Prediction of total protein percentage from partial sampling and adjustment of environmental effects

Objective. To find a mathematical expression for calculating the average percentage of total protein (APTP) from the partial production (PP) and other environmental factors that affect this characteristic. Materials and methods. The study was conducted on three dairy farms in the department of Antioquia, Colombia. 182 Holstein cows were sampled; the data capture was carried out through monthly sampling of the two daily milking sessions. Milking activity beginning time, milk production, number of births produced, season of parturition and days in lactation were recorded. The statistical analysis was conducted using a multiple regression model which identified significant sources of variation on the total daily protein content. Based on the estimated regression coefficients a prediction model for the aforementioned variable was generated. Results. The number of births produced, days in lactation, partial protein production (pm), milk production, quadratic expression for partial protein production am, partial protein production (am) and milking intervals had significant effects (p<0.05) on the total daily protein percentage content. Two models for APTP prediction from a partial sample (am and pm) were generated from these effects. Conclusions. The APTP is affected by various environmental factors that must be adjusted for prediction models. These models can be implemented to adjust data from a partial sample (am or pm) and further adjusted to a reliable average daily protein production in genetic breeding programs.