Heritability estimates of liabilities to digestive diseases in heavy racehorses

Data on 1410 heavy racehorses were taken from the Hokkaido Banei Horse Racing Result Book published during 1997–2000. Heritabilities were estimated using the restricted maximum likelihood method with a full data set of the accumulated prize winnings and the frequency of medical treatment against digestive diseases. Heritability estimates for the same traits were also obtained using a partial data set from affected animals only. In addition, a ‘two classes, one threshold’ model was applied to the full data set for the heritability estimation of liability to the diseases. Heritabilities estimated by restricted maximum likelihood with the full data set were quite low: 0.09 ± 0.02 for the prize winnings and 0.08 ± 0.03 for all three digestive diseases – colic resulting from flatulence, constipation and overfeeding. Low heritabilities were also estimated for each of the three kinds of colic. Medium heritabilities were estimated for colic resulting from constipation (0.30 ± 0.19) and overfeeding (0.46 ± 0.22) using the partial data set taken from the affected animals. Using the threshold model, higher heritabilities from 0.42 ± 0.04 to 0.57 ± 0.03 were estimated for the liabilities to all three digestive diseases and to the individual digestive diseases. These results suggest that liability evaluation of the frequency of medical treatment is insufficient because all non-affected animals are scored as 0. The high heritabilities estimated by the threshold model with half-sib relationships should be verified in future studies by comparing them with estimates from other genetic relationships and with realized heritabilities from a small-scale practical selection program.