SELECTION FOR OR AGAINST FACIAL ECZEMA SUSCEPTIBILITY IN ROMNEY SHEEP, AS MONITORED BY SERUM CONCENTRATIONS OF A LIVER-ENZYME

Genetic selection for or against susceptibility to facial eczema (FE) was begun in Romney sheep in 1975, with the establishment of a resista nt (R) selection flock, a susceptible (S) selection flock, and later a control (C) flock. For all but the initial years, rams were identifie d by performance testing with a sporidesmin challenge, ranking them on relative elevation of the liver enzyme, gamma glutamyltransferase (GG T) measured in serum. A different dose rate of sporidesmin was used fo r performance testing in the R and the S flocks, with a balanced half of the C-flock animals being tested at each dose rate; in some years R -flock animals showing no elevation of GGT were re-dosed later at an e ven higher rate. Mixed-model methodology was used to determine respons es to selection, expressing results as a breeding value for log(e)GGT. Analyses took account of one dose rate used in the R flock and in hal f of the C flock, and a second (lower) dose rate used in the remainder of the C flock and in the S flock. The heritability estimate for log( e)GGT was 0.45 +/- 0.03. Results through to the 1993-born lamb crop sh owed that the R flock became more resistant (i.e. showed reduced log(e )GGT) at a rate 1.77 times faster than the rate at which the S flock b ecame more susceptible. The two flocks differed in log(e)GGT in 1991-9 3 by 2.73 +/- 0.28 log(e)GGT units (3.03 phenotypic standard deviation s or 4.52 genetic standard deviations). The C flock mean was closer to that of the S than the R flock, indicating slower progress in the S t han the R flock. From the relationship between the percentage change i n animals resistant and log(e)GGT (i.e. -38% per unit increase in log( e)GGT), it was shown that a dose which just failed to induce any R-flo ck reactors would lead to all S-flock animals reacting to the dose. Co mparing annual rates of response by theoretical calculation and by bes t linear unbiased prediction breeding values showed that the latter we re 60-70% of the theoretical value, and possible reasons for this are discussed. The realised response represents a major difference between flocks in likely liver injury from natural FE challenge in New Zealan d's at-risk regions.