Genetic parameter estimation and implementation of the genetic evaluation for gaits in a breeding program for assisted-therapy in donkeys.

Genetic analyses in donkeys are likely to face compromises in terms of sample size and population structure. This study aims at implementing a suitable model to estimate breeding values and genetic parameters for gaits in Andalusian donkeys. Empirical observation revealed that ambling donkeys (showing a slightly uneven, non-isochronous 1-2, 3-4 lateral sequence gait) did not walk (i.e. presented an isochronous, even 1-2-3-4 sequence gait) and vice versa. However, the two donkey groups could trot, equally. In this study, 2700 gait records were registered from 300 donkeys. The sample included 1350 gait records from 169 ambling/trotting donkeys and 1350 gait records from 131 walking/trotting donkeys. Fixed effects included year, season, sex, farm/owner, husbandry system, weather, ground type and appraisers. Weight and age were included as covariates. MTDFREML software was used to estimate (co)variance components, genetic parameters and predict breeding values and their accuracies in both sets, separately. Gaits' heritability ± SE estimates were 0.56 ± 0.155, 0.53 ± 0.317 and 0.67 ± 0.166 for amble, walk and trot, respectively. Genetic correlations were 0.31 ± 0.216, 0.42 ± 0.115 and 0.28 ± 0.178, for amble and walk, amble and trot and walk and trot, respectively. Not all gaits are suitable to treat every human sensomotor condition. We developed a locomotion selection index, assessing the relative loss/gain in index accuracy when each gait modality was excluded to develop different gait specific therapeutic lines to genetically select the best performing donkeys from each gait modality. Our results suggest that gait genetic lines could be developed and may be potential selection criteria to consider in assisted-therapy donkey breeding programs.