Statistical analysis Cow Dataset The association between each SNP and performance was quantified using linear mixed models in ASREML [41] with an additive genetic effect and permanent environmental effect included as random effects.
Relationships among individuals were accounted for using a numerator relationship matrix. Fixed effects included in the model were contemporary group defined as experimental treatment-by-year, parity and breed. The ratio of the additive genetic to phenotypic variance and permanent environmental variance to phenotypic variance were fixed to give a heritability and repeatabil- ity of 0.35 and 0.5 respectively for milk, fat, protein, and lactose yield; 0.52 and 0.6 for fat and protein percentage and 0.11 and 0.13 for SCS, subclinical mastitis and peak SCS. SNPs were included in the model as either contin- uous variables or as class effects. SNPs were included in the model either individually or together. Significance was based on the F-statistic.