not been highly effective.
As the lead graduate student, I was responsible for data cleaning, analysis, and final interpretation. This project was a portion of my M.S. thesis. I have presented these findings at national scientific and producer-oriented meetings. A manuscript is currently in process. The project involved collaborating with various stakeholders, including station managers, accessibility advocates, and passenger groups.
1,860 beef on dairy heifers were fed and harvested in Kansas. The heifers were a part of a West Texas A&M study, so all were assigned a feed-additive treatment. Treatments were not of statistical interest in our study, but accounted for in contemporary group. Liver abscesses were assigned at slaughter using the Elanco scoring system (0, A-, A, A+, A+AD, and A+O). Heifers without liver abscess score and sire information were removed from analysis, leaving 1,492 heifers remaining.
A 3-generation sire pedigree was used in analysis. All heifers had recorded sires, but no recorded dam. There were only 19 unique sires across the 1,492 heifers. The lack of dam information and limited sire variability were significant limitations of the study, but are reflective of common dairy management protocols.
Data analysis was conducted using the Linux-based Colorado State University Center for Genetic Evaluation of Livestock high-performance computer, Animal Breeder’s Tool Kit, and ASREML 3.0. Additional data analysis and visuals were produced using R. The methodologies included the assessment of assumptions, mixed-effects models, and variance component estimation.
For Base Model. A. Residuals versus fitted values plot for linearity and homogeneity of variance assumptions. B. Sample quantile deviations versus standard normal distribution quantiles plot for assumption of normality of residuals.
For BRD Model. A. Residuals versus fitted values plot for linearity and homogeneity of variance assumptions. B. Sample quantile deviations versus standard normal distribution quantiles plot for assumption of normality of residuals.
The best fit model was the BRD model. While the Base model had a lower AIC value (AIC = 4996.07), the significant partial F-test for number of BRD treatments inclusion as a predictor variable (P < 0.05), and the decrease in normality of residuals violation indicate the BRD model may be a better fit. The BRD model is the model used throughout this study. When fit, contemporary group and age in days were not significant in the overall model (P > 0.05). Contemporary group and age in days were still included in the model to account for environmental differences between animals. Number of BRD treatments was significant in the overall model (P < 0.05).The model included fixed effects of contemporary group, age in days, and number of BRD treatments, with a random effect of sire.
The heritability estimate for liver abscess severity was 0 in this sample. In this data set, liver abscesses are not heritable and did not have a genetic component. Limitations revolve around the lack of sire variation, uneven spread of sires, and general difficulties in determining genetics of disease traits. Based on the results of this study, genetic selection tool development is still to be determined.
One significant challenge was the lack of variability and uneven distribution of sires in the sample. Producers of beef on dairy cattle are typically using artificial insemination, so the lack of variability was expected based on common dairy management protocols. To work around this challenge, I tested different subsets of data to better account for sires that had a similar number of progeny. The results and comparison of subsets can be seen in my Thesis.
Our study had, at most, 1,492 beef on dairy animals from the same feedlot and looked solely at phenotypic data with only sire information. The animals from the same originating producer contributed to the limitations in sire variation in this study. The difference between the results of these studies indicates that there is a need for further exploration of liver abscess genetics to determine if there is a non-zero genetic component.
The project highlighted the need for continuous research into traits that impact animal welfare, sustainability, carcass quality, and profitability. The data reflected common management practices, which provided difficulty in analysis. However, this project pushed me in my analysis skills by forcing additional checks and differing methods of verification due to the challenges. Disease traits are difficult to measure, especially when they are not diagnosable prior to death. Further research in the field is needed to comprehensively estimate the heritability of liver abscesses and determine future selection possibilities.
Incorporate Genomic Data: Integrate genomic data into the analysis for more accurate analysis.
Increase Sample Size: With a larger data set that has more variation in sires, the results would be more accurate and representative of beef on dairy cattle.
The results of this study indicate no genetic component linked to liver abscess phenotype. Genetic selection feasibility must be determined following the continued study of liver abscess heritability.
For a deeper dive into the methodologies and data, please feel free to contact me at mirandazuvich@gmail.com. Data and code are not publicly available.