Multitrait estimation of genetic parameters of lactation curves in Holstein heifers.

Weekly milk yields of 1022 Holstein heifers from 61 sires were used to derive coefficients of the lactation curves using modified gamma and inverse polynomial functions. The natural logarithm of a modified gamma function was ln(yn) = ln (a) + b ln (n) + cn + u sin (x) + v cos (x), where a, b, c, u, and v are coefficients to be estimated; n is the day of lactation; and x is the day of year. Estimates of a, b, and c were combined to define persistency [-(b + 1) ln c], week of peak yield (b/c), and peak yield [a(b/c)be-b]. The inverse polynomial function was n/yn = A0 + A1n + A2n2, where A0, A1, and A2 are coefficients to be estimated. Variance and covariance components for the coefficients of the lactation curve were estimated by the multitrait restricted maximum likelihood method using canonical transformation. Heritability estimates were ln (a) .11, b .07, c .04 u .01, v .04, A0 .28, A1 .26, A2 .21, persistency .21, week of peak .18, peak yield .23, and 308-d milk yield .41. Genetic correlations indicated that selection for faster rate of increase to peak production would result in higher 308-d milk production, higher peak yield, and greater persistency.