Table 2 Trait definition for the resilience traits

lnvar_weight

The natural logarithm of the variance of pigs’ daily differences between observed weights versus expected weights via Gompertz modeling of weight versus age (example shown in Fig. 5). A higher value indicates more deviations and, hence, a lower resilience

lnMSE_weight

The natural logarithm of the mean squared error (equivalent to variance) of pigs’ daily differences between observed weights versus expected weights via linear modeling of weight versus age. A higher value indicates more deviations and, hence, a lower resilience

lnvar_{weight_standardized}

The natural logarithm of the variance of a pigs’ standardized weights versus age (mean is zero, standard deviation is one; Fig. 4b, Fig. 5c and g). A higher value indicates more deviations and, hence, a lower resilience

Skew_weight

The skewness of pigs’ daily differences between observed weights versus expected weights via Gompertz modeling of weight versus age

Lag1_weight

The lag1 autocorrelation of pigs’ daily differences between observed weights versus expected weights via Gompertz modeling of weight versus age

Straightness

The straightness index, estimated after trajectory analysis of a pigs’ observed weight versus age. Straightness index is estimated as the Euclidean distance between start and end point divided by the total path length covered by the weight trajectory. Maximum value is one (straight line), minimum value is zero (infinite body weight deviations). A lower value indicates more deviations and, hence, a lower resilience

Mean speed

The Mean speed, estimated after trajectory analysis of a pigs’ observed weight versus age. Mean speed is estimated as the total path length covered by the weight trajectory divided by the age difference (d) between end and start. A higher value indicates more deviations and, hence, a lower resilience

lnMSE_FI

The natural logarithm of the mean squared error (equivalent to variance) of pigs’ daily differences between observed feed intake versus expected feed intake via linear modeling of feed intake versus age. A higher value indicates more deviations and, hence, a lower resilience

lnMSE_dur

The natural logarithm of the mean squared error (equivalent to variance) of pigs’ daily differences between observed visit duration versus expected visit duration via linear modeling of visit duration versus age. A higher value indicates more deviations and, hence, a lower resilience

lnMSE_{n_visit}

The natural logarithm of the mean squared error (equivalent to variance) of pigs’ daily differences between observed number of visits versus expected number of visits via linear modeling of number of visits versus age. A higher value indicates more deviations and, hence, a lower resilience

QR_FI

The number of off-feed days, calculated as the number of days during which feed intake was in the 5% lowest quantile using quantile regression on age over all pigs. A higher value indicates more off-feed days and, hence, a lower resilience

QR_dur

The number of off-feed days, calculated as the number of days during which visit duration was in the 5% lowest quantile using quantile regression on age over all pigs. A higher value indicates more off-feed days and, hence, a lower resilience