The use of fixed and random regression test day models in genetic analysis of daily milk yield of Mashhad's Holstein cows

To compare genetic evaluation methods for milk yield of Mashhad’s Holstein cows, two fixed and random regression test day models were used. The data set consisted of 95,510 test day milk records collected from 11,054 first parity Holstein cows calved between 1994 and 2007 Mashhad’s dairy farms. The number of sires, dams and total animals in pedigree file were 807, 8,835 and 11,054, respectively. Average daily milk was 26.92±6.91 Kg. Fixed and random regression test day models had fixed effects of herd – year – season of recording - milking times - sperm origin, as well as covariables of Holstein gene percentage, stage of pregnancy, age of cow at recording, and daily temperature. To take into account the shape of the lactation curve at phenotypic, genetic and environmental levels, orthogonal Legendre polynomials (order 4) was implemented in the random regression test day model. The models were fitted to the data by DXMRR software. There was a high positive and significant (P<0.001) rank correlation coefficient (0.939) between predicted breeding values (PBV) obtained by two models. Averages PBV were found to be 0.1695 and 0.1524 Kg for fixed and random regression test day models, respectively which were significantly (P<0.0001) different from each other. Genetic trends for daily milk were 0.044 and 0.077 Kg annually for fixed and random regression test day models, respectively.