Evaluation of classifiers that score linear type traits and body condition score using common sires

R.F. Veerkamp, C.L.M. Gerritsen, E.P.C. Koenen, A. Hamoen, G. de Jong

    Research output: Contribution to journalArticleAcademicpeer-review

    33 Citations (Scopus)


    Subjective visual assessment of animals by classifiers is undertaken for several different traits in farm livestock, e.g., linear type traits, body condition score, or carcass conformation. One of the difficulties in assessment is the effect of an individual classifier. To ensure that classifiers rank animals consistently, i.e., the repeatability between classifiers and within classifier, genetic links across routinely scored observations may be used to validate scoring of individual classifiers. Eighteen classifiers of NRS scored 18 traits, and body condition for 91,589 first-lactation heifers, daughters of 601 sires. Genetic parameters were estimated in a series of bivariate analyses. In turn, observations of each individual classifier were trait 1 and all observations of all other classifiers were grouped as trait 2. Likelihoods were used to test whether additive genetic or residual variances for each classifier (trait 1) differed significantly from the grouped records (trait 2), and to test whether the genetic correlation between trait 1 and trait 2 was significantly smaller than unity. Arbitrary criteria were set to mark traits for individual classifiers when a significant deviation was found: genetic correlations of 0.40, and more than 15% deviation for the standard deviation. One classifier had relatively low heritabilities, but high genetic correlations with the others. This might indicate that the repeatability within classifier should be improved. Another classifier had high genetic correlations with the others, but his sire variances were significantly higher than average for most traits. For the genetic correlations, each classifier averaged 3.3 traits marked, ranging from 0 to 9. Overall feet and legs, rump width, central ligament, and foot angle received most marks (12 to 6 classifiers), but no disagreement existed on the definition (i.e., no mark) for body condition score, stature, rump angle, teat length, overall udder, and teat placement. These simple and cheap marks can be used in training sessions to improve the quality of the scoring system.
    Original languageEnglish
    Pages (from-to)976-983
    JournalJournal of Dairy Science
    Issue number4
    Publication statusPublished - 2002

    Fingerprint Dive into the research topics of 'Evaluation of classifiers that score linear type traits and body condition score using common sires'. Together they form a unique fingerprint.

    Cite this