Horn flies are a major nuisance to cattle and induce significant economic losses. Fly abundance varies within and across breeds and genetic analyses have shown sufficient genetic variation to permit selection. A major bottleneck for selecting against horn fly abundance is the complexity of measuring fly attraction phenotypes. Easy-to-measure proxy phenotypes could be an attractive option to indirectly estimate fly abundance. In the current study, thrombin was investigated as a potential proxy to assess fly abundance. Fly counts and blood samples were collected on 355 cows. Pearson correlation between subjective fly count and thrombin was −0.13, indicating a decrease in fly abundance with the increase in thrombin concentration. When thrombin was discretized into three classes, there was a 22% difference in fly count between the top and bottom classes. Heritability estimates of thrombin were 0.38 and 0.39 using linear and threshold models, respectively. The correlation between estimated thrombin breeding values and fly count was around −0.18. There was a noticeably lower density of high fly counts among animals with high breeding values for thrombin. These results indicate that thrombin could be used in combination with other biological factors to estimate fly abundance and as a proxy for selection against fly abundance.

The genetic basis of horn fly abundance remains largely unknown and only few heritability estimates are available based on small scale studies. In this study, one subjective and two image-based fly abundance phenotypes were analyzed. Each animal was assessed subjectively for fly abundance by at least one trained agent (SUB). Additionally, several digital images were taken and the best image for each animal was used to assess fly abundance. A box that runs roughly from the withers to the hook and from the chest floor to the fore of the udder was established and all horn flies (PA) and sampling-based estimated number of flies (PR) within the box were calculated. Horn fly counts (SUB, PA, and PR) were discretized into 4 classes based on the quartiles of their respective distributions. Heritability estimates ranged between 0.10 and 0.15 and between 0.14 and 0.16 for the continuous and discrete cases, respectively. The genetic correlation was 0.66 between PA and PR and decreased significantly between the image-based (PA and PR) and the subjective (SUB) assessments for the continuous case. For the discrete case, the genetic correlations between the three traits were very similar, indicating a high concordance between traits.